Insights Are For Brands, Data Is For Direct Response: The IMP Is Born
- First is a validation function. Brand advertisers buy audiences. When they buy at the top of the funnel, they should be able to validate that the audience they’ve bought is the audience they are actually reaching. Today, we can make that validation, thanks to audience measurement and validation tools delivered by the likes of comScore, Nielsen, Mendelsohn, Moat and many others. Validation TV for advertising is both standard and analog. When an advertiser buys a nightly show with a specific audience, the audience is measured for size and demographics and is scored accordingly. If the score or reach was not met, the TV advertiser gets a make-good for the next night. Get ready for the validation era.
- The second function of an IMP is measurement. Where many legacy platforms have been built and measured around click-through rate, the insights management platform is built around brand metrics, such as engagement or interaction rate. With these measurements, success usually requires engagements or interactions greater than 30 seconds in duration, which is much higher than the rate expected for TV ads or even the length of a TV ad itself. Branded response measures are a necessary and integral part of an IMP.
- The third function of an IMP is as a data marketplace. It’s critical to ensure that, as an advertiser, you’re getting insights for your own brand and products as a first party, not only the insights from a second-party partner. Brand advertisers should also be vetting third-party insights. The marketplace is a clearinghouse to connect to all the data available to advertisers in an open data exchange or marketplace. This is not simply third-party data, but insights from players like eXelate, BlueKai, and Axciom. A marketplace is necessary to aggregate those data in order to turn it into insights, and then distill and validate them via the IMP.
- The fourth piece, which is most important, is the ability to predict. The platform must be able to determine which sites, which pieces of content, and which formats, channels and creative elements will perform well in a brand campaign. That prediction is an effort to determine the additional audiences with which a brand’s message might resonate, as well as new formats and channels they might consider. A prediction engine contains an optimization algorithm driving the brand advertising to the right audience, at the right time, to the right screen, with the right creative and content.
So how does an IMP differ from a DMP? A data management platform houses first- and third-party data, and gleans behavioral and other types of data that can be placed against a campaign that’s about to run. The DMP allows advertisers to select inventory more strategically, so there’s less waste in each campaign. It typically sits at the bottom of the technology stack, looking for relevant data.
On the top of that stack is the CRM system. The CRM, which might be an SaaS platform like Salesforce, is where an advertiser or its partners typically house information about customers, including contact history.
The IMP lives in the middle, between the DMP and the CRM. Its role is to reveal general trends that help fill the top and middle of the funnel. By no means am I suggesting that an IMP should replace a DMP or a CRM system. However, for several years, many companies have relied on cloud-based, data-driven CRM systems, and today a growing number of companies license or build DMPs. The IMP embraces and extends the functions of the DMP and CRM.
If “Data 101” was leveraging behavioral data to improve direct response, I believe we’ve graduated to “Data 201,” which is all about insights. It’s now necessary to develop that IMP to help brand advertisers finally zero in on the context they want, working with the insights they need, within the formats and channels that will reach their audiences most efficiently. Actionable insights are the future of digital brand advertising.