StoreGround floorG03 · Work · Vans
Which trends should Vans chase?
When a competitor's shoe takes off, a brand has to decide fast: answer it with one of its own shoes, or let it pass. I built Web Catcher as a Product & Merchandising intern at Vans in Shanghai in 2024 to help make that call, and rebuilt it in 2026 to run right here. Read how hot a trend is, try it on Vans' three core shoes, and see what backing it would take.
Web Catcher, playable
The trend you're trying on
Read the trend
How hot is it? Growth over the last year counts for half, buzz for 30% and size for 20%, each on the same 0 to 100 scale first. It needs 50 to count as hot.
How long it should last
Try it on our shoes
Rate how close each core shoe is to the trend on six traits, from one star (far off) to five (right on it). Fit is the average, out of 100. The closest shoe is the one Vans would answer the trend with.
Back it, or not
Each lever step moves the ratings it touches by half a point. Nothing changes a shoe's shape quickly, and product work only counts if it lands before the trend ends.
The whole wall
Every trend on one chart. Higher is hotter, further right is closer to one of our shoes, and the corner a trend lands in is its call. Hollow dots show where a trend sat before the levers. Click a dot to try that trend on.
See the numbers
Every number here is illustrative. The signals, the ratings and the calls are mine, set to show how the tool works, and none of them is Vans data or anyone's real sales. Your setup stays in this tab while you look around the store, and nothing leaves your browser.
How the call is made
Two scores, each out of 100, and a line at 50.
- Momentum is how hot the trend is. Growth over the last year counts for half, buzz for 30% and size for 20%. Each signal goes onto the same 0 to 100 scale before it's weighted: growth runs from −50% to +50%, buzz and size from 1 to 5.
- Fit is how close a shoe is to the trend: the average of its ratings on six traits (price, function, shape, material, color and pattern), on the same 0 to 100 scale. Of Vans' three core shoes, the Old Skool, the Authentic and the Slip-On, the closest one counts.
| A core shoe is close (fit 50+) | No core shoe is close | |
|---|---|---|
| Hot trend (momentum 50+) | Back it | Test small |
| Not hot | Hold | Pass |
How long the trend lasts decides what backing it means. Under six months, new product can't reach the shelf in time, so the play is marketing and the stock you already have. Six to twelve months leaves time for new colors and materials. Over a year is long enough to develop the shoe.
The three launch levers move the ratings half a point a step. Product moves material, color and pattern (at half strength for a six to twelve month trend, and not at all for a shorter one). Marketing moves function, and money moves price. Nothing moves shape quickly.
What I changed in the 2026 rebuild
Coming back to it two years later, I fixed the gaps in the first version's logic:
- The shoe ratings now decide where each trend sits on the chart. In the original they were scored, but the chart used a fixed fit for every trend.
- Every signal is scored out of 100 before it's weighted. Sales used to go in as raw units and drowned out growth and buzz.
- A hot trend that none of the core shoes fits gets its own call, Test small, instead of the same answer as a dead one.
- The levers move the shoe on the chart instead of only changing the wording, and product work only counts if the trend lasts long enough for it to land.
- A coach line under the call says the cheapest lever move from where you are that would turn it into Back it, or why nothing will.
Where it comes from
I built Web Catcher during my Product & Merchandising internship at Vans in Shanghai in 2024, in Python with Streamlit, Plotly and pandas. The question it answers comes up every time another brand's shoe takes off: is this one ours to chase?
The five trends are the ones I built it with, as they looked in 2024. This page runs the rebuilt Python rules step for step in JavaScript, so both give the same numbers. The code is on GitHub.