Divisional revenue up £100m.
Through the pandemic’s toughest two years.

Ten thousand staff, dozens of data silos, one pandemic.

Leadership needed to steer a £1bn division through the pandemic, and the data to do it was scattered across departmental silos with inconsistent definitions and variable quality. Applying machine learning and predictive analytics meant reaching beyond internal records into external data — and not all of it was usable. Every source had to be vetted before it could be trusted.

The organisation itself compounded the problem: more than 10,000 staff, spread across countries with different regulatory regimes, languages and operational practices. Country-to-country differences in regulation, enrolment and local operations added layers of complexity that had to be engineered into the analytics from the outset, not bolted on afterwards.

Three constraints were non-negotiable:

• Data locked in departmental silos with inconsistent definitions
• Not all external data sources were reliable enough to use
• 10,000+ staff across multiple countries and regulatory regimes

One governed source of truth, built for a decade of growth.

Siloed data was consolidated into a single, governed and compliant source — the authoritative foundation everything else was built on. A governance framework standardised definitions, lineage and access controls so the platform stayed compliant across every jurisdiction it touched.

Internal operational systems were integrated with a tightly curated set of external data feeds, each one evaluated for reliability before it was let in. On top, machine learning and predictive analytics ran at a strategic level — forecasting demand, modelling cost streams and simulating scenarios — with country-specific regulatory and operational differences parameterised in so results respected local realities while staying comparable group-wide.

Scalability was planned before it was needed. The team worked directly with vendors to guarantee capacity ten times beyond initial requirements, future-proofing the platform against volume the organisation hadn’t hit yet.

The division that grew through the sector’s worst two years.

The tool did more than survive the pandemic — it turned it into a growth period, taking the division from £650m to £750m while workforce costs fell and student experience held up. What’s left behind is longer-lived than the crisis that prompted it: a governed, compliant single source of truth, a proven method for vetting external data, and architecture built to scale well past what the organisation needed on day one.