Case studies, with what I'd do differently on every one.
Real numbers, the engagement shape, and a what-I'd-do-differently section on every one. If a project didn't earn its money back, you'll read that here too.
tracking apps retired
How BoxQuiz stopped paying for tracking the platform was already doing.
A Danish puzzle shop was running two AdNabu apps to cover Google tracking — work the native Google & YouTube channel already does, at the cost of duplicate event fires. I audited the full tracking and attribution layer, completed the Google tag migration, switched on the Merchant Center and Customer Match pieces that were off, and retired both apps only after two weeks of live data proved the native channel had it covered.
and reconciled
A custom ops layer that replaced broken automation and filled the B2B gaps Shopify doesn't natively cover.
A multi-channel retailer migrated to Shopify in late 2025. The catalog moved cleanly, but a layer of operational logic had no native home — a broken bundle-recompute automation, B2B order classification done by eye, and customer-segment data trapped outside Shopify. One operations layer closed all three, hosted on infrastructure the merchant owns end to end.
across 7 active POs
Accurate per-location stock and pre-order dates for Fathom's 2,238-variant catalog.
Fathom sells cookware and tableware across three inventory locations. Bundle pages showed "ships in X days" placeholders, and the stock math missed components split across locations. I built two custom Node.js apps that write per-location stock and pre-order arrival dates to the storefront — verified with a full production dry-run resolving 74 of 74 SKUs across 7 active purchase orders, zero errors.
Have a problem that looks like one of these?
The 30-min discovery call is free. The first ten minutes go to whether your problem is the problem you think it is.