Guts vs. Graphs: Design in a Data-driven World

Data-driven design: it’s all the rage. Also, design thinking is all the rage. In fact, design itself is all the rage these days, it seems.

Everywhere you look, businesses are touting the benefits of another design process or championing another design concept, and data-driven design is among the frontrunners in this craze. But is it all that different from the design processes we already have? What makes it so special?

It’s a worthwhile question. After years of being mistreated, misunderstood, and misrepresented, design and all its variant components are finally getting a moment in the sun. Questioning which components we champion — and why — is crucial to ensuring that the momentum of the design moment isn’t lost, and that design gets a permanent seat at the table.

So before we can defend, support, champion, or foster data-driven design in our workplaces, we need a clear understanding of what we’re talking about.


What Is Data-driven Design?

“There are far too many terms, abbreviations and nonsensical jargon in our field these days. It essentially boils down to engaging in some form of user-centric design process and applying any learnings that come from that into the product or experience that is being created.” — Andrew Semuschak, Design Practice Lead, Myplanet

As is often the case with emerging concepts in the digital sphere, data-driven design doesn’t have a clear, agreed upon definition. But with data-driven design, it’s not an industry-only source of confusion. People hear the word “data” and their brains, for better or worse, move to Business with a capital B.

In some respects, this is good. It means when we need to get clients on board with the design process, stakeholders are more receptive. If data is involved, it’s often assumed that designs will be more reliable, offer a better ROI, and generally be a better guarantee of a “sure thing”.

But it’s also not good. Data-driven design is an approach to design, not to data. The expectation of data as the silver bullet is not only unrealistic, it’s detrimental. Design isn’t a cut-and-dry process. Good design — innovative, exciting, engaging design — is messy. It’s not a science, it’s not perfect, and most importantly, it shouldn’t be.

And while none of this gives us a clear answer for what it is, understanding the perception of data-driven design helps us to understand it at least a part of it.

So, for our purposes, we’ll use Remy Rey de Barros’ definition of data-driven design: “[I]n essence it is basing design decisions on actual data (qualitative and quantitative).”


Data-Driven Design Goes Mainstream

Why is this so important now, especially for those of us in enterprise spaces? Partly, it’s down to design thinking. As design thinking gains prominence and moves into more and more corporate spaces, enterprise is taking notice of design more generally. You can’t mention design in the corporate space and not see this valuation index, for instance:


But it’s also partly down to the particular technological moment we’re in: new technology allows us to collect more and more data to be used in decision-making. Big companies, with their broad customer bases and greater resources to draw from, are in the best position to collect that data.

Combine these two factors and it makes sense that a shift to data-driven design would emerge, especially in corporate environments.

But that doesn’t solve for how we should be using data and a data-driven design approach in our work.


Data-driven Over Intuition?

Proponents of data-driven design will tell you that backing up your choices with data is the only sensible way to design. Why wouldn’t you use hard facts to help you decide what to build and how to build it? Why trust a hunch when the data will tell you what’s what? What could be better than taking the data your users provide as gospel and using it in your designs?

That may all sound good, but as Remy Rey de Barros says, “Data can inspire and inform, but it can never drive something innovative, it can make the product great, but more is needed to make it awesome.” That spark of something great — the intangible, je ne sais quoi that makes a design something truly special — can’t be found in data.

Moreover, as analytics guru Avinash Kaushik notes, “All data in aggregate is “crap”.”

“The best-case scenario for data-driven design: using quantitative data to identify issues and to benchmark current performance, and then using real-time qualitative user testing to understand why you’re seeing those numbers and how to improve them.” — Ashley Moreno

Numbers are meaningless until we give them meaning. To make effective use of data in our design process, we need to have a comprehensive understanding of how to conduct, measure, and analyze data sets as well as harness a great deal of knowledge around the application of design theory. And all of that needs to be synthesized into effective testing, validation, and iteration efforts.

“I’d argue there are three critical points for data-driven design to applied,” says Andrew Semuschak, Design Practice Lead at Myplanet. “1 — Getting inputs prior to discovery / starting the project. 2 — Confirming assumptions and early executions in the middle of the define phase. And 3 — Vetting the first full iteration at the end of the first release.“

Andrew’s points may not sound all that different from any other design framework, because testing and validation should always be a part of the process. But data-driven design requires both a rigour and a flexibility in our approach that may not always be required in more generalized design thinking frameworks.

For instance, the use of “big data” (what we’re often relying on when we’re dealing with data-driven design) as compared to the types / amount of data we would normally rely on requires two seemingly opposite things from us: The first, a strict adherence to best practices when it comes to interpretation of the data. The second, an enormous amount of flexibility in the application of that data to our designs.

We need to be open to altering, changing and re-jigging our designs like never before. In many ways, when we look to data-driven design as an approach, we’re designing for data — for its incorporation and influence — and this is a beast of a different burden.

“Designing to blindly satisfy a number almost always leads to a poorer experience, a poorer product, and ultimately the company getting poorer.” — Dan Turner

When a CEO asks why we would spend precious resources on data-driven design, we need to be ready to answer them with sound reasons about why bringing data analysis in at this stage is worthwhile.

And when they ask why we wouldn’t spend those same resources on data-driven design, it becomes even more important to have a firm understanding of when to turn to data and to take the analyses offered as the best path forward, and when to rely on your own expertise as a designer.

“Ultimately, we should be driving towards the “anything I make has to serve a beneficial purpose and be used” mentality.” says Andrew. Which means when we think of data-driven, we need to be embracing an ongoing, iterative use of data collection to impact our design decisions. How will data help me improve the functionality and usability of this product? And how do I know data can answer it better than I can?

“Too often, testing and analysis are one-off activities, providing plenty of important-looking numbers but not lot of context or specific direction,” — Ashley Moreno

Design has never been about just “making things pretty”, but this new era of data is a sizeable shift in how we can approach our work. Striking the balance between what experience, knowledge, and intuition can offer versus what data-informed approaches can do is a delicate balance — and increasingly, part of an essential skill set for designers.


We have open positions— apply to come work with us now.

Written by

Leigh Bryant

Leigh Bryant

Sign up for our newsletter