Homelessness prevention by Maidstone Borough Council and xantura

Using predictive analytics to target earlier help

Return to map

The context

The Scottish Government’s Action Plan commits to preventing homelessness for groups at highest risk. Many services already target activity at groups who are over-represented in our homelessness system, on the basis “if you can predict it, prevent it”. There’s rightly a focus on improving the efficacy of prevention activities. But perhaps services should also consider getting better at prediction itself?

One way some agencies have addressed this challenge is through ‘predictive analytics’: solutions ethically linking data from across council services and third party agencies to build up holistic views of households at risk. This allows agencies to target advice and support to individual households upstream, when problems may only just be emerging. Maidstone Borough Council was one of the first to team up with public sector data specialist xantura to test the use of predictive analytics designed specifically to prevent homelessness.

The intervention

In the five years before England’s Homelessness Reduction Act came in in 2018, Maidstone recorded a 58% rise in applications and a doubling of households in temporary accommodation. Inspired by a conference talk, the Council funded a pilot to test whether predictive analytics could increase prevention. xantura provide an enabling infrastructure to analyse structured (i.e. statistical) and unstructured (i.e. case note) data, enabling risks and trends to be better understood by frontline teams and help them target support.

Aided by EY, the Council and xantura agreed information governance and secure GDPR-compliant data-sharing via a process of pseudonymisation across 15 internal services and external agencies, including housing register; council tax and housing benefit data, tenancy debt data from Golding Homes (Housing Association); domestic abuse sanctuary scheme and ‘troubled families’ data from Kent County Council.

In partnership, they designed a suite of dashboards and a textual case summary, accessible through existing housing software and enabling staff with the correct data-sharing protocols and legal gateways to obtain a holistic view of households. They agreed ‘risk thresholds’ that would trigger alerts to the housing team, wherein an officer with a financial inclusion background was appointed to review and act on the alerts, getting in touch with households to offer advice, assistance and support.

The outcome

In year one, 650 alerts were generated for households at risk of homelessness in three to six months, according to agreed thresholds. Due to capacity, the officer could only attempt contact with 260 of these, offering, for example: income maximisation; budgeting and debt support; Discretionary Housing Payments and/or mediation. 0.4% of those households went on to present as homeless. In contrast, during the period, 40% of alerted households the officer did not have capacity to contact went on to present as homeless. A further 30% presented as threatened with homelessness.

Many households for whom the Council received alerts were unexpected, being identified three to six months before crisis point. The Council estimated the intervention saved over £225,000, with potential to save over £550,000 if they had been able to respond to all the alerts. The unintended ‘control group’ caused by limited capacity enabled them to make a case for further investment. The Council has now recruited a second officer to enhance their ability to act on every alert for a household at risk.

Key insights

• predictive data doesn’t just point you to households you already know/can tell are at risk: it enhances a local authority’s existing abilities to target proactive advice and support
• the Council overestimated customer concerns on data-sharing across different departments: most customers expected the Council already internally shared their data this way
• the potential of predictive data can be optimised if a service commits to new
  ways of working, and ensures that it has capacity to act on the data

Find out more…

Natalia Merritt, Housing Advice Manager, Maidstone Borough Council
nataliamerritt@maidstone.gov.uk

Tom Davies, Commercial Director, xantura
tom.davies@xantura.com