Shipping a store in twelve weeks

Arden came to us with a store that took six seconds to load on a phone and a launch date twelve weeks away. We could not pause sales, and we could not move the launch.

Weeks one to three: decide what not to build

We listed every feature of the old store and asked the team which ones customers used. Nine of twenty-three made the cut. The rest were removed, not rebuilt.

Weeks four to nine: build in the open

The new front end ran alongside the old store from week four, on a private address the team used for their own orders. Real orders found real bugs faster than any test plan.

Weeks ten to twelve: switch in stages

We moved one product category at a time, starting with the smallest. The last switch was the home page, on a Tuesday morning, when traffic is lowest.

Pages now load in under a second. Conversion went up by a fifth in the first month, which Arden puts down mostly to the ingredient lists being visible again.