Skip to content
ParamergeParamerge
Stephen Lieberman
Through Paramerge
Contact

PAN Lab example

Xantura OneView (predictive homelessness flagging)

The flag no one can reach: a predictive homelessness-prevention platform

A vendor platform integrates fifteen-plus council data feeds into one household view and flags who may become homeless months ahead; the flags outnumber the single officer who can act on them, so most are never contacted. Modeled on Xantura's OneView. Watch the bottleneck: the model's real reach is set by staffing, not accuracy, and the households it may have named correctly but no one reached never appear on this diagram.

Stylized model of a documented deploymentHousing & homelessness services

Open this example in PAN Lab v0.1 to apply pressures and levers and watch what the system does.

What this models

This example runs on the OneView-class predictive homelessness-flagging platform network: 6 components and 12 pathways between them. Every context in the Lab is a stylized model, never a reconstruction of any actual deployment, and each assumption behind it carries a provenance label.

Evidence base: 4 assumed · 2 published baseline. In the Lab, the shaded evidence band behind each headline readout draws its width from the least-established class below.

  • assumed

    This models the predictive homelessness-flagging pattern documented in the Xantura OneView case file — not a reconstruction of the actual platform or its models.

  • baseline

    The defining dynamic is capacity rationing: the ranked alert list holds far more households than the single officer can contact, so a flag becomes an intervention for only a fraction. The households the officer never reaches are where the harm lands, and that harm is recorded in the case file, not computed here.

  • assumed

    The independent-evaluation node carries a check pathway that is dormant at baseline: before the commissioned randomised trial reports there is no standing external check on the alerts' accuracy, and a single officer meets little internal challenge. These are the latent check pathways levers can open; the case file explains why the trial was commissioned.

  • baseline

    The multi-agency feed and the record-to-model loop are drawn privacy-sensitive: the flags are computed by integrating sensitive non-housing records (offending, health, benefits, debt) under a statutory data-sharing agreement, on residents not told they are scored. What that integration means for those residents is recorded externally in the case file, never computed in these dynamics.

  • assumed

    Peer pathways are authored on both signs: shared triage habits and a single platform's systematic skew reinforce, while both peer-check pathways (an independent accuracy check, and internal challenge to the list) start closed — the documented scrutiny arrived from an outside ethnography and a not-yet-reported trial.

  • assumed

    Every effectiveness figure this class of tool is cited with (a reported 40 percent reduction, savings, ROI, and the contacted-versus-uncontacted contrast) is vendor- and council-reported from a single COVID-affected pilot year and is not the result of a controlled trial. This Lab models institutional propagation only and estimates none of those figures.

What this example does not show

  • The people this system flags — households at risk of homelessness — are not in this diagram. The Lab models how the tool moves through the council's own workflow, never who becomes homeless or is prevented from doing so. The capacity-rationing harm (the households an officer never reaches) and any pattern in who gets flagged are recorded in the case file and measured outside any diagram like this one.
  • Every headline effectiveness figure for this class of tool — the reported reduction in homelessness, the accuracy claim, the savings and the return on investment (ROI) — is a vendor- and council-reported number from a single COVID-affected pilot year, not an independent finding. The randomised trial built to test the causal claim was still in progress. Nothing in this Lab estimates or endorses those figures.

Sources and evidence

What this example rests on, claim by claim. Every entry resolves to the same ledger the Evidence Registry publishes.

  • Xantura's OneView integrates more than 15 multi-agency data feeds into a single household view and flags residents as likely to become homeless months ahead. In Maidstone's pilot year it produced 650-plus alerts that a single financial-inclusion officer could contact only about 260 of. Its headline effectiveness figures - a reported 40 percent fall in homelessness, savings and an ROI over 600 percent, and the widely quoted contrast between contacted and uncontacted households - are vendor- and council-reported pre/post numbers from one COVID-affected pilot year; the contact-versus-no-contact contrast reflects capacity-driven selection rather than a randomised comparison, and the independent randomised controlled trial commissioned to test the causal claim was still in progress into 2026.

    empirical
    • Advocacy Crisis UK, Homelessness prevention by Maidstone Borough Council and Xantura (2023) https://www.crisis.org.uk/ending-homelessness/homelessness-prevention-guide/maidstone-borough-council-and-xantura/
    • Vendor Xantura, Maidstone Borough Council - Preventing Homelessness (vendor case study, 2023) https://xantura.com/maidstone-borough-council/
    • Trade press Government Transformation Magazine, How predictive analytics reduced homelessness by 40% (2023) https://www.government-transformation.com/data/how-predictive-analytics-reduced-homelessness-by-40
    • Government Ministry of Housing, Communities and Local Government, Using data to prevent homelessness - privacy notice (GOV.UK, 2024) https://www.gov.uk/government/publications/homelessness-and-rough-sleeping-using-data-to-prevent-homelessness-privacy-notice/homelessness-and-rough-sleeping-using-data-to-prevent-homelessness-privacy-notice
    • Government evaluation Centre for Homelessness Impact, Can we predict and prevent homelessness? (2024) https://www.homelessnessimpact.org/news/can-we-predict-and-prevent-homelessness

  • OneView's single view of vulnerability is built by integrating sensitive multi-agency records - including offending, health, benefits, and debt data - under a statutory Digital Economy Act 2017 data-sharing agreement with named public-body controllers and processors. An independent ethnography of an early deployment (its fieldwork centered on children's social care and the COVID-19 response) found frontline staff could not see which factors drove the tool's alerts and were not all convinced it was as accurate as described, and a separate NGO investigation characterised the vendor's COVID-era model as operating without residents' knowledge.

    empirical
    • Government Digital Economy Act Register, LBBD OneView - Single View of Vulnerability (data-sharing agreement 376, 2023) https://www.digital-economy-act-register.data.gov.uk/agreements/376
    • Advocacy Ada Lovelace Institute, Critical analytics? Data analytics in local government (research on Barking and Dagenham OneView, 2024) https://www.adalovelaceinstitute.org/report/local-authority-data-analytics/
    • Advocacy Big Brother Watch, The Poverty Panopticon: the hidden algorithms shaping Britain's welfare state (2021) https://bigbrotherwatch.org.uk/press-releases/councils-hidden-algorithms-profile-millions-on-benefits-big-brother-watch-investigation-finds/

Where this connects

Institutional pressures in this domain

  • Workload surge — Demand outruns staffing; per-case attention shrinks and review becomes triage.
  • Austerity & recovery incentives — Cost-cutting and overpayment-recovery targets tilt the system toward denial and enforcement errors.
  • Vendor opacity — The deploying institution cannot inspect the model, data, or update pipeline it is accountable for.
  • Data & policy drift — The world, the intake process, and the rules change under a system trained on how things used to be — two mechanisms with different remedies: the statistical properties of what the system processes move (concept drift), or the mixture of inputs arriving in deployment differs from the mixture it was trained on (covariate shift).
  • Compliance over substance — Paper controls (sign-offs, checklists) satisfy audits while the behavior they describe erodes.

All of them in context on the Housing & homelessness services domain page.

Levers available here and the patterns behind them

Documented case histories