Research and design that grew six-month retention from 39% to 62%

The Context

hiku was a connected device and companion mobile app that helped busy families stay on top of grocery shopping. Living on the kitchen counter, it scanned barcodes and recognized voice input to build a shared grocery list on everyone's phone — one-button simple, so the whole household could keep the list current. I worked on hiku from early product development through launch and into post-launch optimization, living with the full arc of a hardware-plus-software product as it met real kitchens.

The product earned a small but fiercely loyal user base; in-kitchen research consistently surfaced genuine emotional attachment to the device. But the business model depended on that loyalty. Long-term engagement was what produced the household consumption data our retail partners valued, which made retention the central design problem to solve.

The Work

At launch, 39% of users were still active after six months — strong for a brand-new product category, but not enough for a model built on sustained engagement. I partnered with engineering to implement Mixpanel, giving us real visibility into behavior and drop-off across the customer journey. Two problems stood out: onboarding, where many users never connected the device to Wi-Fi or used only scanning or voice rather than both; and habit maintenance, where users who formed a genuine habit still drifted away over time.

To fix onboarding, I watched people set up the device and documented friction at nearly every step — not realizing the app had to be downloaded first, getting blocked by unclear instructions or forgotten Wi-Fi passwords, stalling on ambiguous calls to action. Working directly with the CEO and engineering team, I redesigned the Quick Start guide around an unmistakable "download the app" call to action, reimagined the in-app onboarding flow, and embedded video showing exactly how to scan and speak with hiku.

To build the habit, observational research explained why onboarded users still fell off. The most common reason was almost absurdly simple, and fatal: the device ran out of battery and never got recharged. We added clear battery-level indicators and proactive in-app low-battery notifications, optimized the Add Items flow for adding directly in-app, and redesigned the Quantity Picker for quick increments that still respected voice-entered quantities.

The Outcome

hiku was a popular holiday gift, which made post-holiday retention the metric that mattered most. Over two years of incremental, research- and analytics-informed improvements, six-month retention climbed from 39% to 62%. We also pursued distribution through retail partners who connected hiku to their e-commerce platforms—pilots with Peapod, Walmart, HEB, and Chronodrive in France—but those partnerships proved hard to sustain.

hiku is where I learned to see retention as a design goal—pairing behavioral data with in-context research to find the unglamorous, decisive friction (a dead battery, a missed download) and designing it out. Working shoulder-to-shoulder with a founder and a small engineering team to move a single number over two years is exactly the kind of focused, evidence-driven ownership I bring to startup teams now.