Insight Analytics
Service

SQL Development

The unglamorous layer everything else depends on. Get it right and dashboards are fast; get it wrong and no BI tool can save you.

Most slow dashboards are slow because of what's happening in the database. A missing index, a scalar function called row by row, a view stacked on four other views — the report tool gets blamed and the real cause never gets looked at.

We do two kinds of SQL work. The first is remediation: profiling your slowest queries, reading execution plans, and fixing the root cause rather than adding hardware. The second is development: building the models, procedures, and views that new reporting sits on, written to be read by the next person.

Everything is version-controlled and deployed through scripts, not by editing objects in production.

Problems this solves

What we usually walk into

  • Reports time out at month-end when volumes spike.
  • A view references a view that references a view and nobody can trace the logic.
  • Nightly jobs overrun into business hours.
  • Database changes are made directly in production with no history.

Technologies

What we build with

SQL Server / T-SQLPostgreSQLAzure SQLSnowflakedbtRedgate / Flyway

Industries that benefit

  • Banking
  • Healthcare
  • Manufacturing
  • Logistics
  • SaaS platforms

Implementation process

How the engagement runs

  1. 01

    Profiling

    Capture actual workload, rank queries by total cost, and find the handful causing most of the pain.

  2. 02

    Plan analysis

    Read execution plans, identify scans, spills, and bad estimates, and confirm the cause before changing anything.

  3. 03

    Remediation

    Indexing, rewrites, statistics, and set-based replacements for row-by-row logic — measured before and after.

  4. 04

    Modelling

    New tables and procedures designed for the queries that will actually run against them.

  5. 05

    Source control and deployment

    Scripts in Git with a repeatable migration path to each environment.

Sample screens

What the finished work looks like

Representative layouts using demonstration data — client work is never shown without written permission.

Query cost ranking

Top query

42s → 0.6s

CPU

-71%

Reads

-88%

Query cost ranking

Top workload consumers before and after tuning.

Nightly job timeline

Window

5.2h → 1.1h

Jobs

28

Overruns

0

Nightly job timeline

Batch window before and after set-based rewrites.

Index coverage

Added

14

Dropped

22

Writes

+3%

Index coverage

Which workloads are covered, missing, or over-indexed.

Illustrative example

How a sql development engagement typically plays out

Anonymised scenario · not a verified client record

Third-party logistics provider

Challenge

The operations dashboard timed out every afternoon once daily shipment volume passed a threshold, and the vendor's suggested fix was a larger server.

Solution

Workload profiling traced 80% of the cost to two views built on scalar functions. We rewrote them as set-based inline table functions, added three covering indexes, and updated statistics maintenance.

Result

The dashboard now returns in under a second at peak volume on the same hardware, and the planned server upgrade was cancelled.

Illustrative figures

42s → 0.6s

Peak query time

$0

Extra hardware

-71%

CPU at peak

This is a composite illustration of the scope, approach, and range of results this service is designed to deliver. It does not describe a specific named client, and the figures are demonstration values rather than audited outcomes. We're happy to talk through real references under NDA on a call.

Deliverables

What you receive

  • Performance baseline and after-state comparison
  • Indexing and rewrite change scripts
  • Documented stored procedures and views
  • Version-controlled database project
  • Maintenance plan for statistics and index upkeep

FAQs

Questions we get asked

Can you tune without changing application code?
Often, yes. Indexing, statistics, and plan-level fixes go a long way before any code change is needed.
Which platforms do you cover?
SQL Server and Azure SQL most often, plus PostgreSQL and Snowflake.
Will you touch our production database directly?
Not without a change process. We work in a copy, produce scripts, and deploy through your approval path.
How much improvement is realistic?
On untuned reporting workloads, ten to fifty times on the worst queries is common. We baseline first so the number is honest.

Related services

Often scoped alongside this

Data Warehousing

Dimensional models on Snowflake, Fabric, SQL Server, and Azure.

View data warehousing

ETL Development

Reliable ingestion pipelines from any source to any target.

View etl development

Power BI Dashboard Development

Executive dashboards, semantic models, paginated reports, RLS.

View power bi dashboard development

Talk through your SQL Development project

A 30-minute call is usually enough to tell you whether this is a two-week fix or a two-month build — and roughly what it costs.

Book a consultation Browse all services