Big data, data-driven, real-time, analytics. Enough to make your head spin because information services is an area where utilizing a mess of individual point-solutions may leave you more clueless than when you started. Imagine reading The Great Gatsby in nine languages, a different one for each chapter -- the themes, writing style and expression may not come through as the author intended. In the same manner, the story you tell of your customers, campaigns, or community is more effectively told when you invest in a common "language." When it comes to digital marketing and information services and analytics, I'd argue that the data methodology is more important than the solution itself. I'll explain.
Let's say you have a dozen data companies, all of them able to target whom they call (the favorite industry cliche) "soccer moms." The obvious question here is to ask why these companies identify specific users as soccer moms. Let's explore some possible reasons why a data company would place a user in that segment:
- Visitors of "soccermoms.com"
- Or the "mom" section of soccer.com
- Or the "soccer" section of mom.com (you get my point, content-based)
- Purchased kids soccer balls, cleats, shinguards, or other gear online.
- Belongs in a group of like-minded households that have a higher propensity to have kids playing soccer
- Searched for "kids soccer league" online
- Watched a "how to cut orange slices" video online
- Subscribes to Grass Stains and Goalposts magazine offline
- Responded to a survey asking why they bought their minivan with "to haul the entire team to games"
- Looked at pictures of kids playing soccer on a photo sharing site
- Downloaded soccer game apps, or an MLS app
- They are female in the age range of 25-45 living in the suburbs
- They previously responded to similar ad campaigns
- They live in an area with a high concentration of soccer leagues, or kids
This random handful of examples outline how little we know about a "soccer mom" segment at face value. If you don't care where your data comes from, you're essentially saying that any of the above reasons in an "or" statement" suffices. You're basically saying that as long as your client pays you and a 3rd party tells you that your segment(s) are accurate, you don't care. Also, I've heard that "where the data comes from doesn't matter, just as long as it performs." To that I'll ask you these questions:
- As long what you eat is called "food" and cures your hunger, so you not care what it is where where it came from?
- As long as you get from point A to point B, does it matter where you sit on a plane? What if you had to stand the whole trip? Sit in the wheel well?
People in this industry love to consider themselves "scientists," and these days specifically "data scientists." When a scientific experiment is conducted, you measure cause and effect of changing one or a few variables. After a set of results you can start to conclude that your hypothesis was aligned with reality or not. How then does "where the data comes from doesn't matter, just as long as it performs" have a place in your "science?" This is like performing titrations in chemistry class, throwing any mix of solutions in a beaker until PH reaches 0.
Do yourself (if you're a marketer) or your client (if you're an agency or platform) a favor and try to understand the sources that your 3rd party data providers are using. Are the sources authoritative? Refreshed often? Does your partner own the data? These are all important to know as you could be paying a pretty penny for people that are "photo enthusiasts" that simply visited flickr, looked at a camera accessory on eBay, or something related and as immaterial.
I see data the same way as I see remnant inventory monetization -- "garbage in, garbage out." Look for companies that build off of an infrastructure you can trust, a segmentation methodology that spans across multiple industries, and would perform multiple tricks if it were a pony. Especially in this time where there are as many as 80 companies that Forrester is considering for their upcoming DMP Wave report, remember that not all pixels are created equal.
Recent Comments