by Chuck Moxley, CMO
attribution, "campaign measurement", cross-screen, "mobile marketing"
Shaking my head when this happens, but it does sometimes … In a client meeting, I’ll present how we’re doing something in mobile advertising that wasn’t possible before in any medium — tying impression data to actual in-store sales transactions and enabling third-party sales lift measurement (i.e., full closed-loop attribution) — when, surprisingly, the discussion devolves into picking at certain elements or certain exceptions (e.g., cash transactions in fast food) and how sales lift measurement won’t work because it isn’t perfect.
Who doesn’t want absolute attribution precision when delivering an ad? But is it wise to refuse to engage your customers on their mobile devices because the technology doesn’t guarantee attribution perfection when 4INFO’s mobile attribution is the closest thing to perfection that any media planner, buyer or marketer has been able to experience in the history of advertising?
Consider your measurement alternatives. In TV, radio and print, attributing a purchase to a specific ad is an educated guess. TV has evolved some, from total guesswork where, at most, you could hope to identify a vague sense of brand recall or awareness, to smart TV sets with automatic content recognition … but still not enough scale to tie ads seen on linear TV to in-store sales transactions.
Even the bulk of the most sophisticated digital advertisers still rely on clicks (which give no attribution relevance or even negative correlation, as stats show) or store visits (you know the downsides), or not measuring attribution at all.
But we’ve come light years ahead with mobile/digital attribution. We can concretely tie exposure to an ad to a sale. First time in history!
Which means that, for the first time ever, an advertiser can concretely identify ROAS. Companies like 4INFO make it possible for an advertiser to know that, “For every dollar you spent on that campaign, you saw a $3.82 gain in sales.”
Exciting stuff, right?
So why aren’t more advertisers embracing it? Are you missing this attribution revolution while waiting for perfection? Sure: a digital ad can never be perfectly attributed if you focus on the exceptions to what actually works, but it sure beats relying on unreliable clicks or store visits!
Attribution perfection is likely never going to happen because no one company has all the sales data you want to perfectly capture every member of a target audience. There will always be exceptions. For example, no one has Walmart’s data but Walmart. No one has Costco’s sales data but Costco. And so there’s no way to get it all.
That said, there are companies solving for this. Nielsen Catalina Solutions and Kantar Shopcom use a panel of 100,000 shoppers who scan all of their purchases, and use that to make an “all outlet” adjustment to account for purchases at all stores. It’s a statistically valid and proven method.
Rather than hold back and not measure your mobile campaign, take a cue from modern software development: test and learn and adjust. Many companies, like Facebook, have adapted the “Agile” style of project management, which uses the “fail fast” strategy; instead of developing a product on the basis of one big project, do it in pieces, take what you’ve got and run a test. If it fails, then you learn from it and fix your method, and put it out there again. Test, learn, adjust.
“It’s a journey,” noted Ravi Parmeswar, VP of global strategic insights and analytics at Johnson & Johnson, during a recent industry panel at a recent measurement and attribution event. Another panel member, Pravin Chandiramani of Simulmedia, summed it up nicely: “If you wait for the end state, you won’t have a business for 10 years and by that time the world will have moved on.”
In short: If you’re waiting for perfect targeting/attribution, you’ll always be waiting. And you’ll be missing out on the closest thing to perfection you’ve ever had access to.
For additional insight, check out Ad Exchanger's coverage of the idustry panel at the measurement and attribution event mentioned above.