Making Websites Win CRE
Making Websites Win CRE
Making Websites Win CRE
Why this is the only book of its kind (and why it will transform
your website, your business and your career, like it has ours) ..............30
Choose your role models wisely: Take inspiration from this awe-
inspiring story of how one of our clients went from zero to half
a billion dollars .....................................................................................................60
Why the world’s top websites are winning (and how yours can
too)............................................................................................................................62
Why you should avoid meek tweaking (and not just because it
sounds creepy) ......................................................................................................72
Installing a tag manager (to help you activate and deactivate tags
without having to speak with your IT department each time) ...............88
Diagnose by…using live chat (to let your visitors tell you what’s
wrong with your pages) ......................................................................................97
Diagnose by…using exit survey tools (to ask your visitors why
they didn’t take action) ................................................................................... 107
Diagnose by…using on-page survey tools (to ask questions at
exactly the right moment) .............................................................................. 109
Our podcast
You can hear some of our most popular talks by subscribing
to the Conversion Rate Experts podcast. You can find details
at www.conversion-rate-experts.com/podcasts/.
The free guides accompanying this book help you get started quickly.
“They helped us get our CRO program off the ground. We now
have a much deeper understanding of some of the key customer
pain points and, more importantly, a well-defined road map on
how to address them.” —Hertz, the car-rental company with
locations in 150 countries.
363% increase in conversion rate for Crazy Egg, the leading click-
mapping platform. “Instead of saying, ‘We want to do this, or we
want to do that, or I think this would be best for the business,’ they
got data to back up their decisions and they did that. Which is
why we had such a substantial lift.” —Neil Patel, founder. “They
focused on understanding our customers. We believe that led to
the increase in conversions.” —Hiten Shah, founder.
“If you don’t have a solid conversion rate plan, I would definitely
advise you to get in touch with Conversion Rate Experts.” —Dell,
the multinational computer technology company.
“Conversion Rate Experts are by far and away the number one
in the world. They have helped us transform our business and
our processes around testing. As a direct result of working
with Conversion Rate Experts, we’ve doubled our revenue.”
—HomeFinder, a leading real-estate website.
“I went into this project fairly skeptical that we’d be able to achieve
a 30%+ improvement. At 50%+, it far exceeded our expecta-
tions.” —AAG, America’s largest provider of government-insured
reverse mortgages.
“Sales have increased from £16 million to £31 million per year.”
—Sunshine.co.uk, an online travel operator.
“They are brilliant. The results speak for themselves. Sales nearly
tripled.” —Broadband.co.uk, a leading comparison website for
broadband services.
foreword 25
power dynamic in most organizations, where decisions are
based not on customer needs but the opinions of the person
with the most senior job title. (In a speech in 2006, I framed
this as HiPPO-driven decision-making—HiPPO standing
for the Highest Paid Person’s Opinion.) Experimentation
reduces that power dynamic with amazing results.
In addition, I deeply believe in heuristic evaluations, in
which a broad cross section of company employees use a
website or mobile app as customers. Heuristic evaluations
create a democracy of ideas, capturing insights from beyond
the User Experience team and Digital team. The only way
to allow this democracy to thrive is to build an experimen-
tation model—to put all the ideas through testing.
In my experience, most companies operate as if the
world is static. As if tastes don’t change, preferences don’t
evolve, and disruptive competitors don’t show up. In reality,
change is the only constant. Experimentation assumes this is
true and fosters a culture of mental agility that constantly
explores boundaries (and pushes against them).
Finally, most people and companies are risk-averse.
What I love the most about experimentation is that it allows
us to accommodate for the short-term risk we are willing
to tolerate. You may be deeply shy, your company may be
a cash cow, and even the slightest change to the ecosystem
may mean death, but that’s no problem. Experimentation
allows you to dial-in the level of risk you want to take (you
may even choose to show a test to only a small fraction of
your visitors). On the other hand, if you realize that taking
foreword 27
SECTION 1
Why most
web design
is done
wrong
—and how to do it right,
like the winners
Sales went through the roof the first time we applied A/B testing to a company.
Rather than rescaling the axes, we drew on the wall, hoping to penetrate the ceiling
tiles. (Those dashed lines represent interim forecasts we kept beating.)
• In nine languages
• In twenty-two countries worldwide: Australia,
Canada, China, Denmark, Finland, France, Germany,
India, Ireland, Israel, Italy, Japan, Lithuania, Malaysia,
New Zealand, Norway, Serbia, Spain, Sweden, United
Arab Emirates, United Kingdom, and United States
• Of all sizes—from startups to large enterprises
• In business-to-consumer (B2C) and busi-
ness-to-business (B2B)
• In more than eighty industry verticals, including
finance, health, retail, travel, technology, leisure, and
food (you name it, we’ve sold it)
• With different types of product: selling physical goods,
services, software, and information
• With all types of business: merchants, affiliates,
publishers, social networks, e-commerce and lead-gen-
eration sites
• In many media and formats: desktop websites, mobile
websites, native mobile apps, email marketing, offline
advertising, and more
In short, the skills you learn from this book will work for
you now, and for whatever you do next.
You could then tally the orders for each headline and deter-
mine which headline brought you the most.
A/B testing software makes it easy to carry out such
tests. Each of your visitors will see a different version of
the page—Version A or Version B, or even Version C, D, or
E—and then the software will work out, on average, which
of the elements performed the best. The winner can then
be promoted.
In most tests, Version A is the existing version, often
called the “control,” which you are trying to beat.
On every webpage, there are many changes you could make to increase your profits.
So what is CRO?
Back in 2005, when we worked in-house, we tried hard to
outsource conversion. It turned out to be impossible. There
were many vendors, but each pushed its own solution:
• The live chat companies claimed that live chat was the
answer.
• The analytics companies claimed that analytics software
was the answer.
• The usability companies claimed that usability was the
answer. CRO não é formado ou definido por apenas
um aspecto, trata-se de uma análise holística
das plataformas
And so on.
What we wanted was an independent advisor who was
effectively a “hub” for conversion, who had used all of the
services, who was completely impartial, and who knew how
to get the best results.
We ended up doing it ourselves. As a result, we learned
what worked. We now spend all of our time designing pages
that convert—and then we put our necks on the line by
insisting that we test our creations, to prove that we’ve
measurably improved things. Over the past ten years, we
have tried an enormous number of techniques, some of
which were real gems. We have pulled together the best
ones (along with practical, easy ways to get them done) and
added many techniques that we developed ourselves. At
each stage, we have shared our findings on our blog, which
has rewarded us by attracting kindred spirits, people who
share our passion. Hopefully, you’ll be one of them.
We coined the term conversion rate optimization (CRO)
in 2007 to describe this process of pulling together all of
the available tools, techniques, and skills—with the goal
of improving a website’s conversion rates. CRO takes the
The Power Law of CRO, which explains why so many of our clients have won awards
for fast growth.
(2) Prove you’ve done so. A/B testing allows you to do this.
As such, CRO makes for a fantastic career choice.
There’s a serious shortage of people who have proof that they
can grow businesses. Amazingly, most people’s résumés
have no quantifiable evidence that they have ever created
value. Thanks to CRO, many of our clients, followers, and
team members have incredibly impressive success stories
on their résumés.
Plus, if you’re anything like us, you’ll find it end-
lessly fascinating.
• ads by 19%;
• landing page by 19%;
• shopping cart by 19%; and
• checkout by 19%.
“Being able to figure out quickly what works and what doesn’t
can mean the difference between survival and extinction.”
—Hal Varian, Google Chief Economist
Diagnosis
The ultimate guide to tools and
techniques to understand how
your website can be improved
diagnosis 81
A goldmine of techniques
In 2008, we published the most comprehensive article about
how to understand your website’s visitors. Back then, there
were hardly any tools for doing it: our list contained just
fourteen. As evidence of how much the web has changed
since then, this section contains 200 resources—software,
techniques and UX tools for finding out exactly why your
potential customers aren’t converting.
It’s pure gold.
diagnosis 83
How we’ve categorized the following techniques
It used to be that each software solution carried out a specific
function. Now, they all tend to be turning into all-purpose
suites, making them hard to categorize. We have grouped
each software solution into the functionality for which
we most often use it.
Mobile web usage has skyrocketed, but the tools for
understanding visitors have been slow to catch up. It’s finally
happening, though. Throughout this series, we will label
some of the tools as being “mobile-friendly.” That doesn’t
mean that the others aren’t; it just means that either the tool
has been specifically built for mobile, or that we have expe-
rience using the tool on mobile devices, and it works well.
Some of the techniques require a lot of visitors. So before
we dive in, let’s take a quick tour of which of them you should
focus on if you have a low-traffic website.
diagnosis 85
test it on people who are from your target demographic
and psychographic.
• Watch session recordings of the visitors you have.
Doing so will give you insight into how web visitors
see your website. Plus, you’ll see your creation through
fresh eyes.
• Speak to salespeople (what we call “VOC Aggrega-
tors”)—people who have sold face-to-face the same
type of product—or similar products.
• Analyze competitors’ websites. Or, if you don’t have
any obvious competitors, look at companies that are
successful within adjacent fields. For example, if you
sell B2B software, look at other B2B software vendors.
• Add your phone number prominently to the top of
every page. Even if you have no plans to encourage
phone calls on an ongoing basis, it can help to get at
least a few of them. In fact, you may be able to charm
your early callers into becoming long-term user testers.
• Increase the incentives for visitors to complete surveys.
The more you offer as an incentive, the higher percent-
age of responses you are likely to get.
diagnosis 87
Installing a tag manager (to help you
activate and deactivate tags without having
to speak with your IT department each time)
This section will no doubt persuade you to install multiple
tools on your website. You may get tired of asking your
developers to activate, manage, and deactivate tags. So
before we continue, you may benefit from installing a tag-
management solution. Tag-management solutions provide
marketers with an easy-to-use interface that doesn’t require
the user to have IT skills or IT permissions.
the visitors came. They show the path that visitors took
through the store, but they don’t reveal what the visitors
were thinking. They show you exactly where and when the
visitors left the store, but not why. Or what to do about it. For
that, you’ll need qualitative tools, which are described later.
You’ll find web analytics most useful in the early stages
of a project, when you are seeking to identify on which
pages to start work. It will also inform the pages on which
you should implement the tools described in the rest of
the chapters in this section. If a page gets no visitors, then
changing it will have no effect. Nor will changing a page that
already has 100% conversion rate. Web analytics software
will help you to identify the arteries of the website—the
high-traffic flows that lead to successful conversions. Along
with other tools, it can also help you to spot the aspects of
those flows that are currently underperforming.
diagnosis 89
over a long period. Amplitude helps marketers to under-
stand how users behave within a website or app.
Crazy Egg’s confetti map shows exactly where visitors clicked—even if it wasn’t on
a link.
diagnosis 91
of the page—then a scroll map will reveal that visitors
aren’t scrolling. (You then need to work out whether
that’s because they didn’t realize that the page could
be scrolled, or because they weren’t interested enough
to scroll.)
Hotjar’s session recordings allow you to see how your visitors struggled their way
through your website.
diagnosis 93
on attributes such as the visitor’s country of origin, how
much time they spend on the site, or the number of
pages they visited. You may choose to watch videos of
visitors who appear to be struggling—for example, those
who visit the same page several times.
2. Get a feel for how people use websites: Session-re-
cording tools are not a substitute for carrying out user
tests, which are described later. However, watching a
few videos will give you a better idea of how people
interact with websites.
3. See errors: The software can display a report of errors
that users have encountered.
4. Analyze funnels: Get to see where your visitors are
dropping off. Clicktale, in particular, makes it easy to
study funnels for opportunities.
5. See scroll maps: Scroll maps reveal how far down your
page visitors scrolled.
diagnosis 95
motivated—it may fatigue them, causing them to aban-
don later.
• Which fields tend to get left blank. A blank response
often indicates that a field is confusing or intimidating.
Such fields reduce a visitor’s resolve to complete the
form.
• Which fields result in error messages, which the visitors
then need to edit before they can resubmit the form.
• Which browsers and devices are performing poorly.
Maybe your form is hard to use on small mobile devices.
Live chat software helps in several ways: you can convert visitors via the
chat conversations, you can discover their objections, and you can test your
counter-objections.
Live chat can allow you to hear from visitors who wouldn’t
phone you. Such visitors might prefer live chat for some of
the following reasons:
diagnosis 97
• They appreciate that, unlike phone calls, live chat doesn’t
cost money.
• They don’t want to be stuck at the end of a phone waiting
for someone to answer.
• They feel that a live chat session is less of a commitment
than a phone call.
diagnosis 99
Awesome, tested questions to ask your
customers
The following questions can provide invaluable insights.
We recommend you identify the ones that will populate
the gaps in your existing knowledge.
100 m a k i n g w e b s i t e s w i n
would you describe us to a friend?” reveals why your
customers like you. Similarly, you could ask a question
along the lines of the following: “Which other options
did you consider before choosing our product or
service?” or “Why did you decide to use us?” It’s
particularly important to ask questions from this group
before you undertake any re-branding exercise, so you
understand what your existing positioning is.
• “Do you use us for all your [ProductType] or do you
also use alternative companies? If so, why?”
• “Why do you use [YourCompanyName or YourPro-
ductName] rather than the alternatives?”
• If you want your customers to use you more often, you
could do worse than to ask them “What would per-
suade you to use us more often?” Can you see how
survey questions can save you a lot of trial and error?
• “How could we persuade your friends or colleagues
to use us?” “If you were in charge of our company,
how would you persuade people like yourself to use
us?” and “If you were in charge of our company, how
would you spread the word about us?”
diagnosis 101
fact, it almost always reveals an easy way to increase a com-
pany’s sales. It works like magic on any business—online
or offline.
We arrived at it after seeing a TV program about a
military assault course—a series of rope swings, climbing
nets, and muddy pools. The TV presenter was interview-
ing people who had just finished the course, asking them
what had been the worst obstacle. Several of the finishers
mentioned a tall wall that had been tough to climb. The
scene then cut to a view of that wall, revealing that, sure
enough, many people were still stuck behind the wall. After
several attempts at climbing it, they were giving up and
dropping out.
We wondered how we could adapt that question to sales
processes, to identify the conceptual “walls” that prevent
prospects from buying.
We devised the question, asked it of our clients’ cus-
tomers, and discovered that it has an incredible ability to
identify obstacles—and hence opportunities for growth. The
question is so valuable, it deserves its own page (but it’s not
going to get one). The question is…rabba dabba dabba (this
is supposed to be a drum roll) dabba dabba dabba…TISH!
102 m a k i n g w e b s i t e s w i n
You can also ask the question in many ways. For example,
you can use a survey on your website’s thank-you page or
even email the customer a link to the survey. In other cases,
it’s worth asking the question when speaking to custom-
ers—either face-to-face or on the phone.
There are some important subtleties and realizations
as to why the question works so well.
diagnosis 10 3
seconds. His or her opinion—however strong—is much less
likely to be accurate. When your buyers criticize you, they
are likely to be right.
104 m a k i n g w e b s i t e s w i n
• design the most effective sales funnel, so your most-
liked products aren’t hidden away; and
• improve your existing products, to make purchasers
more likely to buy from you again.
diagnosis 10 5
Learn more about questions
For more about the subject of asking questions, visit
our website and watch the talk “Golden questions” that
reveal exactly why your visitors aren’t converting (slides
and video).”
106 m a k i n g w e b s i t e s w i n
Diagnose by…using exit survey tools (to ask
your visitors why they didn’t take action)
Visitors don’t know why they abandoned until after they abandoned. Exit surveys
allow you to gather insights that couldn’t have been gathered any sooner.
diagnosis 10 7
If your website has multiple distinct segments of visitors,
you may choose to add a question that reveals the visitor’s
segment too.
108 m a k i n g w e b s i t e s w i n
Diagnose by…using on-page survey tools (to
ask questions at exactly the right moment)
Exit surveys allow your visitors to tell you why they abandoned your website.
d i a g n o s i s 10 9
Some information can best be obtained by surveying your
visitors while they are on a specific page. In doing so, you
can ask questions at the exact moment that the visitors are
thinking the thoughts you want to hear. On-page survey
tools let you add surveys to the corner of a page or, more
prominently, to its middle, obscuring the other content.
Several options are available:
“Give Feedback” buttons allow your visitors to leave feedback at a time of their choosing.
diagnosis 111
Tools for “Give Feedback” buttons
You can easily add a “Give Feedback” button using one of
the following solutions: Survicate, Feedback Lite, Feedback-
ify, Usabilla, Qualtrics, Medallia Digital (mobile-friendly),
and SurveyGizmo (the “Give Feedback” button is one of
its many features). For Conversion Rate Experts’ website,
for reasons of leanness, we don’t use a hosted software
solution; we simply link our “Feedback” button to the form
on our “Contact Us” page.
The software hosts the feedback forms, and then allows
the website’s creators to view—and manage—the responses.
(In several of the apps, the feedback is managed using
an interface similar to that of an email client, having an
inbox and folders.) If visitors leave their email addresses,
the website’s customer-support team can easily reply to
the feedback.
diagnosis 113
Diagnose by…using search engines (to be
notified when people say things about you)
Search engines allow you to track what the world is saying about your company and
products. Moz Fresh Web Explorer is great for this.
diagnosis 115
surprise, the other tuning fork had started to vibrate and
was giving off the same musical note.
Caples observed that emotional vibrations work in the
same way; they travel from one person to another. When
you write, the tuning fork within you must vibrate at the
same frequency as the tuning forks within your visitors.
So before you can write compelling copy, you must emo-
tionally empathize with your visitors.
It isn’t easy, though. As a web marketer, your work is
almost intrinsically ivory-tower work. It’s hard to even meet
your visitors, never mind empathize with them. So how do
you develop this emotional resonance?
Quantitative feedback tools—like clickstream analytics—
don’t help, because they just give you numbers. Qualitative
feedback tools—like surveys—can give you the voice of the
visitor, but they aren’t sufficient to turn you into a tuning
fork. Several of the following techniques (method marketing
plus to a lesser extent face-to-face selling, encouraging vis-
itors to phone you, and talking to “VOC aggregators”)—are
what we call ultra-qualitative. They are at the core of how
we approach conversion. You don’t hear much about them
because no one has found a way to package them into neat
little subscription services. You’ll never see ads for them.
That doesn’t make them any less valuable.
• being a parent;
• having a migraine;
• orgasm;
• being drunk;
diagnosis 117
• flying;
• being blind;
• having a loved one die;
• spiritual enlightenment;
• meditation;
• hallucinogens;
• being gay;
• holding your newborn child;
• being underwater;
• the tastes of certain foods;
• being dumped;
• having an addiction;
• having depression, anxiety, stress, or any other type of
mental illness;
• being an identical twin; and
• falling in love (incidentally, Norwegians have a word,
Forelsket, which means “the indescribable euphoria
experienced as you begin to fall in love”).
To sell Dysons, use Dysons. As part of our research program, we use our clients’
products.
diagnosis 119
Getting under the skin of the customers.
1 20 m a k i n g w e b s i t e s w i n
about going anywhere near them. But during the unboxing
process, we noticed some valuable information that was
wrongly absent from the website.
diagnosis 12 1
If we aren’t the target customer, we find someone who is. At this point, marketing
becomes anthropology.
diagnosis 12 3
When growing a company that sold houses, we visited one of the company’s sales
offices and followed the customer journey of buying a house. (We stopped short of
buying the house, a decision we now regret.)
diagnosis 12 5
The great thing about weight-loss groups is that you can
speak with real customers. In fact, that’s what the meeting
is—a one-hour discussion group.
The following photos show Karl starting and finishing
a weight-loss program. He adopted the traditional before-
and-after poses.
Left: At the first weigh-in, doing the traditional “before” pose. Right: Not so humble
now, Karl strutting around like he owns the place, after proudly receiving his Slimmer
of the Month award. The activity gave us insights that we have used to grow many
clients in the weight-loss industry. (Karl’s subsequent fall from grace is not pictured.)
diagnosis 127
By opening and running Japan’s first Nokia store, we learned rich insights that
helped us to triple the sales of an online phone store.
1 28 m a k i n g w e b s i t e s w i n
your phone number until you arrive, so your friends and
family won’t know how to contact you.”
If you don’t have an offline store, create one. Flea markets are an easy way to get in
front of users quickly.
diagnosis 12 9
eficácia das ferramentas pois o
visitante/cliente não vai forjar um
comportamento que julga ser
melhor aceito socialmente
At first, we told our stall’s visitors that we were carrying out
market research and that we wanted to hear their feedback.
This had two problems: (1) people didn’t want to speak
about market research, and (2) those who did tended to
give responses that were polite and false. Only when we
tried to close the sale did we hear their true objections. So
from then on, we tried to sell in earnest (even though our
prime motive was to gather feedback). If a visitor showed
interest, we would take down their name so we could notify
them when the product became available.
This activity became a core part of our process for devel-
oping new products. Because our product was aimed at
travelers, we soon graduated from flea markets to airports,
where we would rent space for an exhibition stand.
1 30 m a k i n g w e b s i t e s w i n
which is particularly important for products that are
technical or complex. We have seen several products
for which the Wikipedia page was clearer and more
persuasive than the company’s official landing page.
(We once tried, unsuccessfully, to persuade a client to
point some of its AdWords traffic at its Wikipedia page.)
• Read reviews: Reviews on online stores like Amazon
reveal a lot about the buyers’ psychology—particularly
their likes and dislikes. Pay attention to the wording
used in the reviews.
• Look for places where customers sell the product to
others: There’s no purer copywriting than when buyers
recommend products to their friends. Search Twitter
and Facebook for instances of customers recommending
your product to their followers. If your company has
a tell-a-friend program that allows the customers to
send a personalized message to their friends, then those
messages can be a great source of sales copy. (Check
first that you are OK to analyze those messages without
breaking any laws or terms and conditions.) Not only do
the messages reveal the persuasive arguments, but they
also tell you the exact wording that customers use. By
analyzing the frequency of words used, you can establish
the lexicon you should use in your copywriting.
diagnosis 13 1
Diagnose by…talking to a “VOC aggregator”
(perhaps the fastest way to understand users)
What VOC aggregators are, and why you need to
find yours
According to Paul Graham of Y Combinator, to grow an early
stage web business, you should spend all of your waking
hours on the following three activities:
diagnosis 13 3
While we were working with Sony, we visited a store that
sold its devices. The shop assistant was great at selling
the product.
She sold several of the devices each week, and she knew
the answers to all the questions that visitors asked. We asked
her for her views on the product’s website, which she knew
well because she had studied it when searching for answers
to her visitors’ questions. She described twenty-two facts
that her visitors needed to know but that weren’t mentioned
on the company’s website. We incorporated these facts into
the webpage, making it much more persuasive.
For many products, the salespeople hold decision
trees in their heads. What they say depends on how visi-
tors answered the previous questions. In such cases, you
should map out the decision trees, consolidate them, and
then turn them into conversion flows. We did exactly this
for a blue-chip financial services company. After speaking
with its call center team, we made changes that resulted in
a 214% increase in orders.
1 34 m a k i n g w e b s i t e s w i n
should be a distilled version of words and logic that the
salesperson says to the callers.
If the price of your product is low, you may find that it’s
economically unviable to invite phone calls. Regardless,
we still recommend you add a phone number for just a
day or so, for research purposes. The first few phone calls
can reveal breakthrough insights. Not only do phone calls
help you to understand your visitors, but they also help you
to understand yourself. A phone call can be an “intuition
pump”—on the phone, you’ll find yourself intuitively using
words and logic that you hadn’t thought to write on your
pages. Just as stand-up comedians come up with many of
their best lines while performing in front of an audience,
you’ll find that much of your best sales copy comes while
you are selling on the phone or in person, one-to-one.
diagnosis 13 5
Tools for knowledge bases
The following knowledge base solutions are popular: Help
Scout, Zendesk, KnowledgeOwl, Moxie, Intercom, Help-
juice, and Freshdesk.
1 36 m a k i n g w e b s i t e s w i n
your target market. However, at first, you can get great
results by testing on whoever’s at hand. Those people
sitting near you right now—they’ll do. Or a family member.
This practice has a name: hallway usability testing—which
makes it sound more legitimate than it feels when you’re
doing it. Hallway usability tests are effective because many
usability problems are so obvious they could be detected
by anyone.
Once you sense that you’re getting diminishing returns
from your hallway antics, and your website’s remaining
problems are too sophisticated to be detected by a layman,
start to look for test participants who are from the website’s
target demographic. For example, when we doubled the
sales of a web app for photographers, we recruited pho-
tographers to test the website. You can recruit visitors
directly from the website using an invitation powered by
Ethnio or Hotjar.
It can be surprisingly effective to carry out what we
call retrospective moderated user tests, in which you contact
someone who has just completed your website’s goal
(e.g., made a purchase), and ask them if they’d be willing
to retrace their steps. Such people are, by definition, qual-
ified, and have recently gone through all of the thought
processes required to buy. Take notes as they show you
the path they took through your funnel, and ask them to
describe what they thought at each stage. They tend to be
excellent at recalling the hoops they had to jump through
and the obstacles at which they nearly fell.
diagnosis 137
Counterintuitively, regardless of which of the previous
methods you choose for recruiting, you tend to learn more
from users who aren’t web-savvy. People who use the
web a lot tend to be better at coping with pages that contain
errors. People who don’t use the web much are more easily
derailed—and can therefore provide more insights.
Then, give them the task you’d like them to complete. The
task is usually a typical goal of the website. For example, if
your website sells quirky gifts, the task might be “Add to
your basket some smoking mittens, some metal-detecting
sandals, and a photo-realistic bacon scarf ” (those are all
real products, amazingly). You may choose to set the goal
in general terms or specifically, depending on whether you
want to exclude certain possibilities:
1 38 m a k i n g w e b s i t e s w i n
• To consider all possibilities, you may state the goal
implicitly: “Imagine that your car has broken down.
What would you do next?” (About ten years ago, we
carried out a user test in which the participant, to our
surprise, asked for a Yellow Pages, and didn’t even con-
sider using a computer.)
• To be more specific, you may choose to state the goal in
general terms: “How would you get your car repaired?”
• You may choose to state the goal specifically: “Assume
you have typed [search terms] into Google and clicked
on the first search result. You arrive on this page. How
would you find the location of your nearest repair
center?”
diagnosis 13 9
Steve Krug, the author of a great book called Don’t Make
Me Think, has a useful script (a Word document) for carrying
out user tests. The script was designed for user tests in labs,
so you may want to ignore the references to microphones
and screen recordings.
The hardest thing about user testing is the emotional
aspect. Not only can they feel awkward, but also they can
make you want to cry. Because the truth hurts. For that
reason, most marketers shy away from them. The best mar-
keters are those who rapidly accept the criticism, use it to
improve the page, and then user test again.
1 40 m a k i n g w e b s i t e s w i n
the participants (you might choose to send your existing
customers to it).
• UsersThink and userinput are low-cost alternatives that
provide written reports instead of videos.
• UsabilityHub allows you to carry out tests for free, pro-
vided that, in exchange, you complete other people’s
tests.
• Watchsend is for iPhone apps.
• Testapic provides a user testing service in French; uxline
does it in Spanish; Testaisso does it in Portuguese.
diagnosis 14 1
assurance process before any page goes live. Smashing
Magazine’s list of “45 web design checklists and question-
naires” can be useful.
Session-recording software can help a lot with finding
such problems. Options include Clicktale, Inspectlet, Hotjar,
and Tealeaf—all of which also help to identify problems
with online forms.
Users’ gaze is influenced by surrounding cues. For example, they tend to look at
whatever the person on the page is looking at. (Image credit: Objective and Tobii.)
diagnosis 143
allows you to get professional eye-tracking reports on
demand. The LookTracker service works much like the
user testing tools mentioned above. EyesDecide provides
a similar service, but the eye-tracking reports are gathered
via the users’ webcams.
Eye tracking can reveal what your visitors see and what they don’t. The headline of
the right-hand page is more compelling, causing visitors to read on. (Image credit:
EyesDecide.)
1 44 m a k i n g w e b s i t e s w i n
EyeQuant and Feng-GUI use a computer algorithm to pre-
dict your visitors’ eye tracking, based on the contrast and
layout of the page elements.
diagnosis 14 5
A overlay inviting users to participate in a user test.
1 46 m a k i n g w e b s i t e s w i n
appear in a table in Ethnio’s or Hotjar’s interface. You can
then decide whether you’d like to carry out a user test
with them.
Ethnio’s control panel allows you to manage your user tests. (Image credit: Ethnio.)
diagnosis 147
Measure your wins…using A/B testing (to
test different versions of your webpages to
see which is the best)
A/B testing software allows you to create different varia-
tions of a page and then measure which converts best. You
can then promote the winning page to become your official
new version.
A/B testing doesn’t generate many insights. However,
it does give the final word on whether your insights and
intuitions were accurate. A winning A/B test is your visitors
voting with their feet (and often their credit cards).
For further information on A/B testing, see a report we
wrote titled “A/B testing 101.” Also, we recommend you
read the help files for whatever A/B testing software you
use. The help files tend to be written well, and their advice
will be correct for that software and whichever methods
and algorithms it uses.
1 48 m a k i n g w e b s i t e s w i n
Our comprehensive summary of A/B testing tools:
www.conversion-rate-experts.com/split-testing-software/
diagnosis 14 9
• Your current conversion rate. The fewer conversions
you get, the longer it takes to detect a doubling.
• The increase in conversion rate that you’re trying
to detect. A 100% increase can be detected about four
times as fast as a 50% increase.
• How statistically confident you want to be. If you
wanted to be 99.99% sure that your new page wasn’t
winning just by chance, you’d have to wait a long time.
1 50 m a k i n g w e b s i t e s w i n
in number and can be measured instantly, but you can’t
be confident that they correlate with overall long-term
success. The less traffic your website gets, the more
you need to rely on “intermediate” metrics toward the
left-hand side of the spectrum.
• Test on major pages. Maybe this one is obvious, but test
only those pages that almost all of your customers see,
like your main landing page or your checkout funnel.
• Combine similar pages into one test. If you have ten
landing pages, and you want to test the call-to-action
button, then apply the same change to all of those pages
and include them in the same test. Some companies
have many more landing pages than they need, maybe
because they wanted to make each landing page bespoke
to a particular keyword. We often consolidate such pages
into one, and then optimize the heck out it.
• Reduce the statistical significance at which you’ll
declare a winner. It has become the norm to declare a
winning test at a statistical significance of 95%, but that’s
not to say you can’t use a different figure. It would be a
shame to conclude that “If I can’t have 95% confidence,
I won’t run an A/B test at all.” That’s like saying, “I’m a
perfectionist. If I can’t do it perfectly, I won’t do it at all.”
Most marketing decisions are made without any mea-
surement, so don’t rule out the possibility of ending a
test at, say, 90% confidence—or even 85% confidence—
especially if the alternative is just to launch the page and
hope for the best. Sure, your chances of declaring false
diagnosis 151
positives increase slightly, but the benefits of being able
to test more ideas usually hugely outweighs the slim risk
of promoting a page that was actually losing.
• Fixed-period testing. Many A/B testing tools now mon-
itor your tests on an ongoing basis and tell you when
there’s a winner. In addition, though, you may choose
to specify a maximum duration for each test, after which
you’ll make a decision regardless. If the control was
winning at that point, you may choose to promote it. If
the challenger was winning, and you’re confident that
it was based on a research-driven hypothesis, then you
may choose to promote it. If the challenger was based
on a risky hypothesis, you may choose not to take the
risk. Either way, this approach is more rigorous than
how most early stage companies make decisions.
• Temporarily increase the amount of traffic to the
page being tested, even if it means sacrificing some
profitability. If the new page wins, the traffic may turn
out to be more profitable than you had expected.
You can estimate how long an A/B test would take by using
a calculator like Optimizely’s.
Regardless of whether you carry out A/B tests, we highly
recommend you use user tests as a way of measuring the
performance of pages. In fact, if the A/B test duration cal-
culator shows that your all of your tests would take more
than six months, we recommend you use only user tests,
and return to A/B testing once your business has grown.
• They are quick to carry out. A user test can take less
than ten minutes.
• They allow you to gather qualitative insights. A
two-month-long A/B test may tell you which page
performed better, but a ten-minute-long user test tells
you why.
• They provide insights that are granular. An A/B test
will only reveal which page is better—that’s just one fact
about the whole page—but a user test will reveal which
parts of the page work.
You may wish to also carry out user tests on your compet-
diagnosis 153
itors’ websites, to get an independent view on how users
perceive them.
You can learn a lot from your competitors. But to beat
them, you’ll need to do things they don’t.
1 54 m a k i n g w e b s i t e s w i n
SECTION 3
Making
websites
win
The most common problems
that make web visitors
abandon—and proven, easy-
to-implement solutions
1 56 m a k i n g w e b s i t e s w i n
and will then describe some proven, effective ways of fixing
it. Of course, there are many way to overcome each problem.
For example, there are over a hundred ways of overcoming
lack of trust, and some of them are much more effective
than others. For each problem, we will describe strategies
that we’ve found to be robust—that tend to work reliably
in most situations.
1 58 m a k i n g w e b s i t e s w i n
• “When there are parentheses at the end of a sentence,
put the period after the closing parenthesis (like this).”
• “If your whole sentence is in parentheses, put the period
inside the closing parenthesis. (Like this.)”
1 60 m a k i n g w e b s i t e s w i n
fused. Teachers and bosses may like intelligent-sounding
text, but readers prefer text that’s easy to understand.
• Record yourself.
• Get the recording transcribed. We highly recommend
Speechpad for fast, accurate transcription.
• Edit your transcript and incorporate it into your website.
• A/B test the new page to confirm that your changes have
increased your profits.
The pipes feel like points at which your brain gets to “take
a breath.” When you read the faded words, you may get
the same panicky feeling that you get when you are diving
underwater and you are starting to run out of oxygen. By
the time you reach the final pipe, your short-term memory
is gasping for breath.
The following text contains another example of the
same phenomenon:
Pulp Fiction, all three movies of the original Star Wars trilogy,
Harry Potter and the Deathly Hallows: Part II, The Social Network,
Avatar, and Finding Nemo reportedly fail the test.|
1 64 m a k i n g w e b s i t e s w i n
(Incidentally, dependency length is the number of words
during which the reader needs to “hold their breath” before
they can reach a resolution point.)
If you aim to go easy on your readers’ memory buffers,
several priceless rules of thumb emerge:
1 66 m a k i n g w e b s i t e s w i n
this case, it’s in the word improvement. You should rewrite
the sentence as follows:
In other words…
1 68 m a k i n g w e b s i t e s w i n
Hemingway “makes your writing bold and clear”
Hemingway highlights long, complex sentences and
common errors. It’s free. You won’t agree with all of its
suggestions, but it provides a fresh pair of (robot) eyes. We
use it regularly.
The Hemingway app suggests how you could have made your sentences easier
to read.
Ten years ago, the best way to create wireframes was with
paper and Post-its. Some people still prefer to work that
way. However, there is now an abundance of good software
for doing it.
The following wireframing tools make it easy to
display your work-in-progress designs. Some of them
allow you to create working prototypes. Many of them allow
you to create pixel-perfect final designs. Some even output
production-ready code:
The tools that we use most often are Sketch, Bal-
samiq Mockups and UXPin. However, there are a crazy
number of good alternatives, including…taking a very deep
breath…Adobe Brackets, Antetype, AppCooker (for iOS
apps), Appery.io (outputs code for mobile and responsive
apps), Atomic.io, Axure (a complex, sophisticated wire-
frame tool suite), Balsamiq Mockups, Canva, CanvasFlip,
Craft, Creately, Demonstrate, draw.io, FileSquare, Fire-
works, FlairBuilder (for apps, mobile-friendly), Flinto and
Flinto Lite (mobile-friendly), Fluid (specializes in mobile),
Framer JS (allows you to prototype and code apps for desk-
top and mobile),…gasping for air…Fuse, Gliffy, Handcraft,
HotGloo, Indigo Studio by Infragistics, iPlotz, iRise, Justin-
mind (mobile-friendly), Keynote, Koncept App, Lucidchart,
Macaw (outputs HTML and CSS), Mockflow (outputs
HTML and CSS), Mockplus, MockingBird, moqups, Naview,
1 80 m a k i n g w e b s i t e s w i n
InVision app (mobile-friendly) also allows other people
to give written feedback on your work-in-progress designs.
You upload your designs, and then invite others to annotate
them with whatever type of feedback you desire. Notable
has similar functionality. Alternatives include Notable
Prototypes (a variation of Notable), Firefly and BugHerd.
Composite connects to Photoshop files, turning them into
clickable prototypes.
To gather feedback on your work-in-progress videos,
you can use Frame.io, a fantastic web-based platform.
Alternatives include Wipster, Symu, Vidhub, Remark, and
Kollaborate. Such services provide great benefits; it’s hard
to gather and record such feedback even when everyone’s
in the same room.
Optimal Workshop provides several tools (OptimalSort,
Treejack, and Chalkmark) to help you optimize your web-
site’s navigation and information architecture. The tools
are described in our article about card sorting. Alterna-
tives for card sorting include SimpleCardSort, UsabiliTEST,
and Xsort.
1 82 m a k i n g w e b s i t e s w i n
“Monopoly is the condition of every successful business.” —
Peter Thiel, venture capitalist and cofounder of PayPal
1 84 m a k i n g w e b s i t e s w i n
Begin by using an on-page survey to ask your visitors why
they visited your website:
1 86 m a k i n g w e b s i t e s w i n
You may find that you satisfy only a small percentage of your visitors’ needs.
1 88 m a k i n g w e b s i t e s w i n
there’s a large correlation between the companies you like
and the ones you give money to.)
m a k i n g w e b s i t e s w i n 189
GoDaddy)—or even the cross-sold companies (like the
office software)—should be cross-selling your products.
1 90 m a k i n g w e b s i t e s w i n
An example of creating additional product ranges
While studying Morphsuits’ analytics, we noticed that many
visitors were searching for a type of Halloween suit that
didn’t exist. Armed with the data, Morphsuits started man-
ufacturing the new type of suit, safe in the knowledge that
there was zero risk in creating it. The new suit became a top
seller. We weren’t surprised. We even knew approximately
how many it would sell.
The Morphsuits team is extremely dynamic and has
taken this principle far. Over the past few years, they have
rebranded into MorphCostumes and increased their number
of suits from 71 to over 300.
1 94 m a k i n g w e b s i t e s w i n
efits, you’ll be able to spot when to state a feature, when to
state a benefit, and when to state both.
and
and
1 96 m a k i n g w e b s i t e s w i n
3. Some companies forget to mention some valuable
benefits
The travel-phone company Mobal gave away a high-quality
travel adapter with every phone. But it forgot to mention
the travel adapter anywhere on its website. When we added
the travel adapter to the website, sales increased. So we
added it to the offline marketing campaigns too. This was
one of the many contributing factors that allowed us to
triple Mobal’s sales in one year.
We used this technique to great effect when we cre-
ated more than $1 million of additional sales for Moz. In
a case study on our website, we describe how we added
to Moz’s landing page many features that previously
went unmentioned.
To ensure that you aren’t making this mistake, list
all of the elements of value that your visitors get, and then
check that your website communicates them all clearly.
It can help for you to order the product yourself, so you
see the whole package with your own eyes. Also, ask your
customers why they bought, and then ensure that all of
their reasons are featured in your marketing materials with
appropriate prominence.
1 98 m a k i n g w e b s i t e s w i n
process is repeated for subsequent trips. The cartoon was
so effective on the Japanese website, we transferred it to
Mobal’s USA website, to similar success.
Such flowcharts are an example of what the hyp-
notism world calls future pacing: The buyer is told what
they are going to experience over the next few days, weeks,
or months. (This is closely related to the sales technique of
making the prospect “think beyond the sale.”) Even though
future-pacing diagrams are less common in the West, they
are extremely effective here, because
This process will ensure that your final video will con-
vert visitors.
The previous techniques provide four opportunities for
optimizing your benefits and pros. In the next chapter, we
will describe how to optimize the other side of your value
propositions: the costs and cons.
20 0 m a k i n g w e b s i t e s w i n
to consider to grow your profits by optimizing your pricing
and offers.
As you read the following points, we recommend you
make notes of how each of them might apply to your business.
20 2 m a k i n g w e b s i t e s w i n
pay it, and then later it can be released in paperback to
mop up mass-market sales. This approach tends to be
more effective with products (e.g., consumer electronics)
than as a strategy with platforms (e.g., online stores). It
also allows statements like “Used to be $XXX” to be
used in the future advertising.
• Communicate that prices will keep increasing:
This approach provides a kind of scarcity. It encour-
ages buyers to act now, because they know prices will
increase. It is useful when the major challenge is getting
visitors to act promptly.
m a k i n g w e b s i t e s w i n 203
able to think creatively about how you could offer it. This
approach led us to create the highly successful $49 world
phone, with no monthly fees, which became the standard
in the world phone market.
20 4 m a k i n g w e b s i t e s w i n
pay a premium for. A good mindset is to imagine that you
will charge a premium price, and then use your marketing
knowhow to justify that price.
Note that many companies take external investment
to fund their costs of client acquisition in the short term.
The investment gives them all the previously mentioned
benefits of high prices and low prices, and is based on the
assumption that the strategic benefits of growing quickly
will be worth it in the end.
Offer strategies
Once you have decided on your pricing strategy, you are
ready to start packaging up your offers. Here are some of
the many options available to you:
m a k i n g w e b s i t e s w i n 205
beans, and charge more for products that are less
easy to compare, like artisanal balsamic vinegar.
° Make your money on the things people don’t
consider when making a decision: Restaurants
often have low-priced meals, then make all their
money on drinks.
° Consider stripping down the features of your
service, then charging for extras. Car dealers often
have a low headline rate, then charge for extras.
• It can help to make the initial purchase free. If you
can’t make it free, make it seem cheaper:
° Offer a “free trial,” which may be a no-strings,
completely free sample, perhaps with a discount
voucher if they decide to continue. This works if your
service is fantastic, and the best way to persuade
buyers is to get them using it.
° Offer a free trial with an ongoing monthly charge
if the customer continues.
° If it’s not possible to offer a sample, consider a simu-
lation of a sample. Freebird is a service that allows
travelers to get an alternative flight if theirs is can-
celed. As part of Freebird’s conversion funnel, it
allows users to experience the whole process as a
simulation, to demonstrate how easy it is.
° Consider offering something small and irresistible
for an amazing price, just to get something into the
buyer’s shopping cart. Once the buyer has accepted
20 6 m a k i n g w e b s i t e s w i n
that they’ll be going through the checkout, they are
more likely to buy other things.
° Offer an initial discount (e.g., only $9.99/month
for the first three months; $19.99/month afterward).
° Multibuy deals: Buy one, get one free is effective—
more effective than “half-price,” because it keeps
a high price on the product (so doesn’t lower its
perceived value) and encourages the customer to
buy twice as much. The same goes for other types
of multibuy deals.
° State that “We won’t bill you until N days after
your purchase.”
° Allow the buyer to pay in installments (e.g., three
monthly payments of $9.99). This works well if your
research reveals that many buyers don’t currently
have the money to pay for the product outright.
• Allow the buyer to “return it within X days for your
money back.”
20 8 m a k i n g w e b s i t e s w i n
how to get a good night’s sleep. You can deliver the
information in a wide range of formats, including audio,
video, electronic or printed reports, software, tools, or
access to websites.
• If some of your offering involves information, consider
how the information can be made to educate the buyers
about how to buy your type of product, and how to appre-
ciate the ways in which your product is superior. This
works well in B2B sales, because B2B buying is often
complex. For example, a seller of web hosting can bene-
fit from giving away a guide called “7 Mistakes to Avoid
When Choosing a Web Hosting Solution.”
• Consider offering particular premiums only to custom-
ers who spend more than a certain amount. One
of our favorite techniques is to offer a valuable free
report for customers who buy one unit, an additional
free report for customers who buy three units, and a
further free report for customers who buy five units.
A buyer who wants that third report can never get it if
they always order one unit at a time. This offer alone
can greatly increase the average order value.
• If a customer is paying with someone else’s money
(such as their employer’s), consider options that will
reward the person personally. Companies that sell to
businesses often use the following rewards such as
meals, events, air miles, cash back, and gift vouchers. (Of
course, don’t offer anything that’s unethical or illegal.)
• Offer premiums to encourage prospects to behave in
m a k i n g w e b s i t e s w i n 209
a certain way. For example, if you want all the orders to
be placed via the web, or via the phone, offer a premium
for people who do so.
• Offer discounted prices if people order quickly (e.g.,
early bird discounts) or in large quantities.
21 0 m a k i n g w e b s i t e s w i n
component that makes it worth it for the buyer to pay
the extra amount.
• Consider a price that’s ten times your current high-
est price, and then decide what people would pay. A
company that advises people on getting work permits
may charge $500 per client. The company might benefit
from exploring what a $5,000 service would entail. Who
would use it? Even if only a small fraction of buyers took
advantage of such an offer, the effect on the economics
of the business could be significant.
• Bestow status levels upon your customers: Have dif-
ferent levels of customers, and publicly reward those on
the higher levels by giving them a higher status. People
are naturally competitive and aspire to become better
customers. This technique is used by every producer of
luxury goods. Even credit card issuers manage to charge
more for premium gold and platinum cards.
• Offer a loyalty program, whereby the buyers get certain
rewards once they have purchased a certain number
of units from you. This encourages them to return for
more, thus forming a habit, and increasing your lifetime
customer value. As every cafe knows, it helps to give
customers free loyalty points to begin with, so their
reward account has value from the outset.
21 2 m a k i n g w e b s i t e s w i n
• Odd prices work slightly better: $9.99 is perceived as
being disproportionately lower than $10.00. This is
presumably for the same reason that people get excited
when their car odometer clicks over to 10,000 (even
though it’s just another mile), and sad when they reach
their fortieth birthday (even though they’re only a day
older than they were the day before).
• If a price has been reduced, put a slash through it to
avoid confusion, so it’s clear that it’s not the current
price. So rather than saying “Was $119, now only $89,”
say, “Was $119, now only $89.”
• If something is free, and the offer sounds too good
to be true, mention that’s it’s free several times, in
different ways. For example: “It’s free, so you pay noth-
ing, no strings attached, no hidden charges, absolutely
no cost to you whatsoever.”
• Many people don’t understand percentages—at least
not at a gut level—so it’s much better to say “half-price”
or “one-third off ” than “50% off ” or “33% off.”
• Use apples-to-oranges comparisons to remind people
what great value the product is. If you’re selling a train-
ing course, compare it to the cost of a college education,
and show people who have gone on to get jobs, to high-
light the net benefit. If you’re selling a shed, compare it
to the cost of having a house extension. If you’re selling
a video conferencing solution, compare it with the price
of international travel. You can make almost any prod-
uct sound like a great deal by comparing it to the price
21 4 m a k i n g w e b s i t e s w i n
how much)—and justify why that price was reasonable. If it’s
cheaper or better than competitors’ products, give details,
maybe with a comparison chart.
Payment options
Different customers tend to have different preferred meth-
ods of ordering and paying. It’s often best to offer the ones
they are most comfortable with:
• Payment method
° Credit card or debit card
° PayPal, Google Checkout, Amazon Pay or Apple Pay
° Direct debit or standing order
° Purchase order
° Credit card over the phone
° 0% finance
° Check
• Payment schedule
° Pay in multiple installments (which is good when
the initial price point is high).
° Till the person cancels: if your service is charged
on a monthly basis, you keep charging the customer
until they cancel it. You may choose to specify a
minimum term, before which the customer may
not cancel. (This is popular with telecoms provid-
ers.) The length of the minimum term (e.g., twelve
months, eighteen months, or twenty-four months)
can affect the take-up rate.
21 6 m a k i n g w e b s i t e s w i n
What kind of proof should you add? There are tens—
maybe hundreds—of ways to show trust and credibility.
Some of them are particularly suited to certain types of
businesses, but most businesses benefit from the following:
21 8 m a k i n g w e b s i t e s w i n
Hand soap manufacturers don’t say that their product
“Kills a lot of germs.” They say, “Kills 99.9% of germs.” It
sounds much more convincing.
And in its 1962 ad, a particular oil refining company
didn’t just say “We supply a lot of oil.” It said, “Each day,
we supply enough energy to melt 7 million tons of glacier!”
With hindsight, maybe that last one was misjudged.
Celebrity associations
Celebrity associations can be more effective than you might
expect. When selling a diet, for example, many people will
be more influenced by a celebrity figurehead than by sci-
entific research.
Over the years, we have paired several of our clients
with celebrity figureheads. Most of the celebrities allow
their image to be used in return for a fixed fee for a specific
duration. In one case, the celebrity was happy to do it for
free, because she was a fan of the company.
A word of warning. Before you commit to a deal, though,
you may choose to first run an A/B test to measure how
the celebrity affects conversions. One of our clients had a
celebrity figurehead whose presence actually reduced sales.
The client’s branding agency hadn’t carried out A/B tests
(we’ve yet to see one that does), and so was oblivious that
they had caused such damage.
A poster showing the space required to transport seventy-two people by either car,
bus, or bicycle. The photos demonstrate, better than words ever could, how single-
occupancy cars take up a disproportionate amount of road space. (Image credit:
Cycling Promotion.)
Social proof
In our case study about how we grew Crazy Egg’s conversion
rate by 363%, our winning landing page featured some of
Crazy Egg’s prestigious clients. Similarly, when we made
over $1 million for Moz, one of our winning pages had a
headline that incorporated the names of some of Moz’s
prestigious clients. Buyers are reassured when they see
that others similar to themselves have already chosen—and
like—a particular company.
220 m a k i n g w e b s i t e s w i n
It can help to mention your most prestigious customers. Readers think, “If it’s good
enough for those companies, it’s good enough for me.”
222 m a k i n g w e b s i t e s w i n
TopCashback’s homepage exhibits many types of proof.
224 m a k i n g w e b s i t e s w i n
° your rapid growth or huge size…then set that as
a goal;
° awards…then aim to win awards.
226 m a k i n g w e b s i t e s w i n
Function 2: A good guarantee self-evidently promises
that your business will be harmed if you don’t honor
your claims. It effectively says, “Our promise must be true.
Otherwise we wouldn’t be in business.” It thus acts as a
kind of proof.
Many people underestimate the importance of Func-
tion 2.
So guarantees are effective when…
228 m a k i n g w e b s i t e s w i n
subject of returns, use a phrase like “In the unlikely
event that…”
• Explain why you can afford to make the guarantee—
because you aren’t really taking a risk; most customers
are delighted.
• Sometimes it helps to put the guarantee in the voice of
the company’s chief spokesperson.
• Consider giving the guarantee a long claim period.
A guarantee with a deadline is effectively a time-lim-
ited offer. The buyer has a certain number of days
to make a decision. If your guarantee’s claim period
is short—fourteen days, for example—the principle
of urgency will be strongly working against you. On
the other hand, increasing your claim period from
twelve months to twenty-four months doubles your
financial risk with possibly no incremental increase in
conversion rate. The optimal period depends on many
factors, including your costs of sales, the quality of
the product, the time a customer perceives they will
need to evaluate the product, the typical lifetime of
the product, and the impracticality of returning the
product. To identify the optimal duration, we find it
useful to use financial scenario modeling supported
by A/B testing data.
• At the point of sale, it should be evident that your guar-
antee is easy to invoke.
• Emphasize that no strings are attached. In particular,
customers often worry that the guarantee will be nul-
m a k i n g w e b s i t e s w i n 229
lified if they use the product—or even take it out of the
box—so emphasize that they are encouraged to use it.
• Promise you’ll give a prompt refund. Customers worry
that it may take a long time to get the money back.
• The guarantee should be featured at all the points
where the buyer is thinking of the risk. That’s usually
near the call to action. Sometimes, it’s up front, too.
You may also choose to mention it when they abandon.
230 m a k i n g w e b s i t e s w i n
be handled operationally (accepting returns, restocking
shelves, issuing refunds, etc.). If you are satisfied that the
guarantee will generate additional profits, then proceed
to the next step.
3. Run the guarantee as an A/B test for a short time—for
just a few days, if that’s all the risk you can bear.
4. Wait for the guarantee period to expire.
5. Calculate the increase in profits, based on the mea-
sured uplift in sales.
6. Calculate the cost of people invoking the guaran-
tee. In our experience, the invocation rate tends to be
lower than companies expect, sometimes by an order
of magnitude.
7. If you need more data (which you probably will),
return to Step 3 and run the guarantee for longer. By
doing this in small increments, you remove the risk
caused by the long feedback loop.
8. If you do have enough data, and the increase in profit
more than offsets the cost of returns, make the guaran-
tee permanent.
9. Return to Step 1, creating a bolder guarantee.
232 m a k i n g w e b s i t e s w i n
return rate turned out to be much lower than our safeguards
had allowed for—so the guarantee was a huge success.
Before long, the conversion rate was so high, Mobal was
able to invest profitably up to a quarter of a million dollars
per month in offline advertising, and the sales tripled within
twelve months to $9.1 million.
234 m a k i n g w e b s i t e s w i n
Your conversion funnel should be like a Hazelnut Trail. Your
visitors must be irresistibly attracted by each step of your
sales funnel. They must believe that every step will make
their lives at least slightly better.
The CRE Funnel Planner: Much more powerful than you’d guess from those three
simple columns.
236 m a k i n g w e b s i t e s w i n
A much-improved Hazelnut Trail for the same company.
238 m a k i n g w e b s i t e s w i n
• Sometimes people become ready because of something
that has happened. For example,
° A cell-phone contract expires, so a person becomes
free to choose another carrier.
° On deciding to move house, a person starts to look
for a moving company.
° After eating too much over the holidays, a person
decides to join a weight-loss program.
• Sometimes people become ready to buy only once they
have started a project. For example, people tend to plan
a vacation on a day of their choosing, not when they
see an ad for travel insurance. As soon as they decide
to start planning, they begin a flurry of buying flights,
hotels, car hire, etc.
m a k i n g w e b s i t e s w i n 239
ple, a company that sells travel insurance could ensure
that its marketing materials appear within the market-
ing funnels of other travel companies (such as hotels,
airlines, travel insurance, airport transport, etc.).
240 m a k i n g w e b s i t e s w i n
Winning websites…manage complexity:
can your visitors not see the wood for the
trees? Here’s an eight-step approach to
managing complex sales
When we ask a website’s visitors why they didn’t buy, we
hear them say the same things over and over again. Lack
of trust comes up a lot, as does lack of understanding and
inability to find a suitable product.
However, some objections aren’t widespread—they
are product-specific. For example, there might be a
hundred reasons why a buyer isn’t ready to buy customer-
relationship-management (CRM) software, and each of the
reasons is specific to CRM software. Each objection has to
be tackled individually. What can you do if your visitors
have tens—or even hundreds—of different objections?
This is where conversion gets hard. For a marketer who
was hoping just to add a few testimonials and a guarantee,
the problem can seem overwhelming. It can only be solved
by a combination of process, diligence, and skill. And the
following steps:
242 m a k i n g w e b s i t e s w i n
You can get an amazing amount of value from populating a simple table of your
objections and counterobjections (O/CO).
m a k i n g w e b s i t e s w i n 2 43
you are creating should be added to the company’s sales
training course.)
• Ask live chat support representatives how they counter
each objection.
• See whether the company has “canned responses”
for live chat. They can be a goldmine of tested
counterobjections.
• Come up with counterobjections yourself. With most
companies, this is the most fruitful method. The people
who are best at coming up with counterobjections tend
to be intelligent and diligent and have an obsessive
knowledge of direct-response marketing techniques,
persuasive copywriting, psychology, and usability—all
the core skills of CRO.
244 m a k i n g w e b s i t e s w i n
thing in one step, selling a three-year enterprise-wide
contract for the CRM, with a multimillion-dollar com-
mitment. This may require months of face-to-face
meetings.
• At the other extreme, you could sell an instant free
trial with no credit card required—and postpone the
rest of the selling to a later stage in your conversion
funnel. It may take you no more than a minute of con-
versation to sell a free trial face-to-face.
If you ask for less, you don’t need to use as many words. So
it pays to explore how you can redesign your conversion
funnel so that each step becomes less commitment.
Of all the possible low-commitment steps, lead-
generation (or lead-gen) pages deserve a special mention,
because they are particularly effective. A lead-gen page is
a page designed to collect the visitors’ contact details—often
their email addresses. The “submit your email address” field
on a lead-gen page represents a low perceived commitment
to the visitor, but a high value to you as a marketer—because
it allows you to lead-nurture the visitor with long copy ad
infinitum (or, at least, ad tedium).
Of course, if you lower the commitment at an early
stage, bear in mind that you may have simply deferred
the commitment—and the persuasion required—to a
subsequent stage. Sometimes a free trial of the product
or service does that persuasion well. Also, it allows you to
collect the visitor’s contact details. Sometimes a free trial
m a k i n g w e b s i t e s w i n 245
doesn’t work, though, especially if the product or software
unavoidably requires the user to commit time or resources.
(Commitments can include money, work, time, and risk.)
Beware of suboptimizing a particular page only to discover
that the problem pops up further down the funnel. A super-
market could get more customers to the checkout by getting
the security guards to drag them there—or by closing off the
rest of the store—but it wouldn’t increase profits.
The most successful online retailers do both: They
reduce the commitment required from the buyers by pric-
ing products aggressively low, and then using many words.
In many situations, it’s most effective to offer a dual
path: to provide direct, low-commitment calls to action,
plus long copy. Successful software companies often pro-
vide a low-commitment way forward (a free trial), plus
detailed information for those who would prefer it. The
calls to action may be scattered throughout the copy, as in
the following image:
246 m a k i n g w e b s i t e s w i n
Many visitors want to know the features of software before they try it out, so it pays
to provide two paths. You can push a low-commitment free trial but also provide
pages and pages of information.
Your call to action may sit at the side of your long page.
248 m a k i n g w e b s i t e s w i n
You can offer as many paths as there are next steps that
you’d like the buyer to consider:
Sometimes it’s optimal to provide multiple paths. This example has three: (1) An
order button for visitors who are ready to buy, (2) a button for visitors who want
to schedule a trial, and (3) pages and pages of information for visitors who want
more details.
m a k i n g w e b s i t e s w i n 249
Step 3: How to organize your persuasive copy
using “separation of concerns”
No matter how long your page is, your visitors should be able
to easily find the information they need. They shouldn’t have
to read the whole page from start to end. A webpage should
be long like a phone book, not like a Russian novel.
Your challenge is thus one of information architecture:
How can you organize all the information so that visitors
see what they need—and at the right level of detail—without
getting bored? This is conversion at its hardest. It’s one of
the biggest challenges for many websites. It’s also extremely
fruitful when you get it right. In the following pages, we
present some techniques and concepts that don’t appear
in copywriting books. They will help you a great deal.
Separation of concerns is a fancy name for when infor-
mation is organized and encapsulated into modules. The
concept is closely related to the phrase “a place for every-
thing and everything in its place.” Separation of concerns
is essential when you’re managing a lot of content. For
website design, the modules can be paragraphs…
250 m a k i n g w e b s i t e s w i n
You can encapsulate benefits into clearly labeled paragraphs.
The navigation bars of many websites clearly encapsulate the content into groups
of pages. (In one of our case studies, we describe how we used this technique to
increase paid memberships for Smart Insights by 75%.)
252 m a k i n g w e b s i t e s w i n
By modularizing, you allow your visitors to easily find the
information that they need, and to ignore the rest.
It’s hard to stress how important it is to organize infor-
mation into an architecture that’s easy to navigate. Once a
visitor is lost, it’s difficult to show them counterobjec-
tions. They’ll never find them.
Separation of concerns may seem obvious and straight-
forward, but once a company does it badly, a mess quickly
ensues. We are fans of a particular brand of sit-stand desks.
However, whenever we recommend them to people, we
struggle to find the version we own (there are over twenty
variations, each with subtle differences). The problem is
common to many e-commerce websites that give each prod-
uct variation its own page. The visitor has to play a game of
spot the difference between different product pages, many
of which are almost identical except for a few differences.
Visitors don’t want to play spot the difference. If they are
buying a laptop, they don’t want to browse product names
like “ABC123-1Tw, ABC123-2Tb, etc.” Instead, they prefer
to see headings that reflect their current mindset and that
narrow down the choices. They want to be given choices
like “Do you want it with a 1 TB or 2 TB hard drive?” and
“Do you want it in black or white?”
254 m a k i n g w e b s i t e s w i n
will be even more likely to deteriorate over time. Poor
separation of concerns tends to snowball. Some telecoms
companies’ websites are messy to the point of being almost
irrecoverably out of control.
To make it clear where one section ends and another starts, try using alternating
gray and white backgrounds. Imagine how much more confusing the page would be
without the backgrounds.
256 m a k i n g w e b s i t e s w i n
It helps to clearly encapsulate content into separate page sections—and it doesn’t
take up any additional space or require any additional thinking.
258 m a k i n g w e b s i t e s w i n
navigation bars, for example, the headings can be under-
stood only by someone who’s already familiar with the
website. For example, companies often use the names of
their products in the navigation, even when the names
aren’t self-explanatory.
You might find it useful to think of it as “surprise nav-
igation.” With surprise navigation, you don’t know what
you’re going to get until you’ve got it. Surprise navigation
is as pointless as a road sign that can’t be understood until
you’ve arrived at the destination. It’s generous even to call
it navigation; it’s more like an in-joke.
Surprise navigation isn’t limited to navigation bars; it
also appears anywhere that has road sign functionality—
including headlines, subheads, the titles of page sections
and other page elements.
Surprise navigation is a common problem; it leads vis-
itors into oblivion, and it kills conversions.
You can eliminate yours by reading all of your head-
ings, including the tabs in your navigation, the headings
of pages, and the headings of sections—and then ensuring
that each heading would be understandable to a newcomer.
Then, to confirm that you were right, carry out user tests
and Treejack tests.
m a k i n g w e b s i t e s w i n 2 59
2. How to fix your headings: Many marketers
mistakenly label modules with “categorizers” when
they should have used “spoilers” or “teasers”
It’s often not enough for a heading to describe what’s in the
module. It should also tease or spoil.
So if your page has a section of media testimonials, intro-
duce it with a spoiler headline. That way, you communicate
the message even to skim-readers:
260 m a k i n g w e b s i t e s w i n
using categorization headlines when they should have used
spoiler or teaser headlines:
Carousels with right and left arrows indicate that more products can be viewed.
262 m a k i n g w e b s i t e s w i n
It’s possible to use horizontal and vertical carousels in combination, to show two
different dimensions of data. For example, in an e-commerce store, you could show
thumbnails of color variations on the horizontal axis and photos of different views
of the product on the vertical axis.
When the users hover over an image, an alternative view can be shown.
m a k i n g w e b s i t e s w i n 2 63
Information can be hidden behind a “Read more” link (which, in the context of
a bookseller, might sound like a subtly placed marketing slogan). When clicked,
the “Read more” turns into “Read less” (which would be a terrible slogan for
a bookseller).
The information in a Help Center can be intuitively categorized into subjects. When
the user clicks on a triangle, the answer is revealed. You can keep the answers short
by linking to separate pages for users who want even more detail.
264 m a k i n g w e b s i t e s w i n
A tooltip can contain any kind of information.
m a k i n g w e b s i t e s w i n 265
There’s no limit to what can be put in a tooltip. A travel website may choose to
display the times of return trips when the user hovers over the outbound journey.
266 m a k i n g w e b s i t e s w i n
By keeping the information in an overlay, sunshine.co.uk was able to counter one of
its biggest objections without distracting visitors away from its conversion funnel.
Each overlay contains enough information to warrant having its own page. But
it’s better for it to be in an overlay, so the user doesn’t lose where they are in the
conversion funnel. (Image credit: SimpliSafe.)
268 m a k i n g w e b s i t e s w i n
intuitive to think of progressive disclosure in the positive
sense—as adding information that otherwise may not
warrant space on the page.
There are so many ways to progressively disclose infor-
mation, how do you know where to start? Should you use a
tooltip or an exit overlay? In the following pages, we describe
the four types of “building blocks” with which you can build
progressive-disclosure elements.
m a k i n g w e b s i t e s w i n 269
• Arrows, triangles, and chevrons indicate that infor-
mation will appear as an expanding element. (The
following image shows multiple types—not that you’d
normally include all of them on one page.)
27 0 m a k i n g w e b s i t e s w i n
• Icons with +, ?, or i on them use little space and hint at
the nature of the content that will appear.
276 m a k i n g w e b s i t e s w i n
• When the user appears to be exiting the page, by
moving their pointer above the viewport of the browser.
278 m a k i n g w e b s i t e s w i n
Building blocks of progressive disclosure: 3. Formats
in which to present the information
The hidden information can appear in several formats:
280 m a k i n g w e b s i t e s w i n
• By a page section expanding within the page. These
are particularly useful when the information doesn’t
need to be hidden again. And when expanding them
won’t mess up the layout of the page.
282 m a k i n g w e b s i t e s w i n
• By users clicking on a “Close” icon. This is often the
best option; users tend to understand it well.
m a k i n g w e b s i t e s w i n 2 83
• By users clicking away from the page element. This
is convenient for users who are “in the know,” but it
isn’t discoverable, so it should be used in addition to
showing a “Close” icon.
284 m a k i n g w e b s i t e s w i n
• Spontaneously after a certain time period. This is often
used with information that appeared spontaneously too.
m a k i n g w e b s i t e s w i n 2 85
• A search box: Search can help users to find what they
need, but only at the expense of “teleporting” them into
a different part of your website. Once they arrive on the
search results page, they often lose their bearings. Once
in a while, look through the queries in your search logs to
identify which information the visitors had failed to find
while browsing. Often, you’ll find that the information
wasn’t where users would have expected it to be.
• A knowledge base: Knowledge bases can help users find
answers to their questions. But, as with search boxes,
visitors who search knowledge bases often lose where
they were in the conversion funnel.
• Live chat: A live chat operator should be able to find
information that the visitor can’t. Whether live chat is
economically viable depends on the economics of your
business (and not on the whims of the customer-sup-
port team).
• A prominently placed phone number: Not all visitors
want to pick up the phone. But for those who do, phone
calls tend to convert extremely well. The main drawback
of phone calls is that they don’t scale easily; call centers
operatives need hiring, training, paying(!), and looking
after. As such, the ideal combination is usually for the
website to do as much of the work as possible.
286 m a k i n g w e b s i t e s w i n
Winning websites…sail past the
competition: if your visitors are choosing
your competitors, here’s how to win
If you don’t have a strategy for winning despite competitors,
you are doomed.
This chapter describes many powerful concepts and
techniques we have used to help our clients dominate in
some of the world’s most competitive markets.
Some of the things you’ll get in this chapter:
288 m a k i n g w e b s i t e s w i n
Niching often wins
When you niche your service to a particular group of needs
or customers, you get several powerful advantages. To illus-
trate the point, consider shampoo. If you study the shelves
of a supermarket, you’ll see shampoo for the following:
m a k i n g w e b s i t e s w i n 2 89
Four reasons why niching usually wins
290 m a k i n g w e b s i t e s w i n
tomer’s need. The niched product effectively says, “I
understand your situation, and I am a ready-made solu-
tion for it.” It’s hard to overlook a product that has
been designed to satisfy your exact need.
2. People are aware that niched things are often better.
A surgeon who specializes in a rare type of gastric opera-
tion will tend to be better at it than a surgeon who hasn’t.
Specialists tend to be more proficient than generalists,
because they have focused their resources on solving a
narrower problem. As a result, most people recognize
that specialization is an indicator of high perfor-
mance. So even if all shampoo bottles were to contain
the same ingredients, customers would still expect the
specialist ones to be better.
3. Niching allows you to become the best in the world.
No one checks into a hotel and asks the concierge,
“What’s the second-best Chinese restaurant nearby?”
Seth Godin argues that buyers always want the best in
the world. Though he admits that “the best” depends on
what the target segment wants (e.g., Chinese food), and
that “the world” may be restricted (e.g., to restaurants
within a certain distance of the hotel). Godin argues that,
to be successful, a person or company must define the
“world” it will be best in, and what it means by “best.”
And then, crucially, to not fall short.
4. Things that are niched get attention. All death-metal
bands have similar logos. Except for the band “Party
Cannon.” See if you can spot Party Cannon’s logo:
292 m a k i n g w e b s i t e s w i n
point, the winning strategy is to become less niched and
more moderate.
You see this happen during TV talent shows and political
campaigns. At the start of a political campaign, a candidate
benefits from being extreme. As the campaign progresses,
and competitors are eliminated, the wise candidate mellows,
sacrificing attention-getting antics in order to appeal to a
broader audience.
m a k i n g w e b s i t e s w i n 2 93
How to spot good niches: Three easy techniques
you can use
• Look back at the success you have had over the past
year. You may find that it has a theme. You may have
already started niching inadvertently. Amplify that
niching by communicating it to your visitors. In his
book Entrepreneurship and Innovation, Peter Drucker
describes how Macy’s department store initially down-
played the growing effect of appliance sales on its
profits. It considered the sales to be an “embarrassing
success.” Macy’s profits rose only once it embraced
appliance sales as a part of its image.
• As markets grow, they fragment. If you can pick the
next dimension along which your market will fragment,
you can get there first. Match.com used to be the only
dating website anyone had heard of. As the industry
grew, many niched dating sites became successful.
There are now successful dating sites aimed at
° long-term relationships (eHarmony);
° time-starved professionals (Lovestruck);
° people who don’t want to pay (PlentyOfFish);
° people of certain religions (ChristianMingle);
° country folk (Muddy Matches);
° and, ahem, Tinder.
• What is luxury today will be mainstream tomorrow.
Study your most advanced, sophisticated, wealthiest
users. The problems they have today—and the solutions
294 m a k i n g w e b s i t e s w i n
they use—will soon be mass market. Such users give you
a glimpse of the future. When plasma TVs first came
out, they were used only by exhibitors at trade shows.
As prices fell, plasma TVs began to be bought by public
venues like bars. It took years before they began to be
bought for domestic use. What are your sophisticated
users buying today? How can you help to make that
service mainstream?
m a k i n g w e b s i t e s w i n 2 95
Six ways to spot niching opportunities your
competitors will ignore
One of the most common mistakes is simply to identify
an opportunity, and then aim to capture it. Just because
something is an opportunity today doesn’t mean it’s
worth chasing.
Instead, look for opportunities that you can seize
without a struggle, because your competitors won’t
stop you.
List your opportunities, and then run them through
the following checklist, to identify which of them your
competitors are likely to avoid:
296 m a k i n g w e b s i t e s w i n
3. Some competitors will avoid an opportunity because it
would disrupt their existing business: When Google
first made its office software free, it was safe to assume
that Microsoft wouldn’t do the same. Such a move would
have undermined too much of Microsoft’s revenue. In
his book The Innovator’s Dilemma, Clayton Christensen
calls this principle dependency. Companies depend on
their existing customers, revenue, and investors and
are unlikely to upset them.
4. You ideally want to focus on an opportunity that will
inevitably grow—to “skate where the puck is going.”
Fortunately, many competitors will avoid such an oppor-
tunity because it isn’t big enough yet. This leaves a
gap that you can fill. Even when an opportunity will
imminently become large, many competitors will ignore
it until it happens.
5. Some competitors will avoid an opportunity because
they have a track record of failing at that activity.
Or not even trying. If a company has always failed at
something, or has always avoided doing it, you can
predict with reasonable confidence that it will continue
to do the same. Often, you’ll never find out the reasons.
Fortunately, you don’t need to. For example, when
Yahoo! acquires a tech company, you can be reason-
ably confident that that tech company will no longer
be a threat.
6. Some competitors will avoid an opportunity because
it doesn’t match their strengths. In this respect, it’s
298 m a k i n g w e b s i t e s w i n
started by being a bookstore and is heading toward being
an everything store.
m a k i n g w e b s i t e s w i n 2 99
Winning websites…keep attention: if your
visitors are forgetting about you, here’s
how to make them keep coming back
Your visitors’ browsing sessions get can sidetracked by real life. This family, for
example, could possibly get distracted by eating, parenting, or the inevitable
breakdown of their relationships.
300 m a k i n g w e b s i t e s w i n
form they were completing, and instead spend the next
twenty minutes mopping an excellently clingy shiraz
with apricot top notes out of their rug.
Dollar Shave Club became successful off the back of its entertaining explainer video.
302 m a k i n g w e b s i t e s w i n
You probably wouldn’t want to replicate it, but LingsCars shows how successful a
company can be just by standing out.
m a k i n g w e b s i t e s w i n 303
your visitors need to make. We described this in detail
in “Step 2: Should you use long or short copy?”
• Persuade your visitors to follow you on Facebook, Linke-
dIn, Twitter, YouTube, or whichever social network is
used most in your industry.
• Use ad retargeting to persuade your visitors to come
back. Retargeted ads are particularly effective. Visitors
who visited once tend to be extremely likely to visit
again.
• Collect your visitors’ contact details—their email
addresses, postal addresses, and/or phone num-
bers—and then create a follow-up flow that keeps their
attention and persuades them to proceed. In many
industries—such as education—this beats everything
else.
304 m a k i n g w e b s i t e s w i n
Many companies ask for contact details as soon as the visitors arrive. It may seem a
bit “forward” but it works.
The university-degrees niche is highly competitive. The winning call to action tends
to be to collect contact details and then follow up via email and phone.
m a k i n g w e b s i t e s w i n 305
By using them, your business becomes resilient to
hungry kids, taxis, wine stains, and whatever else life throws
at your visitors.
306 m a k i n g w e b s i t e s w i n
Notifications often appear over the page a few seconds after it loads, capturing the
visitors’ attention.
An urgency notification.
308 m a k i n g w e b s i t e s w i n
the same message. They manage to express urgency in
many subtly different ways.
m a k i n g w e b s i t e s w i n 309
have bought tickets to see 90 minutes of new romantic
power ballads.
Principles of urgency
Urgency is about time. Explore reasons why your visitors
should act promptly.
31 0 m a k i n g w e b s i t e s w i n
• If there are no inevitable deadlines, consider creating
deadlines. For example, offer a premium or discount
for people who respond within a certain period. (For
example, conferences usually have an early bird reg-
istration period.)
• Look for scarcity in your business. Scarcity causes
an inevitable deadline—the time at which the scarce
resource runs out.
31 2 m a k i n g w e b s i t e s w i n
How can you protect your visitors from the Handover of Death?
31 4 m a k i n g w e b s i t e s w i n
5. Be memorable
Be memorable in terms of your name and branding, so that
visitors think of you next time they have the same need.
31 6 m a k i n g w e b s i t e s w i n
To win at conversion, you must win at LCV. The most
successful companies can outbid their competitors in adver-
tising—not just because they outconvert their competitors,
but also because their customers keep returning to spend more.
That way, the cost of acquiring a customer can be recouped
over the total lifetime of the customer.
Furthermore, existing customers are easier to convert—
provided they had a good experience first time round.
So how can you get customers to spend more, and more
often? We have worked with many companies, and we have
noticed commonalities in the ones that have become most
successful. The following sections reveal our observations
on what works.
31 8 m a k i n g w e b s i t e s w i n
Use Net Promoter Score to improve customer
happiness
How can you measure whether you’re turning visitors into
raving fans? Net Promoter Score (NPS) can be simple and
effective. To measure your raving-fan-ness, simply ask
your customers the following question: “On a scale of 0 to
10, how likely are you to recommend us to a friend or col-
league?” You then calculate the percentage of respondents
who gave scores of 9 or 10 (NPS calls those people “Pro-
moters”), and then subtract the percentage of respondents
who gave scores of 0 to 6 (NPS calls them “Detractors”).
The theory is that Promoters will grow your business via
word of mouth and that Detractors will shrink your business
via word of mouth. An NPS of +100 would mean that every
customer would be a raving fan, evangelizing you wherever
they go. An NPS of –100 would mean the opposite, that
every customer is out there complaining. NPS isn’t just
useful as a metric; it’s useful as a concept. An NPS of +100
is a guiding star that every team member can envisage and
strive for.
Your raving fans are likely to make subsequent purchases,
and they are likely to recommend you to their friends.
320 m a k i n g w e b s i t e s w i n
In buying and constructing this shed, Ben made an observation that led to the
manufacturer redesigning its entire range.
322 m a k i n g w e b s i t e s w i n
SECTION 4
See the
power
of CRO
in action
A case study showing exactly
how we grew a company at
record-breaking speed
324 m a k i n g w e b s i t e s w i n
goHenry’s growth was faster than the winner of the Sunday Times Fast Track 100.
326 m a k i n g w e b s i t e s w i n
mobile visitors were telling us. It sounds obvious, but it’s
often overlooked—even by companies with sophisticated
mobile websites.
Don’t assume that your mobile visitors are just desk-
top visitors on a different device. If you do, you’ll focus on
the user interface as the only difference between the two.
In fact—as we discovered when we gathered data on
goHenry’s visitors—mobile users can have very different
intentions, likes, and objections to their desktop coun-
terparts. You’ll need to address these specifically on your
mobile journey.
328 m a k i n g w e b s i t e s w i n
At 2.5 m tall on mobile, our new page may have problems getting through doorways,
but it had no problem beating the original page by 78%.
s e e t h e p o w e r o f c r o i n a c t i o n 329
There are loads of reasons why our new page beat the orig-
inal. Here are just some of the proven techniques we used:
33 0 m a k i n g w e b s i t e s w i n
Just some of the ways we used to communicate the offer on mobile (all taken from
the same page).
33 2 m a k i n g w e b s i t e s w i n
amazingly effective. It’s worth thoroughly searching your
own business. Often, a fresh pair of eyes will spot something
you’ve overlooked.
33 4 m a k i n g w e b s i t e s w i n
During the A/B test, our new pricing page beat the control by 36%. You’ll see that it
includes more salesmanship than the original. Many companies make the mistake
of assuming that they don’t need to do any selling on their pricing page.
33 6 m a k i n g w e b s i t e s w i n
On the pricing page, don’t forget to remind prospects of the value of your product.
We added a section covering all the great free features bundled with goHenry and
an image showing how much “stuff” you get.
Recycle, recycle, recycle—if you have winning content, look for other places you can
use it.
33 8 m a k i n g w e b s i t e s w i n
Do you know where your customers come from? If you
don’t, ask them the following question: “Where did you
first hear about us?”
You might be surprised at the answer.
s e e t h e p o w e r o f c r o i n a c t i o n 339
Signpost your website so that visitors know they’re in the right place.
340 m a k i n g w e b s i t e s w i n
We created an 11% uplift in card activations from
optimizing the offline journey
Don’t just limit optimization to your website. If part of your
journey happens offline, optimize that too.
That’s exactly what we did with goHenry’s card acti-
vation journey.
Our research told us that half of the children knew about
the card before it arrived in the mail. Now, if you’ve got
kids you’ll know how persuasive they can be, so it came as
no surprise that parents were more likely to activate when
their child knew about the card.
It followed that if we could make more kids aware of
the card, more cards would be activated.
So we sent the welcome letter to the kids instead. After
all, the purpose of goHenry is to teach children finan-
cial responsibility.
The result? Card activations increased by 11%.
342 m a k i n g w e b s i t e s w i n
• If you can’t get friends or colleagues to help, use sites like
UserTesting.com to do remote usability testing on mobile.
(Tip: we love UserTesting.com’s webcam-based tests
for mobile, so look out for this option.)
• Use tools like SurveyMonkey to conduct quick straw
polls. Verify is a good alternative.
• Make sure you can quickly view your wireframes
on mobile. We use InVision to rapidly add function-
ality to our mobile wireframes. Then, we combine the
wireframes with screenshots to create a semiworking
version of the entire funnel. Finally, we use a service
like UserTesting.com to test the prototype funnel.
• We use A/B testing software like Optimizely to mea-
sure which page generates the most conversions.
Optimizely’s advanced custom audience features came
into their own on this project. Tests that would have been
prohibitively complex to run were suddenly within reach.
s e e t h e p o w e r o f c r o i n a c t i o n 3 43
What’s next?
We are now working on the goHenry app itself, aiming to
drive engagement and referrals.
On the website, we’re working on the sign-up funnel,
as well as other key parts of the journey such as forgotten
password and referrals.
Once complete, we will have rebuilt the entire site.
What’s unusual about this rebuild is that it has been
done iteratively, and with conversion at the heart of every
decision. Conversion, by definition, is the reason that any
website exists.
Every new page has been
344 m a k i n g w e b s i t e s w i n
the fact that we’ve worked extensively on goHenry’s desk-
top website. That’s because optimizing for mobile is just
like optimizing for desktop, but with the following addi-
tional obstacles:
s e e t h e p o w e r o f c r o i n a c t i o n 345
Now it’s your turn to win
This book contains enough information to hugely grow
any business.
More importantly, though, it provides a mental model
for improving businesses:
l e t ’ s k e e p i n t o u c h ! 349
Acknowledgements
We’d like to thank the following awesome people for helping
to create this book:
Dave Redfern, Martin Stone, Jonathan Rozek, Tucker
Max, Avinash Kaushik, Ian Claudius, Mark Chait, Adam
Costa, Art Crowley, Bradd Libby, Casey Bell, Darcie Con-
nell, Eoin Edwards, Kamil Ropiak, Karol Barzowski, Lotte
Larsen, Nicolas Fradet, Nina Bordet, Pawel Banhegyi, Peter
Hardingham, Richard Bitz, Vicky Cargill, Will Smith, and
Chelsea Batten.
acknowledgements 3 51