05 · CASE STUDY · UX DESIGN INSTITUTE
A plant shop that tells you what you can actually get, before you fall for a plant you can't have.
Shoppers fall for a plant, then learn at the last step they can't have it.
Buying plants online often ends in the same disappointment: you find the perfect plant, and only at checkout discover it can't be delivered, is out of stock, or can only be picked up in store. We tested three Swedish plant shops, interviewed ten gardeners and designed a shop where stock and delivery status sit on every product card, so users know before they click, not after.
This started as my own problem.
I have a garden, and every spring and summer I buy new plants online. Again and again I'd find the right plant, only to learn at checkout that it couldn't be delivered, was out of season, or could only be picked up in store. Together with fellow UX student Anneli Arn, I set out to find out whether this was just my frustration or a real problem in the market. It turned out to be both.
Availability mattered as much as price.
We ran structured interviews with 10 home gardeners. Six of ten were 30 to 40 years old, and six of ten buy plants online at least once a month, which makes the stock problem a recurring frustration rather than a single bad experience.
“On one of the sites it was really hard. You couldn't see from the start what could be added to the cart. Later I saw there was some small text, but that was after I'd been on the site for five minutes.”
INTERVIEWEE, HOME GARDENER
The empathy map summed up what users expect: to see clearly whether a product is available, how long delivery takes, and whether they need to pick it up.
Two of three shops hid availability until after the click.
In September 2024 we ran usability tests on three Swedish plant shops, Blomsterlandet, Plantagen and Florea, using the same task on each: buy a lavender. We synthesised the notes into affinity diagrams per site, then followed up with an expert review of each shop.
Plantagen showed availability only after the click: a normal looking product page ending in a disabled button and two red crosses. Blomsterlandet signalled it only through the button: a plus for products you could order, an arrow for those you couldn't. Florea was the exception, with stock badges on every card. It was also the only shop where our participants had no availability problems.
Plantagen: the cards show nothing about delivery, and the product page ends in a disabled button and two red crosses. Screenshots from 2024.
| SHOP | WORKED | PROBLEMS |
|---|---|---|
| Plantagen | Member prices, care info icons, pot suggestions | Availability revealed at a dead end; chat covers content |
| Blomsterlandet | Calm layout, clear wayfinding, inline quantity stepper | Availability signalled only by button icon; help leads to a forum |
| Florea | Stock badges on cards, quick buy stepper, editorial imagery | Chat covered product names |
In 2024 we called the project done when the flow felt right. Looking back, that isn't a measure. So we defined success in three ways:
Shoppers can tell whether a plant can be delivered without opening it.
They know roughly when it will arrive.
They feel confident enough to check out.
Every decision traces back to something we saw in research.
Finding: users couldn't tell what was available until they opened a product, sometimes not until checkout. Decision: every product card shows delivery status up front, and the footer repeats the delivery promise on every screen.
Before: Plantagen
Before: Blomsterlandet
After: FLOWER
A participant asked for “simpler symbols to make it clearer”. We tried an icon on its own, but a mailbox can just as easily mean messages. The final card pairs the icon with words: “Available for home delivery”.
The plus icon worked at all three shops, but at Florea users didn't understand how to add more than one. Inspired by Blomsterlandet, our plus expands into an inline quantity stepper on the card.
We went back and forth on where the chat should sit. On the right, it would compete with prices, checkout buttons and calls to action, which usually live there. Florea had theirs on the left, but it covered product names. Our solution: a round “Hello” button on the left that floats over images, but never over text. When you scroll, it jumps to the next free space, so it never hides information. Tapping it opens the chat with an invitation, “Ask us”, like a shop assistant who greets you first and then asks how they can help.
Early iteration: right edge
Final: left, over the image
From the right edge to the left, where the button never covers text.
Two ideas we designed, then cut.
We tried giving each category an image, but one picture can't represent a whole group. One rose for “Roses” misleads anyone looking for another kind. We chose categories in plain text. Images work for narrow categories; for broad ones they promise something the shop may not have.
The mailbox icon on its own looked clean but was ambiguous. Cut in favour of an icon with text.
This was not a smooth project, and the gaps are worth naming.
We ran the project after work, in our spare time. Recruiting participants was the hardest part: we got enough to move forward, but a larger group would have made our findings stronger. Our participants were friends, family and colleagues, all with their own gardens and a real interest in plants. That made them relevant users, but also people who knew us, which may have made their feedback kinder.
We tested the shops on desktop but designed for mobile, using Florea's app as a guide for how features should behave on a small screen. We only looked at three Swedish shops, and the research is from 2024, so the shops may have changed. We also agreed on almost everything. Looking back, that was a weakness: we were so convinced by the idea that we rarely challenged it.
We cut the pickup flow to keep the scope realistic, and a searchable delivery filter didn't make it either. Every product in our prototype was deliverable, so we never designed the harder states: out of stock and pickup only. And we chose not to repeat availability on the product page, since users had already seen it twice. The risk: users who land directly on a product page, from a search or a shared link, miss it.
We haven't tested the prototype yet. Here is exactly how we will.
Five gardeners, recruited outside our own circle this time, testing on a phone. Each gets the same three tasks on an existing shop and on our prototype, in alternating order: find a plant you can get delivered to your home, find out when it would arrive, and add it to the cart. We measure whether they succeed, how long it takes, and how confident they feel on a scale of 1 to 5.
| SUCCESS CRITERION | HOW WE TEST IT | TARGET |
|---|---|---|
| Knows if a plant can be delivered without opening it | Task 1 | 4 of 5 succeed |
| Knows roughly when it will arrive | Task 2 | 4 of 5 answer correctly |
| Feels confident to check out | Confidence rating | Higher than on the existing shop |
Fewer abandoned carts, because shoppers no longer discover a problem at the last step. More completed orders. Fewer customer service questions about when a plant can be delivered or comes back in stock. And by tracking how often shoppers look at plants that are out of stock, the shop learns what customers actually want, and can restock to match demand.
A plant marked “available” that isn't is worse than no label at all. It breaks the exact promise the design makes. The stock data has to be updated often, and when it can't be read, the interface should say so honestly: “We can't confirm delivery right now. Save it to your favourites and we'll let you know.” A final stock check before payment catches anything that changed along the way.
Some shops hide stock problems, hoping customers will buy something anyway. Our research pointed the other way: when availability showed up too late, participants didn't buy something else, they stopped trusting the shop. Being honest early, combined with a good alternative, “Similar plants we can deliver” or “Tell me when it's back”, keeps the customer instead of losing them at checkout.
The design was the easy part. The data behind it is the real work.
The shops already had the information. Plantagen even knows your store and postal code. It just wasn't visible where the decision was made.
The same principle shaped the product card, the decision to cut category images, and the fallback when stock can't be read.
Agreeing easily felt good, but next time I'd build in a moment to challenge our solution before falling in love with it.