The situation
In most Olympic sports, whether an athlete is paid to compete or work in the sport determines how they are classified — and classification determines which events and crew combinations they are eligible for. This governing body runs exactly such a scheme: every competitor holds one of three groups, assessed against detailed regulations covering employment, coaching and paid activity in the sport, with decisions reviewed by a dedicated commission and renewed on a fixed cycle. Careers, sponsorships and race entries hang on the outcome. The regulations themselves lived where regulations usually live: in a policy document, applied by people.
Why the existing approach was limiting
A rulebook applied by hand scales at the speed of the people applying it. Every application — including the large majority where the regulations produce an obvious answer — consumed the same scarce commission attention as the genuinely contested ones. Consistency depended on each reviewer reading the same words the same way, and demonstrating that consistency to an appellant was harder still. Each policy revision meant re-briefing everyone who applied the rules. And the surrounding administration — chasing missing information, tracking evidence, remembering when each decision expired — was exactly the kind of clerical load that quietly absorbs an organisation's time.
Signal North's role
Signal North designed and built the categorisation platform end to end: the athlete-facing application flow behind the organisation's single sign-on, the rules engine at its centre, the commission's review and appeals workspaces, and the production infrastructure — defined as code on the same cloud foundations Signal North had built for the rest of the estate. The starting point was the governing body's own logic specification: the work was to turn a document that described decisions into a system that makes them, without losing the human judgement the hard cases deserve.
Key decisions and intervention
- encode the regulations as data, not code: versioned rulesets of conditions and outcomes, edited through an administrative interface with inline help and a glossary, so a policy change is a ruleset revision rather than a software release;
- make the application form adaptive: fourteen questions that reveal themselves only when the rules need them — an attestation opens its follow-up detail, a date of birth drives the age exemptions — so athletes answer the minimum necessary;
- decide automatically wherever the rules are conclusive, and route only the genuine judgement calls into the commission's review queue;
- build review around evidence and dialogue: commission assignment, information requests with the full staged history of questions and answers, evidence uploads decoupled from the decision itself, and every action recorded in an audit log;
- treat appeals as a first-class process, not an inbox: appellants attach evidence at submission, exchange information with the panel, and an overturned decision re-categorises the athlete automatically;
- issue every outcome as a formal decision document and track validity without human memory — two years as standard, five where Olympic rules apply;
- test the documentation as well as the software: the published help and worked examples are seeded into the test suite and run through the engine, so the guidance athletes read and the behaviour they experience cannot drift apart;
- keep the whole platform on infrastructure defined as version-controlled code, behind the organisation's existing member identity.
What changed
Clear-cut applications are now decided the moment they are submitted, against rules applied identically every time. The commission's attention goes where it is actually needed — the contested minority — supported by a workspace that holds the evidence, the correspondence and the history in one place rather than across inboxes. When the organisation clarified its validity policy, the change shipped as a ruleset and configuration revision, not a redevelopment project. And because more than nine hundred automated tests exercise the engine, the workflows and the documented examples, the regulations' behaviour is now something the organisation can prove, not just assert.
The deeper change is where the rulebook lives. It is no longer a document that people interpret; it is a versioned artefact the organisation can read, amend, test and audit — with a complete record of what was decided, by whom or by what, and why. Athletes get faster answers and a fairer, more transparent process, including a proper route of appeal. The governing body gets a scheme it can defend: every automatic decision traceable to a numbered rule, every human decision traceable to a person and their reasoning.
Signal North was the right builder for this because it treats systems like this the way it treats all operational software: the rules belong to the organisation, not to the developers. A platform whose policy can only be changed by programmers has just moved the bottleneck. By putting the regulations themselves into safe, versioned, testable controls — and reserving engineering effort for the engine that enforces them — Signal North left the governing body running its own scheme, at a pace no manual process could match, without a development queue in the way.