Schema design or review
A model that matches how your business actually works, with the constraints that stop bad data getting in.
What we do
Queries that stay fast as your data grows
Schema design, query optimisation, migrations and reporting pipelines, so your data stays correct, fast and usable past the first ten thousand rows.
The short version
Almost every slow application is a slow database with a web app attached. A missing index, a query inside a loop, or a schema that made sense when there were two hundred records and stopped making sense at two hundred thousand.
We work on the layer underneath your product: how the data is shaped, how it is queried, how it moves between systems, and how it gets reported on. That includes the unglamorous parts, like migrating off a spreadsheet that six people are still editing at once.
Scope
Every item below is in the written scope before work starts, so nothing on this list is an upsell later.
A model that matches how your business actually works, with the constraints that stop bad data getting in.
Slow query log analysis, explain plans, and the indexes that turn a four second page into a fast one.
Versioned, reversible migrations that can run on a live database without a long maintenance window.
Moving off spreadsheets or a legacy system, with deduplication and validation so you do not import the mess.
The numbers your team asks for every week, available without someone writing SQL by hand each time.
Scheduled backups, restore drills and consistency checks so data loss is a scenario you have rehearsed.
How it runs
Five stages, each with something you can see or review at the end of it.
Slow query logs, table sizes and real access patterns, before anyone proposes a rewrite.
The handful of queries causing most of the pain, usually resolved with indexes and small query changes.
Normalise what should be normalised, denormalise what is read constantly, add the missing constraints.
Migrations, imports and scheduled jobs, version controlled and repeatable.
Dashboards and exports so the people who need the numbers can get them themselves.
Questions
If your question is not here, ask it directly. We answer the awkward ones too.
Usually, and we can tell you in a couple of days. We profile the real request path rather than guessing, and the output is a short list of the specific queries and missing indexes causing most of the delay, ranked by how much time each one costs.
Yes, and this is one of the most valuable things we do. We model the data properly, build an interface your team can use, and import the existing sheets with validation so the bad rows are reported instead of silently saved.
PostgreSQL by default: better data types, stricter guarantees, strong JSON support. MySQL or MariaDB is a sensible choice when your hosting is built around it. Either is far better than the third option, which is keeping everything in a spreadsheet.
Most index and query work does not. Larger schema changes are planned as online migrations with a rollback path, and if a short window is genuinely needed we tell you the length before we start.
Also relevant
Cloud infrastructure, CI/CD, monitoring and backups set up so deployments are routine, costs are predictable and recovery is tested rather than hoped for.
Read moreCustom websites and web applications built on Next.js and React, tuned for speed, search visibility and real conversions rather than template looks.
Read moreSend us the situation as it is, including the parts that are messy. We will tell you what we would do first and what it would cost.