How Data-Informed Design can help you navigate these turbulent times
Adding “Just enough data” to your design process can have a big impact
It's been a rough year for Designers, not only because of layoffs but also because of looming fears for the future of Design.
Whether these fears are about AI replacing UX or UX alone is no longer enough, there has been much doom and gloom about UX's future.
While I don't know your situation, I want to introduce you to a niche that's helped me survive rough times and saved my Design career twice: Data-Informed Design.
Instead, it's a lightweight, "Just Enough Data" approach championed by people like Julie Zhuo (Former VP of Design at Facebook). This approach can be used everywhere (especially Low UX/Design maturity environments) to help Design have a more significant impact.
What is Data-Informed Design, and why is it useful?
One of the best descriptions of Data-Informed Design comes from the Fountain Institute:
Their description of "Customer data is used to evaluate design decisions" provides one of the field's central concepts: defining and evaluating success.
Data-informed design can be explained by asking three specific questions:
How do you know you've solved a user's problem six months later?
How do you demonstrate the value of UX to Business?
How can you persuade people to act on your design recommendations?
How does a Design solve a problem?
Design is not art. Design does not exist for design’s sake. Design is about problem solving...I think designers have a responsibility to know whether or not they’ve actually solved a problem. — Jon Wiley, Director of Immersive Design at Google
Designers often like to call themselves problem-solvers, but there's one core question that many designers fail to address:
How do you know that you've solved a user's problem?
The problem, as it turns out, is that many designers can't tell you what happens after they turn over their designs. Engineers begin working on them, and Designers move on to the next thing that they're assigned.
While that's great for building features, without knowing what happens afterward, you have no idea if you've succeeded in addressing your user's problems.
Fortunately, a quick way to determine whether you've succeeded is to check the data. After all, data like Analytics is just a compilation of "What users do" on a large scale.
Executives use data to monitor their company's health and success, and with some basic knowledge, you can check your design's impact after it goes public.
What's more, you don't need an MBA or a Data Analyst certificate to do so.
How do you define the value UX brings?
This is one of those eternal questions about "Return on Investment (ROI)" everyone always brings up.
People often advocate learning about specific case studies that show how a better User Experience results in more money saved (or earned). However, advocating for design by talking about how it helped some companies in the past has limits.
Instead, we should focus on what UX can do for your business.
In the case of Data-Informed Design, this doesn't require learning Business Metrics or KPIs (although it helps); it involves learning Product Metrics.
If the business goal is to gain 10,000 paying customers in Q3, showing how UX had a direct impact is challenging. Improving the UX might be the reason for getting a paying customer, but it could also be linked to Marketing or Sales.
However, Product Metrics are more easily affected by UX and correlate with UX. For example, if we showed that the average time spent completing onboarding, a Product Metric, went down from 6 minutes to 2 minutes, we can demonstrate UX's value to not only Product teams but also the business.
Forming this link is how you persuade people to act.
How do you persuade others of your design recommendations?
Sometimes, you might have valuable user findings, but they may be hard to justify.
For example, you conduct user research and find that 2 out of 5 users need help finding certain critical features in the navigation. However, that doesn't tell the whole story: they struggled for 20 minutes before ultimately giving up, which hints at a more significant problem.
The problem is that 2 out of 5 users don't sound like much. If it were 5 out of 5 users, we could make the case that this is a frequent and common issue.
However, 2 out of 5 users is too small a sample size to claim that "40% of users are having trouble with navigation," and nothing else (by itself) might indicate that this might be a high priority.
But what if you had some additional data to support this finding (and persuade your team)?
Here's how you might do that.
Using Data-Informed Design work at a high level
At a high level, the data-informed design process revolves around three core concepts: creating a hypothesis, thinking indirectly, and generating queries.
Start with a hypothesis draft
At a high level, there are two hypotheses that you can use to either talk about current or past work. If you're working on a project currently and talking about user testing results, for example, you would use this hypothesis, which comes from the book Designing with Data:
"If we [do X], Our users will [do Y] for [Z reasons], which will impact [Metric A]."
On the other hand, if you're summarizing your work in a design portfolio, you would use a slightly different hypothesis, recommended by Google Recruiters:
"I Accomplished [X], as measured by [Y], by doing [Z]."
Both hypotheses clarify your value and contributions to a larger goal. When you start, though, you might not have all the answers. However, just putting together what you do know is the first step.
Here's how it might look for our example:
User testing recommendation: If we [Spend time re-doing our Information Architecture], our users will [do Y] because [our navigation is less confusing], which will impact [Metric A]
Design portfolio summary: I accomplished [X], as measured by [Y], by [re-doing our overall Information Architecture]
If you're unsure how to fill in some of these blanks, that's the next step.
Think indirectly about data to figure out your gaps in knowledge.
After you draft your hypothesis, consider what you don't know and where you might find that data before you look at a data source.
This is known as “Thinking indirectly about data,” and it’s one of the most important lessons about the subject.
For example, in our user testing example, we have our design recommendation (re-doing the Information Architecture) and our reasons (Navigation is less confusing), which we’ve learned from talking with users.
What we need to figure out are a few things:
What are users currently doing, and is it problematic?
What are users likely to do due to this change, and why?
What metrics are likely impacted by this as a result?
We must examine a large repository of "What users do" (i.e., Analytics) to answer the first question. We've seen users struggle in our user test, but are there signs of similar struggles on our website?
We must also know enough about user behavior, usually from user testing, best practices, and additional research, to know what changes our design will likely result in.
Lastly, we need to determine the metrics impacted by this. But before deep-diving into Product Metrics, there’s an easy way to do this: ask your Product team. People have probably spent weeks or months determining the important metrics for a project.
Simply asking them will provide you with the answers you need.
Form your “Queries”, or what you want to ask of Data Sources
There’s a lot of data out there, and it can be intimidating to try and figure out where to turn for data. However, in my experience, I’ve turned to 5 basic data sources to find most of my answers:
Past user research/competitive research
Talking with my team
(Biased) User Reviews/Customer Support Tickets
Existing Customer Surveys/New (1–question) Surveys
Analytics
In our example, there are a few questions (i.e., “Queries”) I want to figure out:
How many users navigate to our product page, and are the problems I see in user testing reflected in the data?
What metrics are this impacting, and how important are they to my team?
For the first question, the easiest place to turn is to Analytics: we can look for web traffic to different website pages and see if there’s a big drop-off getting to the page we care about.
For the second question, we’re likely talking with our team to determine whether this impacts the metrics they care about.
After figuring out these queries, we might end up with something that looks like this:
User testing recommendation: If we [Spend time re-doing our Information Architecture], our users will [be able to find our new product] because [our navigation is less confusing], which will impact [Product Adoption rate]
Design portfolio summary: I accomplished [increasing user adoption by 10%], as measured by [Product Adoption rate], by [re-doing our overall Information Architecture]
By combining data and design, we’ve translated our design recommendations into something that is more understandable and easier for the team to implement.
This skill has allowed me to flourish in otherwise trying times as a designer.
The future of Design may be Data-Informed Design
I know there’s a lot of uncertainty around the field of Design right now. In addition to the poor job market, there are many questions about AI and data and how these things may affect the craft.
This is why Data-Informed Design is a critical skill that Designers will need in the future. When you can craft great user experiences and measure and evaluate what’s going on, you speak to the uncertainty that many businesses may have around design recommendations' impact.
Learning just enough about data allows you to understand the business perspective and what they care about, translating what you’ve found from user research into something they can more easily understand and act on.
So, if you want to future-proof your Design career, consider learning Data-Informed Design.
My Maven course on Data-Informed Design is coming soon! Join the waitlist for updates
Kai Wong is a Senior Product Designer and creator of the Data and Design newsletter. His book, Data-Informed UX Design, provides 21 small changes you can make to your design process to leverage the power of data and design.