“The query review caught a join we had been patching with caches. Once the covering index landed, the checkout timeout alerts dropped within a day. I still wish we had booked the assessment a quarter earlier—we spent too long arguing about whether the schema was ‘fine.’”

— Nattapong K., engineering lead, retail platform, Bangkok

“During the refactor engagement they refused to merge our ‘temporary’ columns into the permanent model. That friction saved us a second rewrite six months later.”

— Mira S., product engineering manager, logistics SaaS serving Thai merchants

“The schema assessment was shorter than I expected and more opinionated than a typical audit. It gave finance a number and gave us a priority order we could defend.”

— Anan P., CTO, education marketplace

Extended story: catalog lookup after a merger

A regional retailer inherited two product catalogs after acquiring a smaller brand. Lookups at checkout joined both catalogs plus a promotions table that still used string matching on SKU codes. Service Riverbase mapped the hot path, introduced a read-model table refreshed by a nightly job for the first phase, and only then proposed renaming the surviving catalog tables. Cutover happened on a Sunday evening with a documented rollback that was never needed. The client kept the migration notes in their runbook and later reused the same staging checklist for a warehouse schema change without our involvement.