The Transformative Power of Unstructured Data


Describing how much data there is in the world is like describing how many grains of sand there are on Earth, or how many stars exist in the universe; it’s a lot — a whole, whole lot. Humans create, give or take, 2.5 quintillion bytes of data every day, and the International Data Corporation, whose job it is to know these things, projects that we’ll be pumping out 163 zettabytes of data annually by the year 2025. For those counting at home, a zettabyte has 21 zeros.

Not surprisingly, it was hard to make heads or tails of it all in the early days of the internet, so digital gatekeepers began archiving and organizing certain data points — things like website content for search engines — to instill some structure and keep a relative lid on things. This is what’s known today as structured data. But — spoiler alert — the internet kept growing, eventually housing massive amounts of audio, video, GIFs, emails, tweets, GPS data, timestamps, and other forms of digital information on top of everything else. This, conversely, is called unstructured data, and it makes up 80% of all the data out there. So if you aren’t leveraging unstructured data, you’re missing out on 80% of the information available to help you create a customized audience and optimize your campaigns with the same element-level control.

The limits of structured data stretch far beyond percentages, though. By its very definition, structured data is information that has been scrubbed and organized and lumped into prepackaged categories. In advertising terms, this means it’s been placed into a nontransparent audience segment — a digital black box of sorts — and sold to marketers and DSPs as indicative of potential consumer behavior. Everyone in that segment is, in essence, the same person, with the same motives, interests and character traits as their segmented brethren. For marketers, whose ultimate goal is to connect with individual consumers on a personalized level, this can be problematic. There’s also increasing skepticism about the validity of the data contained within audience segments. So not only is the data within an audience segment potentially inaccurate, there’s no way for you to know because it’s a black box that you’re not allowed to open.

This isn’t a problem with unstructured data, which offers more information, more flexibility, and more reliability. The freedom to build dynamic audiences with pinpoint precision, right down to the exact mix of desired individual data points, provides advertisers with the ability to create and customize their campaigns to scale. Add things like recency, localization, and other unstructured elements — that 80% of the world’s data we mentioned earlier — to the mix, and the targeting possibilities become truly transformative.

It’s the very reason why took such a unique approach to programmatic targeting, building from the ground up a platform that can consume and act on this vast amount of unstructured data. The benefits to this are numerous, but here are a few:

Deeper insight into dynamic audiences.

The platform optimizes audiences in real time — and it does so at the data element level. So if, for instance, ads served within 30 minutes of visiting a brand’s website have a higher click-through rate than those served outside of a 30-minute window, the platform can adjust accordingly to ensure the most optimal point for delivery. This same ability to adapt applies to each and every data point, so as our platform learns what’s working and what isn’t, the audience will evolve throughout the life of the campaign. Couple that with our completely transparent and highly granular analytics system, and you’ll receive more nuanced and insightful analysis, which in turn enables brands to continually improve their marketing efforts.

Localized audiences for national campaigns.

Personalization as a marketing strategy is as vital and ubiquitous as ever before. So it stands to reason that a brand’s strategy in, say, Kansas might differ from its strategy in California. And the widespread use of mobile devices and location data is helping marketers stretch their budgets by considering local preferences, nomenclature, and buying habits — helping to better inform their campaign strategies and get the most bang for their marketing buck.

Performance on high volumes of localized campaigns.

Because it’s more flexible than its structured counterpart, unstructured data can deliver not only a more efficient campaign performance, but higher volume to boot. Our automated processes for campaign entry, management, optimization, and reporting enable us to deliver robust performance on a high quantity of localized programmatic campaigns, whether you’re running 30 of them or 30,000.


Every programmatic marketer deserves to know who they’re targeting, what it’ll cost, that the results are real and brand-appropriate, and why the campaign did or didn’t work. Without a reliance on prepackaged audience segments, unstructured data illuminates the data source in ways structured data cannot, providing marketers with a deeper understanding of how, why, and where their ad dollars are being spent and enabling them to target, bid, optimize and report at the data-element level. The promise of one-to-one advertising is fast becoming a reality. But to realize the full potential of programmatic advertising — to reach the precise audience at the ideal moment with maximum efficiency — marketers must leverage the power of unstructured data. Otherwise, 80% of all those zettabytes would tragically lay to waste.