This document describes Insights Services, which uses data and analytics to help clients make smarter marketing decisions. It provides three examples of how Insights Services has helped clients: 1) Identifying undiscovered customer segments for a telecom company; 2) Filtering out job site visitors who were not actively job searching; 3) Building a look-alike model to lower customer acquisition costs for a bank seeking new retirement accounts when traditional targeting failed. The document concludes by introducing a self-serve web tool that allows advertisers to efficiently analyze user behavior and target audiences.
2. AGENDA
• WHAT IS INSIGHTS?
• HOW WE DO IT AND WHY IT’S BETTER
• SOME REAL WORLD EXAMPLES
• HOW YOU CAN DO IT, TOO (THE CATHARTIC AND
EMPOWERING CRESCENDO).
3. INSIGHTS: EVERYONE SAYS THEY’VE GOT IT COVERED
MEDIA OWNERS DATA OWNERS RESEARCH FIRMS
RESEARCH FIRMS
MEDIA AGENCIES AD EDIA AUDITORS
M NETWORKS DSP’S
4. EVERYONE SAYS THEY HAVE LOTS OF GREAT DATA
FIGURING OUT WHAT’S GOOD TAKES THREE STEPS.
1. What’s the data asset?
2. How do they make sense of it?
3. How can I use this to plan a campaign?
PLANNING
ANALYSIS
DATA
5. TRIBAL FUSION INSIGHTS SERVICES
• Discover the story behind your
audience
• Learn something you didn’t already
know about your customer, and act on
it swiftly
• Strategic recommendations help you
– Improve response rates
– Extend your audience
– Get more for less
6. EXAMPLE 1: FINDING THE AUDIENCE YOU DIDN’T KNOW YOU HAD
MARKETING CHALLENGE
• A major mobile telecommunications company is tasked with efficient customer acquisition. In
a mature market with over 100% penetration, this means stealing customers from
competitors, with diminishing returns on marketing efforts.
UNDERLYING ASSUMPTION
• Mobile is a commoditized offering. The product is not differentiated, so it does not make
sense to segment the consumer based on lifestyle, underlying needs, etc.
7. EXAMPLE 1: FINDING THE AUDIENCE YOU DIDN’T KNOW YOU HAD
Tribal Fusion profiled converters and uncovered two distinct personas
IMPLICATIONS
• Conversions are actually people with full lives.
• Speaking to the whole person can improve performance.
• Separate creative to these segments leads to higher conversion rates.
• Look alike models built off these behaviors extend reach in relevant places.
8. EXAMPLE 2: WEEDING OUT THE CUSTOMERS YOU DIDN’T WANT.
MARKETING CHALLENGE
• A job site wanted to know why people were visiting their site but not posting resumes.
• How can they do a better job of bringing in the right users, and encouraging them to convert?
UNDERLYING ASSUMPTION
• If no one submits resumes, then employers won’t pay to post jobs. If there are no jobs
posted, then no one submits resumes. Left uncorrected, and the whole business could
implode.
9. EXAMPLE 2: WEEDING OUT THE CUSTOMERS YOU DIDN’T WANT.
TRIBAL FUSION ACTION PLAN
• Stop serving ads to people who are clicking
but not in the job market.
— No more CPA, CPL, or CPC campaigns
Site Visits
• People looking for specific job profiles are
clearly in market, so retarget them until they
convert.
• Recent grads and moms are dipping in and
out of the job market. Hit them with
branding campaigns, so you are top of mind
when they’re ready.
Resumes
10. EXAMPLE 3: DELIVERING CONVERSIONS EVEN WHEN TRADITIONAL
SEGMENT/TARGET/POSITION MARKETING FAILS.
MARKETING CHALLENGE
• A bank was tasked with opening new retirement accounts.
• Market research defined the target audience as
– Aged 17-64
– Above the poverty line, but not super rich (they have other investment vehicles).
– Working or the spouse of someone who is
UNDERLYING ASSUMPTION
• Targeting basically everyone, so a fancy data strategy won’t help.
• Stack ‘em high, sell ‘em cheap
11. EXAMPLE 3: LOOK ALIKE MODELING, BASED ON TRIBAL FUSION DATA
TRIBAL FUSION ACTION PLAN
• TF put a pixel on the conversion page for this
product.
• Mined the data to see which behaviors
indicated lift.
• Built a model to de-duplicate.
• Served ads to people who had those traits,
whether they intuitively “made sense” or not.
RESULTS
• Look alike model cut the effective CPA in half
13. A SELF-SERVE, WEB-BASED TOOL FOR ADVERTISERS AND AGENCIES
• Easy to navigate, intuitive interface
• Quickly surface the user behaviors that drive performance
• Reduce wastage by targeting the right people with the right message
• Extend your audience by defining your own look-alike model
Notes de l'éditeur
Not meant to be an exhaustive list, but most of the players on the previous page are using some combination of the assets described here are building their own analysis on top of it.
ANIMATEDThe ratio of impressions to conversions is key. Here we see our four target groups towards the lower right, and our Placid group on the top. [expand]