In the late 90s he joined Avenue A/Razorfish and much of his time there was spent figuring out how to use data to enhance running campaigns. He then joined Atlas and used the data they were collecting to launch an advertising network - with only $100k investment they launched a $100M business.
Every day we face campaign performance issues with our clients who ask for adjustments to the campaign to compensate for shortage in its success.
Note all data is created equal
In a study of conversion performance (lifet vs. untargeted media) a consistent pattern emerges: contextual advertising best in breed yields 132% lift. Best demographic targeting: 550% and Best behavioral: 1235% (!). These lifts demonstrate the yield improvements if you launch data into your campaigns.
Realistically: if remenant price is $0.50 this isn't such a large impact but it has the potential to be lucrative:
The economics of ad nets and exchanges:
AdNets buys ads for ~$0.5 to sell it for $5 to $8.
Applying the model to a publisher ecosystem can yield the same economics. In a recent partnership with large data providers, ACXIOM was able to leverage behavioral targeting to yield the following:
What should you do about data?
1. Enhance your data with client and 3rd-part information
2. Treat data as an enterprise function. The folks in your company looking at (advertising) data should include the folks who work with remenant advertising, feedback loop to the CMS for improvement to the user experience,
3. Seek partners with data breadth and scale
Scott claims the multi-variate data lift represents a 5,450% lift (!). Multi variate models generate breakthrough results and can become potential rainmaking opportunities, e.g.
brand affinity model
auto country of origin
defector models
mini defector models
ownership models
conquest models
auto consumer dynamics
An example of a integrated data campaign was for a CPG that ACXIOM enhance with their data to better message to their CRM list and determine in0store impact of multi-channel messages. In this example advertiser CRM data was layered with publisher data and other 3rd party data. There was a $20m in store sales lift and 11% share of wallet with a 4% ROI on the campaign
Pitfalls
Tunnel vision (e.g. just remnant inventory, just social, just mobil)
Operational costs: don't try to hire PHd in data or you will hurt the economics in your organization
Partner selection: make sure you are aligning with real companies vs. innovative features. That your partners can handle the data scale, privacy and security of your data.
Privacy: as the world moves toward increased privacy controls over consumer data - consumers become an underrepresented constituency
No comments:
Post a Comment