Microsoft Advertising Insights Exploring Digital ROI For Fast Moving Consumer Goods White Paper
Microsoft Advertising Insights Exploring Digital ROI For Fast Moving Consumer Goods White Paper
Microsoft Advertising Insights Exploring Digital ROI For Fast Moving Consumer Goods White Paper
1
Exploring digital ROI
For FMCG brands; the missing link
In the absence of a proven link between digital campaigns and offline sales, FMCG
marketers have been reluctant to shift marketing budgets to digital.
According to Nielsen Ad Dynamix, in Our study was a meta analysis of all the
the UK, online display currently receives FMCG models of sales built by BrandScience,
approximately 3% of FMCG brands media which included quantification of the sales
allocation, TV gets approximately 60%, Print effect of all of the media channels used to a
23%, Outdoor 6%, Cinema 3%, and Radio significant degree by FMCG brands.
3%*. The current economic climate has
made the need to focus on ROI ever more Where sample sizes permit, we have explored
pressing, with a surge of activity and energy the digital ROI by FMCG sub-category
going into price promotions (Impulse, Food and Drink, Health and Beauty
wherever possible. and Household) and by digital platform
(Standard online display media, Rich media,
Microsoft Advertising, in partnership Pre-Roll Video, Mobile and Social).
with econometric experts BrandScience,
looked carefully at how the case for digital
driving offline sales can best be made. Our
approach is to harness the results that come
from Econometric Modelling, a technique
used for over 25 years by FMCG businesses.
Econometric Modelling provides a credible,
cross-media view of marketing, reflecting
todays media landscape.
Drivers Performance
3
Applying Econometrics
to FMCG sales
Our study aggregates data from over 455 FMCG campaigns across Western Europe (UK,
France, Italy, Germany, Spain, Norway, Sweden and Denmark), representative of the sector
as a whole. BrandScience looked at the effect of all drivers, including media activity on
shortmedium term sales by brand, modelling the historic variation in actual sales figures
collected on a weekly basis over a two-year period.
Our study found that media activity contributes approximately 5% of all sales with the
remaining 95% resulting from fundamentals such as stock, distribution and pricing (in-
store/on pack promotions). This 5% slice of sales variationthe effect of media activity
on salesis the basis of our analysis. BrandSciences Econometric approach quantifies the
effects of different media, delivered at specific times in different combinations and weights,
on shortmedium term sales (i.e. up to 6 months). By comparing the mathematical formula
estimated to the actual sales figures, the modelling process is refined over time.
60% TV
56% TV
23%
Outdoor 12% Print
12%
The media split in our sample is broadly representative of actual spend across
the FMCG sector as a whole, as reported in Nielsen data.
*Sample Set: European FMCG Results Vault : 455 studies as at 18th May 2012
**Sample Set: FMCG advertisers who spend 1m+ on advertising. Online is display only April 2012
4
In order to include as many However, Microsoft Advertising, in partnership with
campaigns with a measurable BrandScience has developed an analysis tool to
isolate the effects by country, by digital platform
online component as possible more on that later.
(54 in total), the data used
throughout this paper is an The sheer amount of data means that the analysis is
extremely robust. On average, each model explains
aggregate of the Western
88% of all sales variations for that brand/model that
Europe findings. range varying from 81% to 96%.
Sales variation
Sales
Model
Sales per week
Time
5
Online delivers
The second greatest ROI for shortmedium term sales
Revenue
gained Cost of
through
marketing
marketing
activity
= RROI
activity
%
media 8%
mix
1
European FMCG Results Vault: 54 studies as of 18th May 2012
2
WARC average media spend figures
6
Online makes other
media work harder
When we compare the difference in RROI performance between studies that have an
online element and those that do not, the results are clearadding online to the media
mix has a positive impact on the campaign RROI, regardless of what media is usedso it
is acting as a good support media. Online not only delivers excellent RROI efficiency
itself, but it makes other media spend work harder.
2.64
No online Online
1.54
1.40 1.34 1.39
1.26 1.32
1.14
0.93
0.74
RROI (Euros)
7
Want an impression that lasts?
Spend more online
Continuing to explore onlines role within the But people will also forget, so the response
media mix, we assessed whether varying the decays at a constant rate into the future. A high
proportion of online spend within a campaign carryover rate implies people will continue to act
had a direct influence on the online ad on the advert a long time after seeing it, while
retention rate, and therefore the overall a low carryover rate implies a less successful
campaign performance. Retention is a measure campaign, which people forget more quickly and
of the effect of advertising over time. When subsequently act upon less.
a person is exposed to an advert, there is a
delayed response.
80
20
0
Time
The illustration above shows how we measure A more successful campaign will have a higher
the life of a campaign. The black area represents peak and a less steep curve, represented by
campaign activity, e.g. TVRs. The two grayed the red area in the chart. A less successful
areas then represent how ad retention is campaign will have a lower peak and a steeper
affected following the campaign. The peak of curve, represented by the gray area. A standard
the chart represents one week after the measure of medium-term effectiveness is by
campaign, the point of maximum impact. The looking at the ad retention rate in week two as
slope of the curve then shows the rate at shown above.
which that impact reduces.
8
As you can see from the chart below, increasing investment in
online increases the longevity of the online impact. Using the
econometric meta analysis, we can see exactly how ad retention
leads to uplifts in campaign RROI
72%
27%
20%
1st Quintile 2nd Quintile 3rd Quintile 4th Quintile 5th Quintile
Under 10k 10-40k 40-120k 120-190k 190-750k
Based on meta-analysis of FMCG econometrics studies by BrandScience: 54 cases, 18th May 2012
9
Online RROI
Varies by subcategory and by platform
When we analyse the online RROI at a sub-category level, we can see that Impulse FMCG purchases
have been driving the greatest RROI at 1.73 for every spent, this is 59% higher than the average
online RROI for Household campaigns (where online returns 1.09 for every spent).
%
media 6% 7% 6% 7%
mix
Based on meta-analysis of econometrics studies by BrandScience: 54 cases who are FMCG and use online as of 18th May 2012
Continuing to view the data at a sub-category level, this analysis enables us to explore the online
RROI compared to the other mediaOnlines RROI of 1.73 achieved for the Impulse category out-
performs all other media. It is 42% higher than its closest rival, radio, which delivers an RROI
of 1.22 for every spent.
%
media 6% 7% 66% 11% 4% 7%
mix
10 Based on meta-analysis of econometrics studies by BrandScience: 8 Impulse cases that use online, 18th May 2012
Digging deeper into the data set, we are able to explore the
performance by digital platform, by sub-category. The chart
below illustrates that mobile leads the way in terms of RROI for
the Impulse sector. We need to bear in mind that the display
media data is an aggregate of all online display assetsthe
rich media display activity performs 1.5 times better than
standard display so here would out-perform mobile if we
applied a further split on the display data.
Rich media
2.36
Based on meta-analysis of econometric studies by BrandScience: 8 impulse cases that use online, 18th May 2012
11
Initial findings
The econometric model meta analysis has provided three key
findings, which hold true across all of the FMCG sub-categories
we were able to analyse:
12
But this is only the beginning...
The data shown in this paper has been scenariosin terms of weights of media
constructed in partnership with BrandScience spendaffect sales. The beauty of this tool is
as the result of a meta-analysis of the that we are able to tailor the inputs based on
econometric modelling studies. However, we client-specific needs and explore the optimal
are constrained by the limitations of the charts media mix for that specific scenario based on
which at best only show how two variants actual campaign performance and sales data.
work together to impact sales. In order to We are at your disposal to assist in advising
really understand how all of these factors how you can optimize your own media
interplay, we have built a multivariate analysis schedules we look forward to
tool that takes all of the findings into account working with you.
and enables us to analyse how different
13
About Microsoft
Advertising
Global Insights
Contacts:
Natasha Hritzuk Anita Caras
Sr. Director of Global Insights Head of Global Agency
natashah@microsoft.com and Accounts Insights
Anita.caras@microsoft.com
While many tech and media companies conduct market research that describes what
consumers are doing, the Microsoft Advertising Consumer Insights team believes innovation
stems from getting at the why. As a result we go beyond behavior to focus on why consumers
do what they dowhether thats choosing one brand over another, or exhibiting a preference
for a specific platform. Our goal is to create more robust insights-driven narratives that reveal
the people behind the data, making it easier for customers to tell creative, relevant and
connected stories across platforms.
2013 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or
trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft
must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy
of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION
IN THIS PRESENTATION.