Risk-Based Modelling: Knowing Before The Failure

There’s been a quiet shift in how asset teams are thinking about condition, compliance, and repairs.

Reactive models are no longer enough. Waiting for complaints, raising works orders, hoping the job gets done that approach is increasingly out of step with regulatory expectations and resident need.

Instead, landlords are being pushed toward something more proactive:

That’s the aim of risk-based modelling. It’s not new. But it’s becoming a baseline.

The Logic is Simple The Execution Isn’t

At its core, risk-based modelling means bringing together the data you already hold from repairs, surveys, EPCs, complaints, and tenant profiles and using it to identify which properties are most likely to develop serious issues.

It’s about scoring properties by risk, not reacting to whichever ones happen to complain first.

For example:

• A flat has two recorded leaks in 18 months

• Its EPC shows poor insulation

• There’s no recent condition survey

• A vulnerable resident lives there

That home should raise red flags. But if those warning signs are split across five systems, it’s unlikely anyone is connecting the dots.

This is the gap risk-based modelling aims to close.

Where It’s Already Happening

Some landlords are starting to get this right.

PA Housing has built a risk model that combines historic repairs, EPC scores, and resident vulnerability to flag homes at risk of damp and mould. L&Q has developed internal dashboards that prioritise fire risk and maintenance need across medium-rise blocks. Peabody is applying similar logic to fuel poverty and overheating.

Even in smaller organisations, basic risk scoring is helping to target capital works more effectively and meet the requirements of Awaab’s Law.

It’s not about AI. It’s about using what you know in one place, at the right time.

Building a Risk-Based Model

You don’t need machine learning or a consultant-built dashboard to start.

Here’s what most models require:

1. Data Integration
Link records across systems housing, repairs, asset, complaints, CRM.

2. Indicator Selection
Choose the right flags: damp repairs, insulation gaps, EPC bands, resident age, unresolved complaints.

3. Scoring Rules
Weight each factor based on how closely it correlates with past failure. For example, multiple damp jobs in 12 months might carry more weight than a single EPC rating.

4. Validation and Testing
Run scores against known outcomes disrepair claims, complaints, legal cases to make sure high scores match high actual risk.

5. Action Framework
Decide what happens when a property hits a red flag: inspection, proactive repair, capital review, or tenant outreach.

The Governance Perspective

This is no longer just good practice it’s a regulatory requirement.

The Regulator of Social Housing’s July 2025 report was blunt: landlords must have a joined-up, property-level understanding of condition. Complaints alone won’t cut it. Neither will occasional surveys or blanket programmes.

Recent C3 and C4 downgrades have cited failures in data integration and risk targeting. The message is clear:

A risk-based model creates a paper trail. It shows that you’re looking at the right indicators, flagging high-risk homes, and acting accordingly. That’s exactly what the new Safety and Quality Standard demands.

What Gets in the Way

Despite the clarity of the logic, many landlords are stuck.

• Repairs jobs aren’t coded consistently
• Survey data lives in PDF reports, not structured fields
• Resident vulnerability data is partial, siloed, or out of date
• Teams operate in isolation, with different risk definitions
• No one owns the response when a property is flagged

Risk-based models reveal these problems fast. That’s part of their value — but also part of the discomfort.

Changing the Culture

The technical barriers are real. But the bigger challenge is cultural.

You can build a scoring tool. You can link the data. But if the organisation doesn’t trust it, use it, and act on it nothing changes.

Risk-based modelling only works when it becomes part of business as usual. That means:

• Letting data guide inspections, not habit
• Accepting that historic survey programmes may have missed high-risk homes
• Giving repairs, assets, compliance, and tenancy teams a shared language of risk
• Moving from chasing backlogs to targeting root causes

A Final Word

Risk-based modelling isn’t a silver bullet. But it’s a step toward a more evidence-led, accountable, and humane approach to property management.

You don’t need to predict the future. You just need to see the signals clearly enough to intervene before a tenant is forced to do it for you.

In an environment of rising scrutiny, rising costs, and rising resident expectation, that’s no longer a bonus. It’s the job.