VP of Data Scientist at ShareThis, Dr. Yan Qu, recently presented at the Global Data Science Conference about our study on the Return on A Share. Here are her slides.
2. With the growth of the social web, we turn more often to friends and strangers for
recommendations on where to go, what to eat, and what to buy.
2
SHARING IS THE CONNECTIVE THREAD OF THE SOCIAL WEB
3. 3
120 SOCIAL CHANNELS
95% OF ONLINE AUDIENCE
ShareThis Sharing Solutions
ShareThis is the largest sharing platform
reaching 211 million people over 120 social
channels on 2.8 MM sites. That’s 95% of the web!
2.8MM PUBLISHERS
4. YOUR AD
AUDIENCE IDENTIFICATION DELIVERY
HOW SHARETHIS WORKS
4
Jessica reads an article
about budget bedroom
ideas & decides to share it
to her Facebook page
using the ShareThis widget!
This is Jessica!
She’s busy chatting and
browsing the web…
ShareThis observes the share
and can then target Jessica
and her friends with
advertising messages tailored
to their interests
All social actions are then cataloged and segmented to build out
custom audience profiles based on social sharing behaviors.
YOU
R AD
6. 6
THE
IMPACT OF
SHARING
IS WELL-
DOCUMENTED
73% of survey respondents process
information more deeply, thoroughly, and
thoughtfully when they share it.
70% of people trust consumer reviews they
found online.
38% During a four-week test, Starbucks’
Facebook fans and “friends of fans” were
38% more likely to make a purchase.
10. SHARETHIS COMMISSIONED THE FIRST STUDY OF ITS KIND
TO MEASURE THE IMPACT OF SHARING ON THE
PURCHASE PROCESS
PURCHASE
INTENT
RECOM-
-MENDATION
PRICE
BRAND
PREMIUM
GENERIC
LUXURY
ECONOMY
TYPE
SOURCE
STRENGTH
Family
Friend
Acquaintance
Stranger
Professional
Excellent
Good
Moderate
Bad
In-person
Social Network
Trade Publication
12. 12
CONJOINT METHODOLOGY USES STATISTICAL MODELING TO
DETERMINE THE RELATIVE IMPORTANCE OF VARIOUS
FACTORS AFFECTING THE CONSUMER PURCHASE DECISION
http://sibmbaw11.blogspot.com/
13. BRAND
PRICE
TYPE
STRENGTH
SOURCE
RECOMMENDATION
WHICH OF THE FOLLOWING TWO PRODUCTS ARE YOU MORE LIKELY TO PURCHASE?
13
CONJOINT METHODOLOGY USES STATISTICAL MODELING TO
DETERMINE THE RELATIVE IMPORTANCE OF VARIOUS
FACTORS AFFECTING THE CONSUMER PURCHASE DECISION
x
Premium Brand
10% higher
Facebook Post
Friend
Economy Brand
10% lower
Car & Driver Review
Professional
Sample conjoint survey screen
14. 14
Study Details
Three industries (consumer packaged goods, mini-tablets
and automobile)
Two studies (traditional Likert rating scales and conjoint
analysis)
Four types of prices (10% lower pricing, comparable pricing,
5% higher pricing, 10% higher pricing)
Four types of brands (economy, generic, luxury, premium)
Four recommendation strengths (bad, moderate, good,
excellent)
Six thousand responses (1,000 for each industry / study)
15. 15
RECOMMENDATIONS ARE MORE
IMPORTANT TO THE CONSUMER PURCHASE
DECISION THAN BOTH BRAND AND PRICE
Recommendations account for 57% of the consumer
purchase decision:
RECOMMENDATIONS
57%
PRICE
28%
BRAND
15%
16. 16
SIMILAR PATTERN ACROSS DIFFERENT
VERTICALS
58%
34%
8%
56%
24%
19%
56%
25%
19%
CPG Electronics Autos
Recommendations
Price
Brand
17. 17
6.3%
7.3%
9.5%
10.6%
ONLINE SHARES ARE NEARLY AS VALUABLE
AS IN-PERSON RECOMMENDATIONS
Consumer Ratings
Consumer Reviews
In-Person Recommendations
Online Shares
Impact on purchase intent
20. 20
THE SOURCE OF THE SHARE IS
IMPORTANT
8.0%
7.7% 7.5%
12.7%
8.5%
9.0%
14.0%
9.6%
10.7%10.5%
11.7%
CPG Electronics Autos
Strangers Acquaintances Close Friends / Family Professionals
21. 21
FOR AUTO BRANDS, AN EXCELLENT
SHARE IS WORTH $3,708
Consumer Ratings
Consumer Reviews
In-Person Recommendations
Online Shares
Professional Reviews
$2,312
$3,277
$3,708
$3,766
$4,702
22. 22
Consumer Ratings
Consumer Reviews
In-Person Recommendations
Online Shares
Professional Reviews
$22.09
$23.56
$24.91
$28.76
$32.44
FOR ELECTRONICS BRANDS, AN
EXCELLENT SHARE IS WORTH $25
23. $0.42
$0.44
$0.92
$1.05
23
Consumer Ratings
Consumer Reviews
In-Person Recommendations
Online Shares
FOR CPG BRANDS, AN EXCELLENT
SHARE IS WORTH $0.92
That’s how much more consumers are willing to pay for the average
supermarket product if it was positively shared about by someone in
their social network.
25. 25
Recommendations have more impact on a consumer’s
decision-making process than brand or price
Highly positive online shares generate a 9.5% increase in
purchase intent, while negative reviews can decrease intent
by 11.0%
Key Findings
26. 26
Recommendations have more impact on a consumer’s
decision-making process than brand or price
Highly positive online shares generate a 9.5% increase in
purchase intent, while negative reviews can decrease intent
by 11.0%
Online sharing has a measurable effect on a product’s
prices, driving incremental value by as much as $3,708 for
automobiles
Online shares are nearly as valuable as in-person
recommendations
Key Findings
28. DRIVE EARNED MEDIA SUCCESS WITH
SOCIAL INSIGHTS
VIEW CURRENT CUSTOMERS AS MARKETING PARTNERS
ENABLE CUSTOMER EVANGELISM
MONITOR YOUR ONLINE PRESENCE THROUGH SHARING
ADDRESS NEGATIVE REVIEWS OPENLY AND HONESTLY
29. Use social signals to determine where and when to spend media dollars
UNLOCK THE VALUE OF SHARING WITH PAID MEDIA
Use social data to identify users who shared relevant content
IDENTIFY
TARGET
Align messaging with sharers’ interests and deliver media on the content
with which they engage
OPTIMIZE
Heavy-up deployment around key
events to drive conversion when
conversations are abuzz
Monitor sharing trends in real time and
capitalize on user intent as quickly as
possible throughout the purchase funnel
30. 30
THE VIRTUOUS CIRCLE OF SHARING
SHARING
INSIGHTS
OPTIMIZATION
EARNED MEDIA
LARGER AUDIENCE
Explosion of social networks, not just in terms of adoption, but also proliferation; from FB & Tw to 4Sq and Pinterest.
What they all have in common is that they are a way for people to share things
This has created an unprecedented amount of user curated content
In fact, there has been so much new content that it has changed the way we make buying decisions.
Effect is well documented; 70% of people trust the consumer reviews they find online – this didn’t exist 10 years ago!
But what really is the monetary impact of all this and how can a brand manager plan for it?
There are a number of factors that play into purchase decisions. Brand, price, recommendation, etc.
Clustered within these are even more parameters (type, strength, etc.)
How do we pick apart the impact of each of these? Can we ask people? Bias from what people say they do vs. what they actually do. Unconscious desires, Freud, etc.
Traditional survey approach (likert) doesn’t effectively weigh the relative importance of all these variables. Moreover, what people say they do (i.e. survey response) is different from what they actually do.
Conjoint analysis is a statistical technique used in market research to determine how people value different attributes (feature, function, benefits) that make up an individual product or service.
Therein lies the power of Conjoint methodology. Instead of asking questions like “Rate the importance of BRAND to your purchase on a scale of 1 to 10”, we lay out a series of choices (tradeoffs) broken out by purchase factor.
With enough data points (~20 tradeoffs per respondent over 6,000 respondents) we can then statistically model the relative importance of EACH FACTOR.
This has never been applied to the concept of online sharing.
Respondents are all recent or active purchasers.
Auto/tech: past 6 months
CPG: primary supermarket shopper in household
Therein lies the power of Conjoint methodology. Instead of asking questions like “Rate the importance of BRAND to your purchase on a scale of 1 to 10”, we lay out a series of choices (tradeoffs) broken out by purchase factor.
With enough data points (~20 tradeoffs per respondent over 6,000 respondents) we can then statistically model the relative importance of EACH FACTOR.
This has never been applied to the concept of online sharing.
Respondents are all recent or active purchasers.
Auto/tech: past 6 months
CPG: primary supermarket shopper in household
For consumer packaged goods pricing, we used $4 for a typical store brand grocery product, $6 for a typical standard brand, $8 for a typical quality brand and $10 for a typical premium item.
For the mini-tablets, where we identified specific brands, we used actual prices of each mini-tablet. Apple’s mini iPad, Samsung Galaxy, Dell Venue and Google Nexus. The other adjustment we made for this test was to use specific price points ($189, $229, $309, $329 and $359).
For the automobiles, we used Kelly Blue Book (www.kbb.com) to price a 2014 Kia Rio LX ($14,100), a 2014 Ford Taurus SE ($26,100), a 2014 Audi A6 ($42,200) and a 2014 Jaguar XK ($76,500).
This chart plots relative importance of each factor. Recommendations (whether personal, online, etc.) drive 60% of a consumer’s purchase decision. A considerable weight.
So we know recommendations are important. What about the type of recommendation. We were particularly interested in the online share
Not all recs are the same. A recommendation, of course, is comprised of many components. We focused on the three most important components: the strength of the recommendation, who made the recommendation, and the source of the recommendation or share. Again, the results were surprisingly consistent across the three product areas: the strength of the recommendation (bad, moderate, good and excellent) is far more important than the source or even the familiarity of the person who made it.
Online shares more valuable due to proximity of relationship. Social networks are usually friends and family. At the end of the day, being told something in person has more weight.
Looking across the verticals, once again we see a similar pattern. The impact is particularly marked for CPG brands where the impact of ratings and reviews has been less
For consumer packaged goods pricing, we used $4 for a typical store brand grocery product, $6 for a typical standard brand, $8 for a typical quality brand and $10 for a typical premium item.
For the mini-tablets, where we identified specific brands, we used actual prices of each mini-tablet. Apple’s mini iPad, Samsung Galaxy, Dell Venue and Google Nexus. The other adjustment we made for this test was to use specific price points ($189, $229, $309, $329 and $359).
For the automobiles, we used Kelly Blue Book (www.kbb.com) to price a 2014 Kia Rio LX ($14,100), a 2014 Ford Taurus SE ($26,100), a 2014 Audi A6 ($42,200) and a 2014 Jaguar XK ($76,500).
For consumer packaged goods pricing, we used $4 for a typical store brand grocery product, $6 for a typical standard brand, $8 for a typical quality brand and $10 for a typical premium item.
For the mini-tablets, where we identified specific brands, we used actual prices of each mini-tablet. Apple’s mini iPad, Samsung Galaxy, Dell Venue and Google Nexus. The other adjustment we made for this test was to use specific price points ($189, $229, $309, $329 and $359).
For the automobiles, we used Kelly Blue Book (www.kbb.com) to price a 2014 Kia Rio LX ($14,100), a 2014 Ford Taurus SE ($26,100), a 2014 Audi A6 ($42,200) and a 2014 Jaguar XK ($76,500).
To equate sharing with a dollar value, we multiplied the percentage impact attributed to shared recommendations by the average price of the products we were looking at. For example, we found that positive online shares drove a 9.2% increase in purchase intent among autos; so 9.2% x $39,000 (the average price of vehicles we surveyed) amounts to $3,708 in additional value.
Of course, we can get even more granular with this data by breaking out these values across price points, recommendation source, and recommendation type. An online share from a close friend (11.2% increase in perception) about a luxury vehicle (avg price: $76,500) yields roughly $8,676 in additional value.
For consumer packaged goods pricing, we used $4 for a typical store brand grocery product, $6 for a typical standard brand, $8 for a typical quality brand and $10 for a typical premium item.
For the mini-tablets, where we identified specific brands, we used actual prices of each mini-tablet. Apple’s mini iPad, Samsung Galaxy, Dell Venue and Google Nexus. The other adjustment we made for this test was to use specific price points ($189, $229, $309, $329 and $359).
For the automobiles, we used Kelly Blue Book (www.kbb.com) to price a 2014 Kia Rio LX ($14,100), a 2014 Ford Taurus SE ($26,100), a 2014 Audi A6 ($42,200) and a 2014 Jaguar XK ($76,500).
For Brand Strategy & Social Media Teams – provides tangible metrics for earned media strategies.
Valuate shares of brand content to better understand success of earned media strategies and optimize accordingly.
Provides basis for understanding brand perception around the web, and generating new and shareable content to drive engagement and advocacy
For Digital Media Planners – provides justification for identification, targeting, optimization based on sharing
Recommendations drive a vast majority of the purchase decision, meaning sharing is an inherently powerful indicator of interest and consideration.