Case Studies

Engagements, measured against where they started.

Real programmes with baselined numbers. Where a client cannot be named we describe the work without them, rather than inventing a logo.

180+

Engagements delivered

94%

Clients who return

12

Countries

0

Abandoned programmes

Selected work

Six engagements in detail

Each links through to the sector page covering how we work in that industry.

Financial services

Core ledger migrated with no maintenance window

A nine-month parallel run moved a retail bank off its legacy ledger with automated reconciliation throughout. The close window fell from overnight to two hours.

9 monthsCore banking
Explore the sector
Healthcare

Forty per cent less clinical documentation time

Redesigning the clinical note workflow around single-entry capture removed duplicate data entry across four downstream systems in a hospital group.

7 monthsClinical workflow
Explore the sector
Manufacturing

OEE up 23 points across three production lines

Automatic downtime capture replaced clipboard logging and revealed that a third of recorded changeover time was actually material starvation.

5 monthsMES & IIoT
Explore the sector
Retail

Twelve times peak traffic with no degradation

A re-platformed storefront with performance budgets enforced in CI, load-tested to twelve times the previous peak before the trading event.

6 monthsCommerce platform
Explore the sector
Public sector

Counter visits halved for six high-volume services

Services redesigned around the applicants who previously had to attend in person, with assisted digital routes deliberately retained.

11 monthsCitizen services
Explore the sector
Technology

Deploy frequency tripled in a single quarter

Rebuilt the release pipeline for a Series B SaaS company and eliminated the flaky test suite that had made every deployment a judgement call.

3 monthsPlatform engineering
Explore the sector

Measured outcomes

Every figure against the baseline it improved on

A result without a stated baseline and a verification method is a claim. These are the numbers, where they came from, and how they were checked.

Core ledger migration

Financial services
Baseline measured
Overnight close window, 8–11 hours
Result at handover
Two-hour close window
How it was verified
Nine-month parallel run with automated reconciliation

Clinical documentation

Healthcare
Baseline measured
Time-and-motion study across four wards
Result at handover
40% less documentation time per shift
How it was verified
Repeat study on the same wards post-rollout

Production line telemetry

Manufacturing
Baseline measured
Clipboard downtime logs, three lines
Result at handover
OEE up 23 points
How it was verified
Automatic capture reconciled against manual logs

Commerce re-platform

Retail
Baseline measured
Previous trading-event peak
Result at handover
12× peak with no degradation
How it was verified
Load test before the event, telemetry during it

Cloud estate optimisation

Technology
Baseline measured
Twelve months of billing data
Result at handover
38% run-cost reduction
How it was verified
Unit-cost dashboards sustained over two quarters

Security posture uplift

Public sector
Baseline measured
Validated findings, ranked by exploitability
Result at handover
All critical findings closed
How it was verified
Independent re-test on completion

How we report

What makes a case study worth reading

Measurement

The baseline is captured before we touch anything

A result is only meaningful against a number that existed before the work started. We insist on measuring the current state first, which occasionally shows that the process we were hired to fix is already efficient — and we report that too.

  • Baseline agreed with the sponsor before design begins
  • The same measurement method used before and after
  • Instrumentation built into the new process, not bolted on
  • Results reported with the raw data, not only the headline

Honesty

The difficult parts are in the write-up

Every engagement on this page had something go wrong. A case study that reads as an unbroken success is marketing, not evidence — the useful part for a prospective client is what surprised us and what we changed in response.

  • What we got wrong, and when we found out
  • Scope that was cut and the reasoning behind it
  • Estimates that moved, with the revised confidence range
  • Initiatives that missed their target, reported not re-scoped

Verification

References you can actually call

Most of our work sits under NDA, so the studies here are anonymised by sector. What we can do is arrange a direct conversation with a client in your sector and of your size, without us on the call.

  • Reference calls arranged for shortlisted engagements
  • Named references in your sector where the client agrees
  • Full write-ups shared under mutual NDA
  • We do not attend the reference call

Questions

Fair scepticism about consultancy case studies

You should be sceptical. These are the questions we would ask.

About these studies

Why are the clients anonymised?

Most of our work is under NDA, and in regulated sectors the client often cannot be named even where the engagement went well. Rather than publish a shorter list of nameable logos, we publish the engagements that are most useful to read about and arrange direct references for the ones you are seriously evaluating.

Are the numbers independently verified?

They are measured with the same method before and after, agreed with the client sponsor, and in most cases signed off by them as part of the closing benefits review. They are not audited by a third party. Where a figure comes from a client's own system rather than ours, we say so in the full write-up.

Can we speak to one of these clients?

For engagements you are seriously evaluating, yes — we will arrange a call with a client in a comparable sector and size, and we will not be on it. That is usually more informative than any document we could write, which is rather the point.

What is not here

Do you have engagements that failed?

We have engagements that were stopped at a gate, which is different and is designed in. Two programmes in the last three years ended at the diagnostic stage because the findings did not justify the work — in both cases the client kept the findings and we billed only the diagnostic. We consider those good outcomes and we will talk about them on a call.

Why is there no logo wall?

Because a logo tells you a company bought something from us, not whether it worked. The engagement detail and a reference call answer the question a logo is standing in for, and we would rather spend the page on those.

Want the full write-up?

Detailed versions — including the parts that went wrong — are available under NDA, and often with a reference call.