Complex cases in income recovery

Peabody is one of the UK’s oldest not-for-profit housing associations, founded by philanthropist George Peabody in 1862 to help Londoners in need. Today it takes a local approach, working with residents and communities to strengthen its 140 neighbourhoods. Peabody looks after 109,000 homes and 220,000 residents across London and the home counties, and provides care and support to 25,000 customers.

The rent and service charge collection function at a housing association is an area where data quality issues can have far-reaching consequences for residents.

Peabody defines cases that fall outside the standard pattern as “complex cases”. These are precisely the cases where the consequences of getting it wrong are most serious for the resident. They also do not lend themselves to the same workflows or KPIs as conventional arrears.

Peabody’s collections team needed to track every complex case through a multi-stage workflow – including cases that did not progress in a ‘straight line’: new information could legitimately push a case back to an earlier stage. The system of record captured what happened as a stream of timestamped events in a database, but reconstructing the workflow journey for any given case needed to be done manually and could take a long time.

The team wanted an easier way to answer questions that were critical to delivering a high-quality service, including which cases had got ‘stuck’, and which apparent arrears cases were actually complex cases that needed to be handled differently. Pursuing a complex case as if it were a straightforward arrears matter can produce outcomes that materially affect a resident’s life – but spotting the signs is hard when data is locked up in inflexible or siloed source systems.

Our team led the work, which involved reconstructing case journeys from an event-level data model that did not natively expose workflow stage. We modelled the events so that for any given case, we could establish exactly where it sat in the workflow. This included correctly handling cases that had “moved backwards”, i.e. there was now more work required to conclude the case as a result of receiving new information.

We then built a dashboard to make these journeys visible: by case, by workflow stage, and by category. Our design meant we could bring out complex cases as a distinct cohort to be managed differently from conventional arrears.

A common hub for all data use cases

Architecturally, although the collections use case sat in a different area of the business to Neighbourhoods and Repairs, it shared dimensions through a common hub we had built. This preserved the data mesh structure Peabody had chosen, while ensuring that residents, homes and other shared entities had a single consistent definition across the estate. This conformed dimension strategy meant that we were developing all the distinct use cases coherently, rather than as siloed solutions – often a risk with visualisation projects. Peabody’s solutions architect cited this as evidence that the use cases were being built coherently rather than as a set of disconnected solutions.

The same ways of working applied across all our use case delivery: close engagement with the business, iterative review at every stage, and a tight feedback loop between visualisation and data engineering. Given the stakes, the quality of the dialogue with the business was particularly important: numbers on the screen had to be ones the collections team could stake recommendations on.

The biggest impact has been better visibility of complex cases. The collections team can now see, for any case, where it has reached in the workflow and how long it has been there – and which ones should not be pursued as conventional arrears cases. Our solution also shows where operational bottlenecks are occurring so the collections team can fix issues before they escalate into complaints.

Feedback from the workshops we ran included phrases like ‘this is a quantum leap’, and ‘it’s weird to be excited about data’ – clear signs of the impact our work has had for teams that were starved of the insights they needed to improve outcomes for Peabody and its residents.

This project benefited from our joined-up approach to modelling: the collections picture of a resident can be reconciled with the rest of the data estate so that teams have a complete view. This is critical in complex cases where patchy information can lead to incorrect decisions on how to progress a case. As Peabody continues to improve services through its use of data, the collections team will benefit from a richer picture of their residents, ensuring they can take the right action swiftly and confidently.