In the past month, I’ve heard lots of complaining about data. I’ve been guilty of it myself, and (gulp) will be now. I recently saw something that said that if you don’t have first-party data, you are nowhere, a loser in the great data race. That is simply not true. Second- and third-party data can be great, and first- party data can be garbage.
It’s like diversity. We should stop judging data based on who collected it, and look at what it really is.
(A note here on second-party data: It’s just third-party data you didn’t pay cash for, or maybe, somebody else’s first-party data.)
True, first-party data performs better. The reason is that category interest is likely going to be higher. Contacts within it are people who care about the category enough to have visited your Web site.
But there is more to dataset effectiveness than category interest. There’s reach (to humans). You want both — and more.
If you want to preach to the unconverted, you are kind of stuck with second- or third-party data. Don’t despair, though, until after you read the next paragraph, after which you may discover a whole new level of despair.
There are at least three very real problems with first-party data: 1) Folks identified may already purchase your brand, so they are not someone who needs to be persuaded. 2) They may have looked at your site and decided against your brand, fully informed, which means they are a poor prospect. 3) They may have come looking for a discount, in which case they are cherry-pickers — that is, not loyal, and lo- margin.
Add these up, and first-party data, already lacking reach, might not have the marketing muscle you expect. These are not the droids you’re looking for.
Also, regarding first-party data, people over-value what they already own. After all, it makes a good story (Competitive advantage! Free!). People feel less trustful about data they bought, but sharp marketers ignore the romance. Data works or it doesn’t.
But, even if you know a dataset worked, you only get a C+ unless you know why.
“Who” is the key to “why.” Consider this: The real, recency-adjusted prospects in your data might be 20% — or 80% — of the audience in the dataset. Nominally, talking to 4x more prospects would yield 4x the return. This is somewhat fuzzy, but intuitively obvious. Prospect density times reach will predict success pretty well — except you probably don’t know what the prospect density really was. Value, then, is reach to qualified humans whose purchase behavior an ad might change. No surprise, I hope.
Free data, if it does not contain the right audience, stops being free when you use it to buy media. Wrong data wastes more media than robots, I suspect.
As to the “why,” absent the knowledge of just whom you are talking to, if something good happened, you might attribute your success to the wrong thing (Creative, maybe?) and do more of the wrong thing as a result. As we know, marching efficiently in the wrong direction is way worse than going nowhere at all. A course correction requires you know where you are.
So, until you know about the people you reached, you can’t know what caused other measured outcomes. Awareness, for example, is pointless among people who don’t buy the category, and success has a thousand fathers.
Data has transcended being the new black. It’s here to stay, embedded in the learning loops held dear by brands, media companies, intermediaries, marketing clouds, and so on. Contact data (a cookie, mobile ID) is particularly powerful because it can be activated directly from learning. Your media, so carefully chosen, is only as good as your data.
Sometimes the origin of data gives clues about its fitness for a certain use, but if you characterize its marketing clout based only on that, there’s a disappointment in your future. Characterization based on an inference made by whoever collected it (first, second, third) is sketchy.
Sadly, characterization based on a whiff of truth is about the best we can do today, but as with all black holes in programmatic media, a little focus can go a long way. Focus on data.