Research by Social Embassy - Facebook ROI for brands – a study that looks at the impact of brand engagement on Facebook in relation to brand commitment, brand preference, NPS and spend.
Regression analysis: Simple Linear Regression Multiple Linear Regression
Facebook ROI for brands - The value of fans, engagemend and conversation
1. FACEBOOK
ROI FOR
BRANDS
THE VALUE OF ENGAGEMENT,
FANS AND CONVERSION
FACEBOOK ROI FOR BRANDS
2. WHAT IS THE VALUE
OF A FACEBOOK
FAN PAGE
FOR A BRAND?
FACEBOOK ROI FOR BRANDS
3. Facebook marketing is popular amongst brands.
67% of top brands run a fan page (Source: Social
Media Monitor 4). This percentage has once again
risen in recent months.
However, in practice, attracting and activating
Facebook fans can require considerable investment.
Brands are therefore looking to discover what the
true value of a Facebook fan page is.
FACEBOOK ROI FOR BRANDS
4. FURTHER 10
DEVELOPMENT OF
SCIENTIFIC RESEARCH 8
Academic research* based on a fictitious brand
has already shown that interaction with a brand
on Facebook has a positive effect on the brand 6
commitment that the fan feels for the brand.
And as well as the brand commitment, the
intent to purchase and the community feeling 4
are also positively influenced.
BRAND COMMITMENT
As a follow-up to this initial study, additional
practical research has been carried out for 2
six brands.
0
DEGREE
OF
PASSIVES INTERACTION CONTRIBUTORS
CONSUMER
No interaction ‘Liking’ fan pages and Liking fan pages,
reading messages reading messages and
Positive reaction responding to messages
Negative reaction
* http://www.socialembassy.nl/bewijs-interactie-met-een-
merk-op-facebook-leidt-tot-brand-commitment/
FACEBOOK ROI FOR BRANDS
5. QUANTITATIVE PRACTICAL
RESEARCH: 6 BRANDS,
3,009 RESPONDENTS, 1,272 FANS,
1,737 NON-FANS
In a study involving six brands, the relationship between brand
interaction (engagement) on Facebook and brand commitment in
practice was investigated.
Via an external panel, an online survey was set up in which fans and
non-fans were asked a series of questions. The fans interviewed for
the study were found via the fan pages of the participating brands.
Total Fans Non-Fans
Total 3,009 1,272 1,737
3FM 772 474 298
Chocomel 436 173 263
Essent 318 53 265
KPN 433 136 297
Optimel 536 242 294
UPC 514 194 320
*The research justification can be found towards the end of this publication.
FACEBOOK ROI FOR BRANDS
6. RESEARCH
QUESTIONS
INTERACTION BRAND COMMITMENT
To what extent does interaction with a
brand on Facebook result in strengthening
of brand commitment?
FANS OF BRANDS...
To what extent is brand commitment
stronger amongst fans than non-fans? ...USE THE BRAND MORE OFTEN.
...SPEND MORE MONEY ON THE BRAND.
...HAVE A STRONGER PREFERENCE FOR THE BRAND.
To what degree does being a fan and ...RECOMMEND THE BRAND MORE OFTEN.
interaction result in conversion?
FACEBOOK ROI FOR BRANDS
7. INFLUENCE OF FANS ON
BRAND COMMITMENT
You could say that fans are more valuable for brands than
non-fans, simply because being a fan is by definition a causal
relationship between interaction and brand commitment.
Within the research structure, the influence of ‘being a fan’ has
been minimised. Incentives, for example, are used to get non-
active fans to participate and hence to realise an appropriate
representation of the fan base.
In this respect, a great deal of focus is given to changes in
behaviour since the respondents became fans of the brand.
FACEBOOK ROI FOR BRANDS
9. This backs up the conclusion that building up a fan com-
THE VALUE OF munity on Facebook does indeed add value. However,
an even more interesting aspect of this study is the
A FACEBOOK FAN PAGE relationship between brand interaction and brand com-
mitment. Fans who interact more with a particular brand
have a higher level of brand commitment. Furthermore,
FOR BRANDS fans themselves say that they have become more com-
mitted to the brand since becoming a fan. This shows
that brand interaction on Facebook can contribute to the
strengthening of brand commitment.
This study, which involved 3,009 respondents, showed
the impact of brand interaction/engagement on brand Furthermore, the research shows that a proportion of
commitment. the fans make more usage of the brand after becoming
fans. This is also explained by the interaction that the
The positive relationship between brand commitment brand initiates with fans. After all, engagement boosts
and trading results such as usage, spend, preference brand commitment, and brand commitment has been
and NPS has been common knowledge for some time. shown to increase conversion. Furthermore, it has been
The results of this study show that brand interaction on shown that Facebook updates increase usage/purchas-
Facebook can lead to an increase in brand commitment ing amongst a proportion of fans. In this way, interaction
in practice, which confirms the correlation between via Facebook has a very concrete effect on the brand’s
engagement and trading results. bottom line.
The results show that fans have a significantly higher This gives greater relevance to engagement/brand
level of brand commitment than non-fans. Fans interaction via Facebook, due to the positive correlation
spend more on the brand, use the brand more often, between brand commitment and results.
have a stronger brand preference and a higher NPS.
FACEBOOK ROI FOR BRANDS
12. PERCENTAGE OF FANS FOR WHICH BRAND
ENGAGEMENT COMMITMENT HAS INCREASED
REINFORCES BRAND
COMMITMENT 18
%
We asked fans to what extent they think their brand commit- 30%
ment has increased since they became fans of the brand on 34
%
Facebook.
3FM (N=474) CHOCOMEL (N=173) ESSENT (N=53)
The research shows that 30% of the fan base thinks that their
brand commitment has increased since becoming fans of the
brand.
For KPN, no less than 56% of fans believe that their brand com- 20%
mitment to KPN has grown since becoming fans of the com-
pany on Facebook. 28%
The research also shows that there is a correlation between
the degree of interaction and the degree of brand commitment. 56%
Brands can therefore significantly boost brand commitment by
KPN (N=136) OPTIMEL (N=242) UPC (N=194)
interacting with fans.
FACEBOOK ROI FOR BRANDS
13. PERCENTAGE OF BRAND-COMMITTED FANS COMPARED
INCREASED WITH THE DEGREE OF INTERACTION
INTERACTION RESULTS NO INTERACTION OCCASIONAL INTERACTION FREQUENT INTERACTION
100%
IN STRONGER
BRAND COMMITMENT 77%
The higher the level of interaction with a brand, the
higher the percentage of fans that claim to be com-
mitted to that particular brand.
43
%
When there is no interaction, 43% of fans say that they
are brand committed, while 100% of fans who say that
they often interact with the brand via Facebook say that PERCENTAGE OF FANS FOR WHOM BRAND COMMITMENT
they are brand committed. HAS INCREASED BASED ON THE DEGREE OF INTERACTION
Furthermore, the research shows that the more that NO INTERACTION OCCASIONAL INTERACTION FREQUENT INTERACTION
fans interact, the stronger the fan’s rise in brand com-
mitment is since he/she became a fan. The increase for
fans who interact a great deal was 77% higher than
for fans who do not interact.
19%
The research therefore shows that there is a correlation 24% 74%
between the degree of interaction and the strength of
brand commitment.
FACEBOOK ROI FOR BRANDS
15. 100
HIGHER BRAND COMMITMENT
80 MEANS MORE FREQUENT
60
USAGE, GREATER SPEND, HIGHER
PREFERENCE AND HIGHER NPS.
40
Brand-committed respondents score higher than non-
brand-committed respondents for all business variables.
20
Fans spend more on the brand, use the brand more often,
have a stronger brand preference and a higher NPS.
90% 66% 30% 18% 93% 48% 79% 26%
0 Achieving brand commitment amongst fans is therefore
Frequency of use / Spend / Preference / NPS / a relevant objective. For this purpose, the encouragement
Weekly Above Will opt for Promoter of engagement amongst fans has been shown to be an
average the brand
again
important tool.
brand committed
not brand committed
FACEBOOK ROI FOR BRANDS
16. MORE PURCHASES SINCE BECOMING A FAN
2%
5%
4%
4%
A PROPORTION OF
5%
16%
94%
%
91
THE FANS SAYS THAT much more
frequently 79%
THEY HAVE BOUGHT/ more frequently
the same
LISTENED MORE SINCE
BECOMING FANS. 3FM Chocomel Optimel
(n=772) (n=436) (n=536)
During this study, we have also researched
changes in behaviour since becoming fans.
1%
19%
10%
2%
11%
4%
Many fans say that they have bought/listened
%
more since becoming fans. 9% of Optimel fans
%
14
14
say that they have used the brand more often.
For Chocomel, the figure is 21% and for 3FM, it 77%
is 6% (3FM already had a very high usage level
75%
amongst fans). 73%
UPC, KPN and Essent fans also indicated that
they had used the services more often.
25% of UPC fans said that they had already
bought a new product/service or package exten- Essent KPN UPC
sion. The same was true for 24% of KPN fans. (n=318) (n=433) (n=514)
Yes, I have purchased new products/services.
Another noticeable point was that of all the
FMCG brands nobody reported purchasing/ Yes, I have broadened my package.
listening less. No, I have not purchased any new products/services.
Don’t know.
FACEBOOK ROI FOR BRANDS
17. The study shows that Facebook posts can result in conversion.
FANS BUY/ For Optimel, 45% of fans have bought at least one product as a
result of a Facebook post. For Chocomel, this figure was as high
LISTEN MORE IN as 53%, and 42% of 3FM fans report having tuned in as a result
of a message on Facebook.
RESPONSE TO Facebook posts were also shown to create conversion for UPC
Essent and KPN. For UPC, 12% of fans have definitely bought
new products/services as a result of a Facebook message. This
FACEBOOK UPDATES figure was 20% for Essent and 12% for KPN.
100 100
80 80
60 60
40 40
20 20
Yes, always Yes, definitely
Yes, frequently Yes, maybe
1% 8% 33% 4% 11% 38% 3% 8% 34% Yes, sometimes 20% 10% 12% 33% 12% 10%
0 0
3FM Chocomel Optimel Essent KPN UPC
(n=772) (n=436) (n=536) (n=318) (n=433) (n=514)
PURCHASED/LISTENED AS A RESULT OF FACEBOOK UPDATES PURCHASED AS A RESULT OF FACEBOOK UPDATES
FACEBOOK ROI FOR BRANDS
19. FANS SCORE HIGHER THAN NON-FANS
WITH REGARD TO BRAND COMMITMENT,
PREFERENCE, USAGE, SPEND AND NPS.
FACEBOOK ROI FOR BRANDS
20. 100
FANS HAVE
STRONGER BRAND 80
32%
COMMITMENT 60
THAN NON-FANS 40
80%
For all brands included in the study, brand 20
commitment amongst fans was significantly
higher than amongst non-fans.
68% 20%
0
Of all of the fans, 68% say that they are com-
mitted to the brand for which they took part FANS NON-FANS
in the study. Amongst non-fans, this figure
was 20%. PERCENTAGE WHO SAY THAT THEY ARE BRAND
COMMITTED N.B. Significance level: p < .001
brand committed
not brand committed
FACEBOOK ROI FOR BRANDS
21. CHOCOMEL AND 3FM (N=772)
81%
21%
3FM HAVE THE CHOCOMEL (N=436)
79%
22%
STRONGEST BRAND ESSENT (N=318)
49%
18%
COMMITMENT KPN (N=433)
57%
18%
AMONGST FANS OPTIMEL (N=536)
63%
27%
46%
UPC (N=514)
Of all of the brands in the study, 3FM had the 15%
highest level of brand commitment amongst its
fans (81%), followed by FMCG brand Chocomel 0 20 40 60 80 100
(79%). Furthermore, 3FM and Chocomel dis-
played the most significant difference in brand
commitment between fans and non-fans. BRAND COMMITTED FANS/NON-FANS PER BRAND
N.B. Significance of differences: All differences are significant at p < .001
A noticeable factor is that brand commitment fans
amongst non-fans is more or less the same for non-fans
all brands. This already indicates that the being
a fan has a significant effect on brand commit-
ment. For all brands, the difference between
fans and non-fans is significant.
FACEBOOK ROI FOR BRANDS
22. FANS SPEND
MORE ON AVERAGE 36%
CHOCOMEL (N=436)
THAN NON-FANS 14%
75%
ESSENT (N=318)
67%
For nearly all of the brands, fans spend more than
non-fans. In all cases, with the exception of Essent, 63%
KPN (N=433)
the difference is significant. 44%
For example, on average, 36% of Chocomel fans spend 56%
OPTIMEL (N=536)
over €11 per month on the product, compared to 14% 20%
of non-fans. Furthermore, consumption is higher
amongst fans than non-fans. 68% of Chocomel fans 86%
UPC (N=514)
consume the product on a weekly basis, compared to 70%
19% of non-fans.
0 20 40 60 80 100
On average, 56% of Optimel fans spend over €11 per
month on the product, compared to 20% of non-fans.
82% of Optimel fans consume the product on a weekly FANS SPEND MORE ON AVERAGE THAN NON-FANS
basis, compared to 43% of non-fans. N.B. Significant differences:
UPC, Optimel, KPN, Chocomel: p < .001 Essent: n.s.
The differences are significant for all brands except fans
Essent. In the case of electricity, spend is dependent on non-fans
many other factors such as size of family, and there is
much less variation in expenditure. To a degree, this is
also the case for UPC and KPN.
FACEBOOK ROI FOR BRANDS
23. FANS ARE FAR MORE FRE-
3FM (N=772)
97%
59%
QUENT USERS OF PRODUCTS
CHOCOMEL (N=436)
68% THAN NON-FANS, ESPECIAL-
19%
96%
LY FOR FMCG BRANDS.
KPN (N=433)
94%
The research shows that fans use the products/services
82% of the brands investigated more often than non-fans.
OPTIMEL (N=536)
43%
For UPC and KPN, there is little to no significance in the
98% differences in usage between fans and non-fans. How-
UPC (N=514) ever, the usage of these brands is so high amongst both
99%
fans and non-fans that becoming a fan cannot really
instigate extra usage.
0 20 40 60 80 100
The FMCG brands Chocomel and Optimel, and the radio
station 3FM, on the other hand, display much more
USE AT LEAST ONCE A WEEK interesting figures. These show that usage amongst fans
N.B. Significant differences:
3FM, Chocomel, Optimel: p < .001 is/can be much higher than usage amongst non-fans.
KPN: p < .05
UPC: n.s. Essent was not included in this part of the research, as
fans neither fans or non-fans have any choice in whether or
non-fans not to use the product.
FACEBOOK ROI FOR BRANDS
24. FANS ARE MORE 3FM (N=772)
95%
31%
LOYAL TO A BRAND CHOCOMEL (N=436)
95%
71%
THAN NON-FANS ESSENT (N=318)
76%
59%
75%
The research shows that fans are more loyal to KPN (N=433)
47%
brands than non-fans. Once again, in all cases,
fans choose the brand in question more often 82%
than non-fans. OPTIMEL (N=536)
47%
Loyalty amongst 3FM, Optimel and Chocomel 79%
fans is the highest. For 3FM and Chocomel, at UPC (N=514)
53%
least 95% of the fan base states that they would
choose this brand again the next time.
0 20 40 60 80 100
For KPN and UPC, loyalty amongst fans is also
significantly higher than amongst non-fans.
WILL CHOOSE BRAND AGAIN NEXT TIME
N.B. Significant differences:
UPC, Optimel, KPN, Chocomel, 3FM: p < .001 Essent: p < .05
fans
non-fans
FACEBOOK ROI FOR BRANDS
25. FANS HAVE 3FM (N=772)
49%
9%
A HIGHER NPS CHOCOMEL (N=436)
64%
20%
Amongst fans, there is a significantly higher 21%
ESSENT (N=318)
number of promoters than amongst non-fans. 8%
This results in a significantly higher NPS score
for fans than for non-fans. 32%
KPN (N=433)
6%
Amongst fans, UPC and KPN have a positive
NPS (+12 and +17), which is rare in the market. 42%
OPTIMEL (N=536)
Amongst non-fans, UPC and KPN have an NPS 13%
of -36% and -41% respectively.
27%
UPC (N=514)
Once again, the FMCG brands achieved an 8%
excellent score. For example, NPS amongst
Chocomel fans is +60%, compared to -14% 0 20 40 60 80 100
amongst non-fans.
WOULD RECOMMEND TO FAMILY, FRIENDS OR
ACQUAINTANCES
PERCENTAGE OF RESPONDENTS WHO ARE PROMOTERS
N.B. Significant differences: All brands: p < .001
fans
non-fans
FACEBOOK ROI FOR BRANDS
26. However, an even more interesting aspect of this
SUMMARY study is the relationship between brand interac-
tion and brand commitment. Fans who interact
more with a particular brand have a higher level of
brand commitment. Furthermore, fans themselves
say that they have become more committed to
This study, which involved 3,009 respondents, the brand since becoming a fan. This shows that
showed the impact of brand interaction/engage- brand interaction on Facebook can contribute to the
ment on brand commitment. strengthening of brand commitment.
The positive relationship between brand commit- Furthermore, the research shows that a proportion
ment and trading results such as usage, spend, of the fans make more usage of the brand after be-
preference and NPS has been common knowledge coming fans. This is also explained by the interaction
for some time. The results of this study show that that the brand initiates with fans. After all, engage-
brand interaction on Facebook can lead to an in- ment boosts brand commitment, and brand commit-
crease in brand commitment in practice, which ment has been shown to increase conversion. Fur-
confirms the correlation between engagement and thermore, it has been shown that Facebook updates
trading results. increase usage/purchasing amongst a proportion of
fans. In this way, interaction via Facebook has a very
The results show that fans have a significantly concrete effect on the brand’s bottom line.
higher level of brand commitment than non-fans.
Fans spend more on the brand, use the brand This gives greater relevance to engagement/brand
more often, have a stronger brand preference and interaction via Facebook, due to the positive corre-
a higher NPS. This backs up the conclusion that lation between brand commitment and results.
building up a fan community on Facebook does
indeed add value.
FACEBOOK ROI FOR BRANDS
27. WANT TO KNOW
ABOUT THIS RESEARCH MORE?
If so, contact Steven
Jongeneel or Martin Kloos.
THANK YOU!
This study would not have been possible without the Social Embassy
co-operation and contributions of UPC, 3FM, Optimel, Huys Azie, Floor 1-a
Chocomel, Essent and KPN. You rock! Jollemanhof 8-a
1019 GW Amsterdam
020 435 0290
SHARE THE KNOWLEDGE www.socialembassy.nl
All parts of this report can be reproduced provided the @socialembassy.nl
source Social Embassy is quoted.
ABOUT SOCIAL EMBASSY
Social Embassy connects brands and people via social
media. We support brands with social-media strategy,
implementation and community management.We are
very proud to work for over 25 top brands including
Rabobank, Unilever and Volkswagen.
Every year, Social Embassy publishes the Social Media
Monitor, which researches the use of social media by the
top 100 brands (www.socialmediamonitor.nl).
FACEBOOK ROI FOR BRANDS
29. RESEARCH
STRUCTURE
This research was carried out in the period May- The fans were recruited via Facebook updates pub-
June 2012. Two groups - fans and non-fans lished on the brands’ fan pages. For each brand,
- were defined for each of the six participating two to three posts were enough to recruit a suffi-
brands. For each brand, the respondents were cient number of respondents via Facebook.
provided with a brand-specific online survey.
Fans and clients of UPC, for example, would Non-fans were located via a Multiscope research
therefore only answer questions relating to UPC. panel.
• Fans are Facebook users that have ‘liked’ the For this publication, the results of the six separate
fan page of the brand in question. surveys were aggregated in order to give us results
• Non-fans are respondents who have not ‘liked’ for all six surveys. The aggregates of the six sets of
the fan page of the brand in question, but who research results resulted in a data set of 3,009 re-
are customers of the brand. For the purposes of spondents: 1,272 fans and 1,737 non-fans.
this survey, the term ‘customer’ applies to people
who have used the product/brand within the last
three months.
FACEBOOK ROI FOR BRANDS
30. of analyses. If the groups differ by a significance less
ANALYSIS OF than 0.05, then the difference is significant. This means
that you can say with 95% certainty (or 99% certainty
in the case of p< .001) that the groups differ from each
RESULTS other and that this can be generalised to the total popu-
lation.
REGRESSION ANALYSIS
SCALE VARIABLES In order to discover the degree to which X leads to Y,
The questionnaire includes a variety of questions that regression analysis, including linear regression analy-
measure a construct. For example, brand commitment sis, is carried out. In order to determine how strong the
is measured using seven questions. By means of a correlation is, the R2 is examined. This figure indicates
factor analysis, it is determined to what degree these what percentage of the variance of Y can be explained
seven questions all measure the same aspect. If they by the variance in X. The α indicates the direction and
are deemed to measure the same aspect, then reliability the strength of the correlation. In this way, for example,
analysis is carried out to determine that the scale varia- you can state that as X increases, Y will increase to the
ble is reliable and therefore whether it can be stated with same degree, e.g. the more somebody interacts with
certainty that when combined, the variables measure, brand X, the higher the degree of fan-page commit-
for example, brand commitment. This is done if a scale ment will be.
becomes more reliable upon removal of one or more of
the variables. The following categorisation is used when The following rules are employed for determining the
applying the reliability analysis and creating the scales. strength of the correlation:
α < .60 = Unreliable 0 – .10 = Little or no correlation
.60 > α < .80 = Reasonably reliable .11 – .25 = Weak correlation
α > .80 = Reliable .26 – .50 = Moderate correlation
.51 – .75 = Strong correlation
.76 – .90 = Very strong correlation
.91 – 1 = Extremely strong correlation
DIFFERENCE BETWEEN FANS AND NON-FANS
In order to establish whether there is a real difference
between fans and non-fans, ANOVA carried out a number
FACEBOOK ROI FOR BRANDS
31. ENGAGEMENT LEADS TO BRAND COMMITMENT
RESEARCH DATA ß p< T F R²
Total .39 .001 14.60 899.55 .15
This slide and the following slides display the most
important results from the statistical analysis. 3FM .34 .001 7.70 59.23 .11
The tables have been divided between the conclu- Chocomel .39 .001 5.40 29.16 .15
sions drawn within this publication. Essent .65 .001 5.69 32.41 .41
KPN .48 .001 6.02 36.24 .22
Optimel .51 .001 8.91 79.33 .26
UPC .44 .001 6.68 44.70 .19
ENGAGEMENT LEADS TO CHANGES IN BRAND COMMITMENT
ß p< T F R²
Total .48 .001 19.17 367.62 .48
3FM .35 .001 8.01 64.23 .12
Chocomel .48 .001 6.92 47.95 .23
Essent .65 .001 5.73 32.86 .65
KPN .61 .001 8.45 71.35 .36
Optimel .51 .001 8.98 80.61 .26
UPC .46 .001 6.90 47.63 .20
FACEBOOK ROI FOR BRANDS
32. DIFFERENCES IN BRAND COMMITMENT BETWEEN FANS AND NON-FANS DIFFERENCES IN BRAND PREFERENCE BETWEEN FANS AND NON-FANS
p< F df p< F df
Total .001 972.98 1, 3007 Total .001 542.57 1, 3007
3FM .001 441.36 1, 770 3FM .001 655.20 1, 770
Chocomel .001 185.78 1, 434 Chocomel .001 46.92 1, 434
Essent .001 27.91 1, 316 Essent .05 4.47 1, 316
KPN .001 91.50 1, 431 KPN .001 44.08 1, 431
Optimel .001 129.91 1, 534 Optimel .001 109.65 1, 534
UPC .001 84.65 1, 512 UPC .001 44.87 1, 512
DIFFERENCE IN FREQUENCY OF USE BETWEEN FANS AND NON-FANS DIFFERENCES IN NPS BETWEEN FANS AND NON-FANS
* This question was not asked for Essent, as electricity is a product that is used every day.
p< F df
p< F df
Total .001 556.85 1, 3007
Total .001 306.77 1, 2689
3FM .001 413.39 1, 770
3FM .001 280.42 1, 770
Chocomel .001 130.80 1, 434
Chocomel .001 174.88 1, 434
Essent .001 18.70 1, 316
KPN .05 5.62 1, 431
KPN .001 80.02 1, 431
Optimel .001 134.98 1, 534
Optimel .001 131.93 1, 534
UPC n.s. .03 1, 512
UPC .001 69.82 1, 512
DIFFERENCES IN SPEND BETWEEN FANS AND NON-FANS
p< F df
Total .001 34.78 1, 2235
Chocomel .001 27.62 1, 434
Essent n.s. .10 1, 316
KPN .001 13.40 1, 431
Optimel .001 128.33 1, 534
UPC .001 30.00 1, 512
FACEBOOK ROI FOR BRANDS