A new storefront for 7,000 sneakers, measured against the store's own register.
Sneaker and streetwear retailer in Woodland Hills. We rebuilt the Shopify storefront, cleaned up a seven-thousand-product catalog, wired search, SEO, and email, and judged the first month against in-store sales over the same weeks.
Growth you can separate from the weather.
Shopify and PostHog, Jul 22 to Aug 26, 2026, against the same window in 2025. In-store sales fell 10% over the same weeks, so the lift is the site, not the season. Conversion rate held near 0.4%; the growth came from traffic. [Figures pending client sign-off.]
Seven thousand products and a checkout nobody could finish.
The week before launch the old site took 2,528 sessions and zero orders. Over five thousand products had no description, 788 had no photo, a 230-product Kobe collection returned a 404, sizes were typed by hand, and every customer email was Shopify's unbranded default.
Rebuilt the storefront on Shopify
New theme, a mega menu over roughly 150 taxonomy collections that had been sitting unpublished, size facets driven by real metafields, a condition scale for pre-owned pairs, and pinch-zoom product photos for a 91% mobile audience.
Cleaned the catalog with a sneaker database
2,155 descriptions and 1,735 SKUs backfilled in two runs. 788 photo-less products pulled off the web and kept in-store only. 39 bad machine-written descriptions caught and cleared before anyone saw them.
Made search the second page on the site
Predictive search, popular chips built from real query data, and a size-intent hint that turns "size 10" into a filter. Shoppers who search buy at 4.3 times the rate of those who do not.
Launched SEO on day one
Breadcrumbs, product and store schema, a 6,859-URL sitemap, corrected store address and hours. 135 clicks from Google by day four; branded queries at number one; "sneaker stores near me" on page one.
Replaced default emails with Klaviyo
Welcome flow earning $14.27 per recipient in its first week. Abandoned checkout, abandoned cart, browse abandonment, and review request flows live within five days of launch.
Read the data and fixed what it said
37% of product views landed on sold-out pairs, so a "more like this, in stock" row shipped the same day. A recently-viewed row underperformed and is on the list to cut. All of it in the weekly report.
“[Quote to request from SoldOut LA]”
[Name, role] · SoldOut LA