How to Squeeze the Data Out of Your Online Media
Today, planners, brands and creative people strive to get insights by looking at all sorts of data, including interactive data.
A.G. Lafley, former P&G CEO, once said that the best managers are the ones that "face reality soonest." That's all good, but the hard part is finding out what reality is in the first place. Many techniques rely on asking people questions, or constructing a contrived situation.
By contrast, on the web, we can get pure observation. The vast amounts of data available are observations of what consumers actually did in the wild.
Albeit messy, electronic voyeurism unquestionably gives marketers a glimpse of reality not available elsewhere.
Even better, on the web, the advertiser owns the data from their own campaigns by default.
Smart direct-response companies conduct experiments using the data they own. Change the font. If the ad response isn't better, change it back. Takes an hour. Costs nothing. Life is good when you have a closed loop, a clear and cost-free performance indicator, and an office full of analytics people.
Branding companies, however, lacking a closed loop, usually don't have this luxury.
What data is available to help those with no loop back to purchase? Options are abundant. Search can, for example, tell you how customers think about your product, and it can yield connections you never imagined via "affinity." Search-affinity analysis can examine behaviors of people who search for "eyeliner." If that group also searches heavily for "Kim Kardashian" you have a candidate for casting your eyeliner commercial.
Cookies can be constructed to cheaply and quickly yield target groups. Put an offer in Spanish and cookie the landing page. Now you have an anonymous list of Spanish-speaking consumers, and you can use cluster analysis to see how those consumers behave differently than other groups.
Which segments in a behavioral campaign hovered over your ad? Those are the people who resonate with your message. Link your Omniture tag to a panel and find out who those consumers really are. Harvest your social data with "listening" methods, and look at the interests of those who "like" you. It goes on and on, but companies don't seem to be chumming up. Here's why they should.
If your website (a sunk cost) tells you that girls with curls resonate with your product (they dwell time on pages about curly hair), you might cast curly-haired characters in your copy. If you are willing to attribute a small increment of lift to the smarter casting, your website data likely earned you way more than the data cost you.
Here's the problem. None of this is easy. It's all a little fuzzy, and it requires that analytic meet strategic. You need serious geeks, computers and a plan to acquire the data you need.
Your digital buyers are unlikely to be rewarded for crafting reach that explores new possibilities for who will resonate with your message, and geeks are famously bad at convincing Harvard MBA's that it's worth it to spend money to wallow in the data. Measurement companies are getting smarter all the time, and are quite capable of using these methods, but who is asking them to do it? We're all learning.
I believe brands should declare what data they need, and then set about getting it by any means possible. Advertisers: Don't just ask agencies for "insights," ask for a learning plan. Ask great questions, and don't give up on data just because it's not available. Curate it. The tools are available. Find a way to get it.
There are, after all, a million knobs to turn.