Etelier Gallery

The store had to behave like the releases do: precise, visual and built around a moment, not an evergreen catalog.

We separated the customer experience from the Shopify theme. The unlock screen, product pages, cart and gallery are designed as one release.

Products, inventory, orders, payments and checkout stay in Shopify.

Drops used to run by hand: someone watching a calendar, unlocking the store and preparing notifications, hoping every step stayed in sync.

Now a drop is configured once and access follows the schedule. The team shares one view of the release before, during and after.

The system defaults closed. If something can't be verified, the store stays shut outside its window.

Three tiers open twenty minutes apart, and the public link opens last.

In August 2026, 72% of orders landed before the public link existed. The third tier’s window carried over half the drop’s revenue in twenty minutes.

Earlier access converts harder: the second tier bought at 12.1% from the same message the third bought from at 1.9%.

Between April and August 2026, drop revenue and orders both more than doubled.

Average order held within $10, so the growth came from more customers, not bigger baskets. Drop-day traffic grew 2.3 times and conversion stayed within a point.

Repeat buyers rose from 18.7% to 28.8% of each drop while the list was still growing. SMS and email carried over half of August’s revenue.

In one measured release window: 2,089 claim events, a 191 ms p95 from request to confirmed claim, and zero oversells.

At the August 2026 peak, 482 people were on the site in a single minute and made 358 claim attempts, with no downtime. One size sold exactly 32 of 32 units and stopped.

Measured under live demand, not in a load test.

The work is ongoing under RILT. We run the releases, analytics, monitoring and upgrades the way software is run, not handed over. After each release the team gets a report on what sold, where demand outran stock and what to change, so the next one doesn’t start over.