I led the redesign of Weedmaps' core search experience, replacing a redundant two-tab results grid with a new map-based paradigm — letting shoppers pick a product and see, in real time, which nearby dispensaries actually have it in stock. It was my call to build this against real internal resistance, including from the ML search team, whose default instinct was to just make the existing results more accurate. I sequenced the rollout desktop-first for four deliberate reasons, partnered closely with lead designer Shawn Yoard (whom I'd recently promoted onto the project as a growth opportunity), and shipped it alongside an AI-powered chatbot and a full navigation overhaul — work that earned same-day CEO sign-off to overhaul the entire desktop experience.

Understanding the real job. Every serious Weedmaps visitor is there to find a product to buy — search queries split between broad category shortcuts ("pre-rolls," "flower") and highly specific asks ("multipack infused pre-rolls," "indica-leaning hybrid premium flower"). Applying a jobs-to-be-done lens to the full user journey, I found the real decision driver wasn't just the product — it was the retailer. For pickup, proximity to where the customer actually is matters most; for delivery, speed matters most. Weedmaps had visibility into every dispensary's real-time inventory and simply wasn't using it effectively.

The problem with what existed. Search results lived in two nearly-identical tabs, "All" and "Products" — the only real difference was that "All" also matched on dispensary name. Neither surfaced where a product was actually available. The opportunity was obvious once I saw it: put the map back into Weedmaps.

The decision — and the resistance to it. I worked with my team to present a new paradigm: select a product, see a map of every dispensary that actually has it in stock, right now, toggle-able between pickup and delivery. There was no real precedent for this — something like shopping through Google Maps crossed with DoorDash, but not quite either. This wasn't a consensus decision. The team, including the ML search team building it, pushed back — their natural instinct as engineers was the more efficient path: keep the existing pattern and just make the matching more accurate. I saw a much bigger opportunity to differentiate Weedmaps and meaningfully increase its usefulness to real shoppers, and I pushed the team past that instinct.

Sequencing it deliberately. Even though we're a mobile-first company — 80%+ of our activity happens on native app and mobile web — the cross-functional team chose to start the redesign on desktop, for four specific reasons: desktop, counterintuitively, showed proportionally more search activity, so we'd get more signal per iteration; it was our worst, most neglected surface, so there was more room to improve; it was our lowest-traffic surface, so the blast radius of any misstep was contained; and our CEO reviews the team's work primarily from his desktop, so shipping visible progress there kept momentum and visibility on the initiative. That decision made execution harder, not easier — I had my team designing for desktop while also thinking ahead to how the pattern would morph for smaller mobile form factors without breaking our shared component system.

Growing my team while building the product. I partnered directly with Shawn Yourd, a Lead Designer I'd recently promoted and moved over from our B2B design team specifically to give her a highly visible, highly complex, consumer project to grow into. (It was a deliberate development decision, not just a staffing one. She hadn’t worked on anything this nuanced or so much in the leadership spotlight before.) Shawn is especially strong at “jobs-to-be-done” thinking and structuring rigorous design-thinking foundations, which is exactly what this bet needed behind it. My own strength is seeing the larger opportunity — the forest — and I work with team members like Shawn to build the path through it.

Shipping it. We combined existing dispensary data — amenities, real-time inventory, full menus — with an integrated Mapbox experience to rebuild the search flow end to end. Once the core map-based product results were validated, we expanded the results experience further with intentional, distinct tabs for Brands and Strains (previously conflated with product results) and a much stronger search-input experience with live type-ahead suggestions.

The moment it became a bigger initiative. Within about 30 days, this search work came together alongside two other projects I was leading in parallel: an AI-powered "Budbot" virtual budtender chatbot, and a desktop navigation and information-architecture overhaul. When I presented all of it together, our CTO was excited enough to bring it straight to our CEO the same day — his response was "how can I have it all," and he authorized the additional complete overhaul of desktop navigation and design, which then had spillover effects into our app and mobile web experiences.

The results, so far. Our retained customers already converted at an unusually strong rate — roughly 65% — and new customers around 25%, both already ahead of typical marketplace benchmarks. This search improvement lifted conversion for both groups by more than 10%.