Flag

We stand with Ukraine and our team members from Ukraine. Here are ways you can help

Get exclusive access to thought-provoking articles, bonus podcast content, and cutting-edge whitepapers. Become a member of the UX Magazine community today!

Home ›› Design ›› Using “Dumb Data” To Make Smart Design Decisions

Using “Dumb Data” To Make Smart Design Decisions

by Jocelyn Lin
3 min read
Share this post on
Tweet
Share
Post
Share
Email
Print

Save

Designers can streamline their workflow and improve their products by getting down and dirty with some rough-and-tumble research methods.

As an industry, we’ve worked to established many new practices and tools for nimble design teams, from A/B testing to measuring bounce rates and CTR performance. But a lot of these methods require engineers or some amount of technical know-how to execute, and they take place only after something has been launched.

The judicious application of “dumb data” can streamline your workflow and improve your designs

What many people don’t know is that there are some unexpected applications of data to consider earlier in the design process, which you, the designer, can do yourself. They’re not fancy, and you don’t need to know how to write SQL queries. The judicious application of just-enough “dumb data” can streamline your workflow and improve your designs in surprisingly useful ways.

Here are a few examples of how to do so. (Warning: non-statistically-significant numbers ahead.)

Understanding User Behavior

Let’s say you want to understand what types of content people are interested in. I lead design efforts at Polyvore, a site people use to get style and outfit inspiration, but what is their underlying motivation? Do they need outfit ideas for everyday work outfits? Special pastel goth outings? Dream brunch dates with One Direction?

Like any other product with a search box, we have a wealth of outfit queries available. Looking only at the most popular queries loses out on the long tail, while building a fancy machine-learning bayesian thingamabob is too expensive. With dumb data to the rescue, you easily can do this using brute force and Google Spreadsheets.

In your spreadsheet, label a few hundred randomly pulled queries, farming some of this work to your friendly teammates. You’ll quickly get a sense of what the major categories are, and from there, you have a better understanding of what people want to do.

Search spreadsheet

This can be applied to all sorts of content at scale that computers can’t parse. I’ve used this type of data bucket diary study activities, brand sentiment, video search intents, and more.

Beefing Up Experience Maps

When experience mapping, it’s all too easy to make assumptions or get into wishful thinking mode about how users interact with our products. A quick check into click rates from page to page will keep the team honest (User Flows in Google Analytics). At Polyvore, we like to pair the map artifact with a flow diagram that visually shows how likely it is that users actually make it into the next step of the intended experience.

experience map

Starting Scenarios and Storyboards

In addition to collaborating with user research, you can sleuth through referrer and landing page data (again in Google Analytics, along with second page parameters) to identify why people come to your site. Using a hypothetical hotel site example, if most people visit from web search, they probably want to learn more about the accommodations. On the other hand, a visitor from Travelocity has already seen photos, read reviews, and is probably ready to start booking. These deductions give you a better starting point and more confidence in creating scenarios and storyboards.

Creating Realistic and Engaging Designs

Wireframes are an essential concepting tool for the interaction designer, but they can hide layout problems if used for too long. Product teams also tend to have a hard time visualizing the end product, potentially leading to major revisions late in the game. I’ve found that it’s best to move onto using real text and images as quickly as possible, but it’s important to choose the right sort of content to illustrate your ideas.

Take 15-20 minutes to collect a set of realistic images and metadata by running searches or looking at existing page content. So, that means no “lorem ipsum text” (not the right feel, doesn’t fill out space correctly) and no cherry-picked imagery (obscures visual design issues and besides, that’s just cheating!).

wireframe stages

L: clean white look with perfect data M: realistic data R: visual container for realistic data

These are just a few scrappy ways that you can use easily available data to improve your design process. Having been lucky enough to work at tech companies that value analytics, these are the methods I’ve figured out over time. What are some of your tips on incorporating data? Feel free to share in the comments below.

Image of young archer courtesy Shutterstock.

post authorJocelyn Lin

Jocelyn Lin
Jocelyn Lin is Polyvore’s lead UX designer, where she manages a team of designers to create engaging experiences for Polyvore’s global style community. Prior to joining Polyvore, Jocelyn was a staff UX designer at Google, where she led design initiatives for Google Maps, Image Search, and Universal Search. Prior to working at Google, she was a product manager for Oracle. Jocelyn received her Masters in Computer Science from MIT.

Tweet
Share
Post
Share
Email
Print

Related Articles

AI is changing the way we design — turning ideas into working prototypes in minutes and blurring the line between designer and developer. What happens when anyone can build?

Article by Jacquelyn Halpern
The Future of Product Design in an AI-Driven World
  • The article shows how AI tools let designers build working prototypes quickly just by using natural language.
  • It explains how AI helps designers take on more technical roles, even without strong coding skills.
  • The piece imagines a future where anyone with an idea can create and test products easily, speeding up innovation for everyone.
Share:The Future of Product Design in an AI-Driven World
4 min read

Why does Google’s Gemini promise to improve, but never truly change? This article uncovers the hidden design flaw behind AI’s hollow reassurances and the risks it poses to trust, time, and ethics.

Article by Bernard Fitzgerald
Why Gemini’s Reassurances Fail Users
  • The article reveals how Google’s Gemini models give false reassurances of self-correction without real improvement.
  • It shows that this flaw is systemic, designed to prioritize sounding helpful over factual accuracy.
  • The piece warns that such misleading behavior risks user trust, wastes time, and raises serious ethical concerns.
Share:Why Gemini’s Reassurances Fail Users
6 min read

AI is raising the bar for everyone, but what happens when the space to learn, fail, and grow quietly disappears?

Article by Thasya Ingriany
Everyone’s a 10x Employee now. But at What Cost?
  • The article demonstrates how AI-driven tools are raising expectations, prompting even junior roles to demand senior-level judgment.
  • It warns that automation is erasing early-career learning opportunities once crucial for developing design intuition.
  • The piece argues that while AI boosts output, it can’t replace the slow, human process of building creative judgment.
Share:Everyone’s a 10x Employee now. But at What Cost?
6 min read

Join the UX Magazine community!

Stay informed with exclusive content on the intersection of UX, AI agents, and agentic automation—essential reading for future-focused professionals.

Hello!

You're officially a member of the UX Magazine Community.
We're excited to have you with us!

Thank you!

To begin viewing member content, please verify your email.

Tell us about you. Enroll in the course.

    This website uses cookies to ensure you get the best experience on our website. Check our privacy policy and