10 insights into user-centered growth design

Illustration by Fanny Luor

At Dropbox, we have a healthy and lean approach to growth design. This is largely the result of a finely-tuned process our growth team has developed over the last couple of years. A common thread throughout is a respect for user-centered design and our belief in the value it adds. Here are the 10 most valuable growth design practices I’ve picked up during my time at Dropbox.


1. Get involved early.

Attend or organize brainstorms with your growth cross-functional partners. This is a great time and place to build on others’ ideas and suggest ideas of your own. But more importantly, it’s a chance to be an advocate for your end users when ideas are still developing. Since designers often work closely with research and customer support teams, we’re uniquely positioned to recall and leverage those insights for the benefit of the user.

2. Get a bird’s-eye view.

Growth experiments often focus on a small set of variables for the sake of clean learnings. When your solution is focused within a single surface, make sure you’re designing with the whole user journey in mind. It also helps to consider how that change compares with other parts of your product or website, to help maintain a uniform experience for your users.

3. Clean as you cook.

When you experiment on old or outdated surfaces, try to bring them in line with the rest of your product or site. A visual refresh is a common way to do this. If your improved version does as well as (or better than) the control, then you can ship it as the new control. If you have a tough time persuading your team to agree to a refresh, I find that it helps to share a mock of the refreshed page.

A visual refresh mock for the Buy page (from 2016)

To give an example, long before our rebrand, our heavily optimized Buy page was looking visually dated and didn’t match newer pages of our site. It had our older-styled buttons, form fields, and menus. Even the illustrations were out of date. When a break between experiments finally opened up, I suggested refreshing the design of the page. To sweeten the pitch, I showed not just a refreshed design, but a few experiment ideas we could run on a refreshed surface. The team rallied around it, and we successfully shipped the new design.

4. Polish your wins.

It can be a challenge to polish every experiment you work on, especially if they have short timelines. One way to deal with this is to align early with your growth or product manager on the constraints and in understanding the initial objective. Usually, it is to get a clear learning, which can inform the solution you eventually arrive at. Once you have learned what you need to, reel it back in, give it the necessary visual and micro-interaction polish, and ship it as the new standard.

5. Don’t experiment on atoms.

In Atomic Design, atoms or primitives are the smallest parts of a design system. Common examples include colors, spacings, border radii, typeface, and typescale. Even small changes at this level can be very disruptive to a design system. You should avoid exposing the foundation layer of a design system to growth experimentation, unless you have good reason to believe that a primitive is somehow at fault. Fixating on atomic level experiments can also distract you from thinking about and solving real problems for real people.

6. Focus on impact.

Your time is valuable. Try to give preference to projects that have the greatest impact. There’s always more growth work than your team has resource to handle. Make sure you understand the potential opportunity and impact of an experiment so you can weigh that project against others. Work closely with your PM to prioritize work that is impactful for the business and the end user.

7. Understand why.

Work with your growth or product manager to understand, and if needed, challenge the experiment hypothesis. I find it helpful to write out the user and business problems in my own words before I begin. With a clear understanding of the problem, you can generate better ideas to tackle it. You’ll also be a better judge of how well a given solution will work.

Take, for example, a recent experiment. We wanted to improve a tool people use to look up charges to their card from Dropbox. This tool got less use than we expected, and people often chose to reach out to customer support for help instead. Our first thought was that the tool itself might be confusing or broken. But by speaking to customer support and reading through actual CX tickets, my PM uncovered the surprising root cause.

Card-owner authentication for the charge lookup tool

When an employee leaves their company, they lose access to that company’s Dropbox account and its content. This is as you’d expect, and as it should be. As it turns out, however, individuals sometimes use their personal credit card for their employer’s business account. If that card remains on the business account after they leave, then it will be charged the next time the subscription renews. When the card owner discovers the charge on their statement, they see the web address of the lookup tool: db.tt/cchelp. Unfortunately, using this tool requires them to sign in to the affected Dropbox account, which they no longer can.

By understanding why the tool was under-used, we could address the real problem. We redesigned the authentication process for the charge lookup tool. Now card owners could access the tool even if they had lost access to the Dropbox account. We also added the ability to remove that card as the payment method for the account.

8. Incorporate winning ideas into your design system.

As you and your team create and improve components, keep tabs on how they perform after they ship. If they do well, fold them into your design system, and try to document how best to use them. When it’s easy for others to discover and use the best patterns, it helps your team stay aligned and always moving forward. In turn, this promotes consistency across your product or site for end users.

9. Avoid dark patterns.

Dropbox relies on continued business from our users, so it’s important for us to be worthy of trust. For us as designers, this means never resorting to dark patterns or other dishonest design practices. This value is well-established on my team, and if an idea falls into a gray area, we turn away from it.

Highlighting yearly billing with cost savings and illustrations

As an example, I once worked on an experiment on our Buy page to increase annual subscriptions. Instead of trying to hide or camouflage the monthly billing option, we decided to clearly position it beside annual billing, and to highlight the annual plan’s cost savings, along with supporting illustrations. The experiment was a success — we hit our target without risking our users’ trust.

10. Keep the team in sync.

You should attend team design reviews or critiques (which we call design sessions) as often as possible so you can share your learnings and stay informed on what others are up to. You can also use this time to get feedback on your work, and generate ideas by leading brainstorms. Design crits also help your team avoid solving the same problems repeatedly. At Dropbox, the growth design team meets twice a week, which is a great cadence for our team.


These practices have helped me approach growth objectives with a user-centered focus. By thoughtfully applying them in your day-to-day work, you can create great experiences for your users and move the needle while you’re at it. Do you have growth design practices you’ve refined on your own? Please share! I love talking about this stuff. 🤓


Want more from the Dropbox Design team? Follow our publication, Twitter, and Dribbble. Want to make magic together? Check out our open positions!