Commentary

Fuzzy Logic: Not All Data is Created Equal

At all of the recent industry conferences, one unifying theme stood out: data obsession is part of a winning strategy. Marketers are now using data to inform every level of their decision-making. And that’s great. But not all data is created equal. 

Cheap data is available in abundance, making it all too easy for marketers to simply go through the motions. Unfortunately, probabilistic data from third-party vendors often turns data buying into a “check the box” exercise, where data is simply bought as part of a CYA exercise.

If a Data CPM seems too good to be true, it probably is

Although the allure is hard to ignore, it’s important to remember that cheap data (even well-branded) is cheap for a reason. The impact of most probabilistic data is nominal at best, with the majority of budgets going to ads that are either served to bots or out of view. It turns out that fuzzy logic -- algorithmic assumptions derived about consumers -- is generally misguided, and priced accordingly.  

Third-party data suppliers in no way guarantee that their math is perfect. As scary as it is to think about, there is no way of knowing how often the data is wrong. What percentage of media plans fail as a result?  Blind faith is a deadly strategy. Don’t trust fancy guesswork to retain clients and deliver outsized results, especially when brands have access to declarative, deterministic data.  

Eliminating the uncertainty

Deterministic data, or data derived directly from consumers without the use of assumptions or algorithms, assures marketers that they are reaching living/breathing humans. It offers some of the benefits of CRM-level data, but with the scalability necessary to target for major campaigns.

Deterministic data is rare, and therefore more expensive than probabilistic data, but the value it delivers for marketers is orders of magnitude greater than what inferred data can offer.

The implications of widespread deterministic data usage for the ad industry are vast. It will create a new group of winners. The data ensures that budgets will be utilized most effectively, eliminating guesswork and extraneous variables. Soon, marketers will be able to focus on defining a core target audience and reaching it with resonant messaging, rather than getting stuck on the tactics needed to reach that audience.  As a result, tremendous economic value will accrue to companies that can deliver first-party data sets at scale.

1 comment about "Fuzzy Logic: Not All Data is Created Equal".
Check to receive email when comments are posted.
  1. John Coctostan from Brightroll, November 13, 2014 at 11:27 a.m.

    Deterministic data is certainly more valuable than probabilistic data, but to call it "orders of magnitude" more valuable implies you slept through statistics class, Ari. Probabilistic data is generally accepted to be capable of 60-80% accuracy. Deterministic data is therefore more accurate by at best 2x, and yes priced accordingly.

    The broader benefit to using probabilistic data is that the data is anonymous and therefore "brand safe", something that agencies and their brands care deeply about.

Next story loading loading..