Growth Conversion Optimization Replatform

Ollie Onboarding

Increasing acquisition funnel conversion by 10% while supporting a FE and BE replatform

Ollie mobile onboarding funnel overview from landing page to plan selection — What started as a front-end and back-end replatform became an opportunity to rethink the entire mobile funnel, from landing page to plan selection.

Ollie onboarding intro with email capture and age versus birthday screens — Email capture was the single largest drop-off point in the funnel, so I combined it with the intro step to get visitors into the experience faster, and shifted repetitive savings banners to targeted placements at key decision points instead. I also proposed moving from age to birth date—with flexible inputs for unknown histories—to unlock future lifecycle marketing moments.

To optimize the onboarding tempo, I consolidated related questions to reduce perceived length and replaced the misleading "three-step" progress bar with a granular quiz progress indicator. This shift, paired with clear navigational cues like "Continue to Checkout," provided users with a more accurate sense of momentum and where they were in the funnel.

Ollie onboarding allergy selection before and after with multi-select chips — To modernize the allergy selection and skip logic, I replaced open-ended inputs with a "top-tier" selection model informed by clinical and support data, ensuring common allergens were accessible while maintaining a global search for edge cases. I further reduced cognitive load by implementing dynamic, context-aware buttons that clearly signaled when questions could be skipped, clarifying the path for users with non-applicable criteria.

Ollie onboarding body type question removed and activity level question clarified — Two questions revealed the same problem: self-reported data doesn't always match reality. Partnering with Strategy, I found a gap between AKC's breed-activity standards and how customers described their dogs, and clarified the question to improve meal accuracy. I also removed the body-type question entirely — it wasn't used in calorie calculations, and pet owners are known to misjudge their dog's weight, making the data unreliable to begin with.

Ollie design system chip component anatomy, variations, and states — The FE manager and I had been in discussion about creating a design system and had a FE intern had been adding foundational styles I had defined to Storybook. The replatform of the onboarding funnel provided the perfect reason to build out additional components that could be used throughout the experience.

The new onboarding funnel combined steps, especially at the top of the funnel, shortening the overall length of the quiz. It also made better use of Ollie's brand, inserting images into certain questions.

Ollie onboarding activity through meal plan selection screens — The new onboarding flow also featured an updated UI and more sophisticated interactions, designed to move visitors from initiating onboarding through to plan selection with less friction at each step.

Ollie onboarding user research feedback alongside order summary screen — While overall qualitative feedback was very positive, there were some areas highlighted in user research for the team to consider post launch for a possible future iteration.

My role

Led evaluative user research to uncover pain points with the current funnel and to validate the new proposed experience.
Consulted with Strategy to define conversion metrics and high-impact opportunity areas to improve within the funnel.
I spearheaded the creation of Ollie's first design system components to ensure the new funnel was both high-performing and scalable.
Infused Ollie's brand refresh into the acquisition funnel through updated copy and images.

The team

Myself as Senior Product Designer
Product Manager
4 FE Engineers
Brand Designer
Content Designer
Consulted with Strategy and CX

The situation

Ollie, a D2C subscription dog food company, had a plan to tackle FE and BE tech debt beginning at the acquisition funnel. Having already lifted homepage engagement in an earlier project, I followed the data downstream to the next-biggest opportunity: a three-week window to refresh the onboarding quiz. Once visitors started the onboarding quiz, the largest area for drop off was within the first three questions. Only 47% of users who initiated onboarding made it to step 4 of the onboarding quiz, but after step 4 drop off was only 1-2% at each step — overall, just 57.5% of users who started onboarding made it through to plan selection. My goal was not only to support the replatform, but increase conversion as well.

The results

After a week of going live conversion from onboarding to the next step (meal plan selection) increased by 10%.
Reduced onboarding flow from 12 steps to 8 steps, getting visitors through the funnel faster, without losing any critical information.
The perceived amount of questions asked reduced from 8.3 to 7.6 questions.
9 new common components were added to Ollie's first ever design system including a primary button, chip, input, select, progress indicator, and banner.

Key decisions & trade-offs

Invest in usability testing tool: In the couple weeks prior to this initiative, I had advocated and gotten buy-in from leadership for a usability testing software. I was able to use this project to showcase how valuable user research is to uncover pain points and be confident in our decisions.
Create a design system: The FE manager was an amazing partner and advocate of the design system. I suggested a list of 9 components based off of the experience created for the onboarding funnel.
Switch from age to birthday: I proposed shifting from asking for a dog's age to birth date and presented the idea to the Marketing Director who loved the idea of this future, personalized marketing moment. We knew from research that many pet parents do not keep track or know their dog's birthday so there was a risk we might introduce more difficulty to move past this step. We decided to take the risk and I tried to make it as easy as possible to enter the birth date by not requiring the specific day.