What Data Science Can Learn From The Slow Food Movement

You may recall a scene in the first season of Portlandia that features a couple at a farm-to-table restaurant asking about the chicken on the menu. 

They don’t just want to know how it’s cooked. They want to know what the chicken was like. The waitress informs them that the chicken is a heritage breed, woodland-raised chicken that had been fed a diet of sheep’s milk, soy and hazelnuts. She also presents the couple with the chicken’s papers. But that’s not enough.

The couple needs to know if the chicken — Colin —  was raised organically, how much acreage he had to roam and whether Colin was happy and got along well with its fellow chickens. Even then, the couple has to trek 30 miles away to visit the farm in person to investigate for themselves. 

The skit is of course a reductio ad absurdum of the idea of local sourcing and the slow food movement. But the idea it parodies has become fairly commonplace: Healthy eating involves knowing where your food comes from, how it was grown, how it was processed, how it got to your plate. 



Marketers can learn a lot from that ethos in the era of data-driven decision-making. In the same way that we “are what we eat,” our data-driven algorithms are only as good as the data that feeds them. We should all apply the same rigor in asking where our marketing data comes from, how it is processed, priced, and optimized. 

Bad-quality data leads not only to wasted ad spend, but further inaccuracies in downstream attribution, leading marketers to believe that their inefficient campaigns proved more successful than they actually were. 

Transparency into the real trade-offs between cost and quality

One shorthand way of ensuring quality is to pay a higher price. But a relatively high CPM (let’s say $8) might not be a marker of quality. Instead, it might indicate that the vendor was bundling disparate elements like data, technology, media and measurements — so that the price of the audience reflects the added cost of service rather than the addressability or accuracy of the segment. 

The truth is that we often don’t know why vendors charge what they do because they don’t offer any transparency into how they arrive at those prices. 

Just as higher prices don’t mean that you’re getting more for your money, lower prices aren’t an indication that the inventory you’re buying is poor. In the culinary world, there’s room for Taco Bell and McDonald’s and also for Benu or Eleven Madison Park. 

A decree for transparency from start to finish

While there’s no panacea for transparency, a more long-term strategy to foster honesty about quality lies in the marketer’s and vendor’s abilities to articulate their data imperatives. 

First, vendors must be forthcoming about how data is sourced and processed, which has absolute influence over the output. Data processors must help marketers understand, in plain English, the meat and potatoes of how data is sourced, cleaned, cross-referenced, organized, and managed. For example, is the raw data unaltered—deterministic—or is it analyzed to derive additional insights—probabilistic? 

In order to build a reciprocal relationship, marketers must know to ask about processes that directly relate to their objectives and may impact their return. For example, understanding the data’s coverage—will the final product actually help you reach the people you expect to reach? How will you determine success?

Measurement is a utility where marketers and vendors often misalign. Marketers must ensure they clearly communicate with vendors on how they plan to calculate success and what attribution methodology they will use, otherwise a campaign may not deliver on expectations.

To ensure high quality, you have to ask some tough questions, such as: Where do you get your data? How do you get it? How can you prove it works in my industry and for my initiatives? 

If you don’t, you could end up with the data equivalent of mystery meat, when—for the same price—you could instead have sunk your teeth into a happy, well-adjusted chicken named Colin.

Next story loading loading..

Discover Our Publications