iHeartJane: Unwind. Wedding Cake 98 percent match, 44 dollars. Gelato 96 percent match, 48 dollars. Cart 92 dollars.

iHeartJane

Santa Cruz

Unwind

Open the budtender
  1. Wedding Cake$4498%match
  2. Gelato$4896%match

Cart$0

I worked on two problems: an inherited product card hid Add to Cart, and pickup shoppers needed help finding products. After three iterations, menu conversion rose from 7 to 10 percent. Jane AI asks how a shopper wants to feel, then ranks products and carries the choice through to the bag.

Role
Founding Product Designer · Cards, grid, discovery
Year
2017-2020
Duration
3 years
Tools
Sketch · InVision · Mixpanel · Google Analytics

I inherited Card 1.0 from an outside agency. Testing showed that it did not support the menu's main job. The horizontal layout broke the grid, the image was too small to evaluate, brand names and THC were missing, and Add to Cart looked disabled. Pickup shoppers needed a visible action and enough product information to decide.

  • OrientationHorizontal only. Broke the grid. Unresponsive.
  • ImageThumbnail too small to describe the product.
  • Add to CartGhost CTA. Weight and quantity looked disabled.
  • I need to find dispensaries near me, that offer the option for pick up or delivery.

Usability testing · pickup shoppers · Mixpanel on the live menu

  • Card 1.0 · inherited, annotated from testing

Card 2.0 separated browsing from configuration. The rest state showed image, lineage, name, price, and Select. The selected state showed weight, quantity, a live total, and Add to Cart as a primary button. Add to Cart was the second most valuable CTA on the site. Tapping Select opened configuration; adding confirmed the choice; closing returned the shopper to the same place in the grid.

  • RestImage, lineage, name, price, Select.
  • Select modeWeight, quantity, live total, Add to Cart.
  • ReturnClose lands back on the same card in the grid.
  • Card 2.0 · rest and select mode
  • Product card interaction flow

Operators asked for more cards per row. Card 3.0 reduced the image and tightened the stack, moving configuration into a product detail page with details, nearby stores, pickup or delivery, and the newly launched recommendation engine. I ran A/B tests across six breakpoints and tracked launch behavior in Mixpanel.

  • GridTighter card. More SKUs per swipe and per row.
  • PDPDetails, store map, pick-up or delivery.
  • RecommendMatch percent on the row beneath the product.
  • 3.0 grid · recommended row with match percent
  • PDP · stores nearby and the recommendation engine

Launch results

  • 7% → 10%Conversion after Card 3.0
  • 15% → 60%Responsive mobile use
  • +$17.00Average order value

Mixpanel · Google Analytics · Product Card 3.0 · early 2019

Mixpanel · Google Analytics · page views +12% · checkout +8% · blind test vs Leafly and Weedmaps

Guided product discovery

Prototype

Santa CruzThe city on the card gives pickup a location.Jane Discovery · location as a given, not a permission wall

For shoppers in Santa Cruz, the experience starts with the market they are in: 3,567 products from 17 dispensaries. I avoided a permission dialog. Showing the city on the card gives pickup and delivery a clear location.

  • CitySanta Cruz, California. Already found.
  • Catalog3,567 products across 17 shops.
  • JobPickup or delivery starts from here.

Jane asks about activities, how the shopper wants to feel, and how they prefer to consume. It is one short interview. Wind Down, Relaxed, and Smoke are inputs; the menu stays out of the way. Research showed that most shoppers found product through friends, not a counter.

  • I need to find an online tool that shows product descriptions, effects, flavors and other info about the products.
  • Target: pickup shoppers who know a little, not beginners
ActivitiesThe first question asks what the shopper plans to do.
FeelThe next question asks how the shopper wants to feel.
PreferConsumption preference filters the results. Smoke keeps flower at the top.

Jane ranks the bag: Wedding Cake at 98 percent and Gelato at 96. The match gives the shopper a reason to open the card. The product page identifies Flower to the People in Santa Cruz and carries forward the store context from Card 3.0.

  • Rank98% Wedding Cake. 96% Gelato.
  • TapCompatibility opens the product, not a ghost CTA.
  • WhereFlower to the People · Santa Cruz.

Open Jane AI budtender

Santa CruzThe city on the card gives pickup a location.
ActivitiesThe first question asks what the shopper plans to do.
FeelThe next question asks how the shopper wants to feel.
PreferConsumption preference filters the results. Smoke keeps flower at the top.
RankedThe match percentage gives the shopper a reason to open the card.
ProductThe product page carries Card 3.0's store context into Santa Cruz.
CartThe cart confirms pickup or delivery and the total.
FT-092Order FT-092 uses studio ID themfa.

Close

  1. Making cannabis shopping easier for new customers

    Instead of starting with a menu, Jane AI budtender asks new shoppers three questions: what are you doing, how do you want to feel, and how do you prefer to consume?

    The answers lead to ranked products, local availability, and a clear pickup or delivery choice, so new shoppers can start without knowing strain names or product jargon.