Skip to content
Systemcraft

System Optimization

A Ticketing Backend Where Slow Checkout Was Costing Sales

Slow API responses were pushing buyers to abandon carts. Database work and caching fixed the cause, not the symptom.

3x faster

API response speed

-70%

Database queries per request

+25%

Completed purchases

Problem, solution, result

Problem
API response times were slow enough that buyers abandoned checkout. It presented as a frontend problem, but the cause was in the data layer.
Solution
Database optimization, a caching strategy for the queries that ran on every page, and refactoring of the endpoints doing too much work per request.
Result
Faster responses across the board, fewer database queries per request, and a measurable improvement in completed purchases.

It looked like a frontend problem

Buyers complained that checkout was slow, so the first instinct was to look at the page. The page was fine. It was waiting on an API that was waiting on a database doing far more work than the request required.

What we found

  • Queries running on every page load that could be cached for minutes at a time
  • Endpoints assembling data for parts of the screen the buyer never scrolled to
  • Missing indexes, so common lookups were scanning whole tables
  • The same data fetched several times within one request by different pieces of code

Why we measured first

Optimization without measurement is guessing, and guessing tends to produce clever changes to code that was never the bottleneck. We recorded response times and the slowest queries before touching anything, then compared after each change.

Two of the four fixes accounted for most of the improvement. The other two we could have skipped.

The general work is described under backend engineering.

Solutions

  • Digital Transformation

    For businesses whose systems worked at half this size and are now the thing slowing them down.

Services

  • Backend Engineering

    The part nobody sees, and the part that decides whether your product survives a busy Monday.

  • Data Engineering

    One set of numbers everyone agrees on, instead of four dashboards that disagree.

Industries

  • SaaS & Startups

    Ship without painting yourself into a corner, and survive the day your launch works.