As a catalogue marketer in the 1990s, I knew which mailing lists would work best for the dress shirt company I represented. We used multiple criteria to measure data quality. But, we always kept the list sources separate, so we knew what worked well and what didn't work as well. Each data set would become its own brand to us, because each data set would perform differently.
Of course I would track the results and optimize my campaign and work with the lists that brought the highest ROI. Since we were selling higher-quality men's dress shirts, lists from other higher quality men's apparel companies would generally perform better, though not every time. We would expect a performance difference between Florsheim shoe buyers and Johnston and Murphy shoe buyers, for example. Each list would elicit a different response, as you might expect.
I would buy these lists granularly and optimize against more precise data lists, which obviously required transparency from data providers, and a certain integrity in the data sets themselves. This may seem as obvious in terms of "best practices" to you as it does to me. So, ask yourself: Why does our online industry today do the exact opposite of this?
We have the technology
We can target and measure audiences in so many ways today; you would think it would make sense in this digital world that we would stay granular – and transparent. But, we're not. Instead, most of the data ecosystem strives for scale instead of precision, blended data sources instead of transparency and quality. Too many of us are making huge, blind buckets of data that are bloated and inaccurate.
Despite the incredible capabilities of today's digital technologies which enable incredibly granular tracking and segmentation, and the ability to build tremendously efficient campaigns based on collections of the smallest segments that work, too many buyers are doing the opposite. Today's buyers seek scale even though the best reason to scale is that it enables discovery of the target audience. Anything else is simple commoditization.
The promise of our data driven, real time ecosystem and the incredible efficiencies that it brings to display advertising is what is at stake. That's where we should be focusing. But today, too much of the audience targeting being done online is far less effective than what we did manually 20 years ago.
This seems pretty simple, right? Do you buy and measure the performance of your data sets this way? Or, do you buy huge swaths of data so you can scale your buy, not knowing which data sources worked best?
If scaling your buy is what you're doing, and it doesn't work, it's your own fault. The fact is that it simply cannot work as well as using better data would, in smaller, more precise campaigns. Unfortunately, this thinking runs completely counter to the fundamental underpinnings of what some of the largest and most acclaimed data marketing practitioners are doing online.Scale: the enemy of precision
A common marketer complaint of the data driven ecosystem is that it cannot scale. But this should not be a complaint – it should be a commandment. See, scaling this ecosystem means commoditizing the data that enables precise targeting by blending it and modeling it with less valuable data. All that does is make what you're doing suck more than it needs to. And that 's bad for everyone.
Many of us like to use the analogy that real time, auction based display advertising operates like the financial markets. But Wall Street is driven by asymmetric information, not commoditized data. The advantage exists in the unique, differentiated information you possess, not the same information that is available everywhere. The greatest asset – the greatest strength of what we do as online marketers - is the granular targeting that our medium's interactivity enables. Maintaining the integrity of the audience – YOUR audience - is paramount, whether this is an audience you have cultivated as a publisher (1st party data), or purchased as a buyer (3rd party data). The more exclusively it is held, and the more robust it is , the more valuable it is .
So, why do some buyers settle for accumulated, undifferentiated data? How can blended or modeled data work as well as exclusively held data that the buyer can identify? The fact is that it can't, and those who think it can are killing this segment.