10,000 Homes, 100 Sensors: What Asset Directors Should Screen For First
By Tricity Labs, Technical Team — Independent retrofit measurement & verification, working to BS 40101 principles
Not every home in a portfolio needs a sensor on day one. Smart meter screening before sensor deployment for housing stock uses the SMETS2 data landlords already hold to flag which properties show retrofit shortfalls, heat-loss anomalies, or unusual consumption patterns -- so hardware only goes where the evidence points. Done properly, this narrows a 10,000-home portfolio down to the roughly 100 properties that need attention this month, instead of guessing or waiting for a complaint to arrive.
Three things to know:
- Screening every connected property against existing smart-meter data costs a fraction of sensor deployment, which is why it should happen first, not as an afterthought.
- A useful screen depends on a weather-corrected, pre/post retrofit baseline -- raw meter readings alone will flag noise, not risk.
- An anomaly flag from screening is a reason to look closer, not proof of a problem; that confirmation is what Tier 2 sensor data is for.
Key Takeaways
| Point | Detail |
|---|---|
| Screening comes before sensing, not alongside it | Tier 1 smart meter screening covers the whole connected portfolio at low cost. Tier 2 sensor gateways are deployed only to properties that screening has already flagged, keeping hardware spend proportionate to actual risk. |
| You likely already have the data you need | SMETS2 smart meters installed across most UK housing stock generate consumption data that can be screened retrospectively -- no new hardware is required to run the first pass. |
| Weather correction is not optional | Comparing raw consumption across two winters without adjusting for degree-days will produce false flags. A defensible screen normalises for weather before comparing pre- and post-retrofit periods. |
| Screening answers 'where to look,' not 'what's wrong' | A flagged property tells you it's worth a closer look. Only room-level sensor data on temperature, humidity, and occupancy can confirm whether that's damp risk, heating underperformance, or something benign. |
| The ratio matters more than the absolute number | Expect screening to flag a minority of stock -- often single-digit percentages -- for Tier 2 attention. If screening is flagging a third of your portfolio, the baseline or thresholds likely need revisiting before you trust the output. |
| Screening data ages | A property that screens clean this quarter can screen differently next quarter after a boiler fault, a tenancy change, or a retrofit measure elsewhere in the block. Screening is a recurring discipline, not a one-off audit. |
Table of Contents
- What Smart Meter Screening Actually Measures Before a Sensor Goes Anywhere Near a Wall
- Why Screening the Whole Portfolio Beats Guessing Which Properties to Inspect
- A Smart Meter Screening Checklist for Housing Stock Before Sensor Deployment
- How to Evaluate Smart Meter Screening Data Before Committing to Sensor Deployment
- Best Practices for Moving from Screening to Targeted Sensor Deployment
- Where Screening Can Mislead You, and How to Guard Against It
- How Verity Applies This in Practice
- Sources
What Smart Meter Screening Actually Measures Before a Sensor Goes Anywhere Near a Wall
Smart meter screening before sensor deployment for housing stock works from data that already exists. Every connected property with a SMETS2 meter is generating half-hourly consumption data whether or not anyone is looking at it -- screening simply puts that data to work before any physical hardware is ordered or installed.
At its core, the screen does three things: it establishes a pre- and post-retrofit baseline for each property, weather-corrects the consumption figures so a cold January isn't mistaken for a heating fault, and flags anomalies where consumption deviates from what the property's characteristics would predict. None of this requires site access.
The output isn't a verdict. It's a shortlist -- properties where consumption patterns diverge enough from expectation that they warrant a closer look. That distinction matters because it's what keeps the whole approach proportionate: screening is cheap enough to run across an entire portfolio, precisely because it isn't trying to diagnose anything on its own.
Why Screening the Whole Portfolio Beats Guessing Which Properties to Inspect
The alternative to screening is usually one of two things: inspecting properties on a fixed cycle regardless of condition, or waiting for a tenant to report a problem. Both approaches have the same weakness -- they're reactive to schedule or to complaint, not to evidence.
A fixed inspection cycle spends the same effort on a well-performing home as a struggling one. A complaint-driven approach only surfaces problems tenants notice and report, which under-counts anything silent -- heat loss through an under-insulated party wall, for instance, rarely prompts a phone call until the damp is visible.
Screening changes the ordering. Instead of spreading inspection effort evenly or waiting for symptoms, it directs attention toward the properties where the data already suggests something is off. For a housing association managing several thousand properties, that's the difference between checking on 100 genuinely flagged homes this month and trying to inspect a few hundred at random.
Run the screen before, not after, planning your retrofit rollout schedule. Properties that screen poorly pre-retrofit are also the ones worth prioritising for post-retrofit verification, since a weak baseline makes any improvement easier to measure clearly.
A Smart Meter Screening Checklist for Housing Stock Before Sensor Deployment
Before committing budget to Tier 2 sensor hardware, it's worth working through a short screening checklist to confirm the groundwork is solid. This isn't exhaustive, but it covers the points most likely to undermine a screen if skipped.
First, confirm data coverage: what proportion of the portfolio has usable SMETS2 data, and how far back does it go? A screen built on six months of data behaves differently to one built on two full heating seasons. Second, confirm the baseline period is genuinely comparable -- same or adjacent occupants, no major void periods, no undocumented works that would distort consumption independent of any retrofit.
Third, check that weather correction is applied consistently across the comparison period, not just at one end of it. Fourth, decide in advance what anomaly threshold triggers a Tier 2 referral -- deciding this after seeing the results invites bias toward whatever number looks tidy. Fifth, agree who reviews flagged properties and on what timescale, so a flag doesn't sit unactioned for months.
Finally, build in a review point to revisit thresholds after the first screening cycle. Portfolios differ enough in age, fabric, and tenancy pattern that a threshold calibrated for one estate may need adjusting for another.
How to Evaluate Smart Meter Screening Data Before Committing to Sensor Deployment
Knowing how to evaluate smart meter screening data before sensor deployment mostly comes down to asking whether the flagged list makes sense against what you already know. If screening flags a block that was fully retrofitted last year, or misses one with a known history of complaints, that's worth investigating before trusting the wider output.
A reasonable evaluation also looks at flag volume. If screening returns a shortlist that's a very small or very large fraction of the portfolio compared with what similar organisations report, the baseline or correction method may need revisiting rather than accepting the number at face value.
It's also worth checking whether the screen distinguishes between different types of anomaly -- a property using unusually little energy might indicate a fabric improvement working well, or it might indicate a tenant rationing heating they can't afford. Screening data that treats both as the same flag loses useful nuance; evaluation should confirm the screen is at least separating these categories before Tier 2 resource is allocated.
Cross-check a small sample of flagged and unflagged properties against any existing complaint or repairs data you hold. Agreement doesn't prove the screen is right, but persistent disagreement is an early signal something in the baseline needs attention.
Best Practices for Moving from Screening to Targeted Sensor Deployment
Following smart meter screening best practices means treating Tier 1 and Tier 2 as genuinely sequential, not parallel. Sensor gateways -- measuring temperature, humidity, and occupancy room by room -- are deployed to flagged properties specifically because that's where the marginal cost of hardware is justified by the marginal value of the evidence it produces.
Good practice also means being explicit about what each tier is and isn't telling you. A Tier 1 flag identifies a property worth investigating. Tier 2 sensor data, once installed, is what actually surfaces damp and mould risk, heat-loss anomalies, heating underperformance, or cold-home welfare concerns with the granularity needed to act.
Another element of best practice is documentation: recording why each property was flagged, what the baseline period covered, and what threshold triggered the referral. This matters for two reasons -- it supports any evidence trail needed for retrofit funding or compliance purposes, and it makes the next screening cycle easier to calibrate because you can see what worked.
Finally, resist the temptation to over-correct by deploying sensors everywhere as an insurance policy. That defeats the purpose of screening in the first place and reintroduces the cost problem screening was meant to solve.
Where Screening Can Mislead You, and How to Guard Against It
Screening is only as good as the data feeding it, and there are a handful of ways it can quietly go wrong. Void periods, meter swaps, and tariff changes can all distort a consumption baseline without any actual change in the building's performance -- if these aren't flagged and excluded, they'll show up as anomalies that have nothing to do with heat loss or damp risk.
Occupant behaviour is a second source of noise. A change of tenant partway through a comparison period can shift consumption patterns for reasons entirely unrelated to the fabric of the building, and a screen that doesn't account for this can send Tier 2 resource to the wrong properties.
A third pitfall is treating a single flagged period as conclusive. Consumption can dip or spike for a month for reasons as mundane as a holiday or a boiler service. Screening is more reliable when it looks for sustained deviation over a full comparison window rather than reacting to a single reading.
None of this is a reason to distrust screening -- it's a reason to build the checks above into how a screening programme is run, so the shortlist it produces is one that genuinely deserves the sensor budget behind it.
How Verity Applies This in Practice
This two-tier structure -- screen everything cheaply, deploy hardware only where it's earned -- is the design principle behind Verity. Tier 1 screens every connected property using existing SMETS2 data, building pre/post retrofit baselines with weather-corrected comparisons and anomaly flags, at a cost that scales sensibly across a whole portfolio rather than per property inspected.
Tier 2 then deploys sensor gateways only to the properties Tier 1 flags, measuring temperature, humidity, and occupancy room by room to surface damp and mould risk, heat-loss anomalies, heating underperformance, and cold-home welfare concerns with the specificity that screening alone can't provide.
For a portfolio of 10,000 homes, that's the practical difference between guessing where problems might be and knowing, from what the data is showing this month, which roughly 100 properties need attention now. Housing associations exploring this approach can see how it fits alongside retrofit evidence requirements, or read more about how the underlying platform is structured before making a hardware decision.
See how VERITY provides continuous monitoring
Book Demo
Photo by Jakub Żerdzicki via Unsplash (source)
Related reading: how the Tier 1/Tier 2 architecture works, housing associations planning a retrofit rollout, retrofit evidence requirements, contractors coordinating sensor deployment, what this means for tenants, more about Tricity Labs
Sources
About Tricity Labs
The team behind Tricity Labs' Verity platform, delivering independent pre- and post-retrofit measurement and verification from existing smart meter data for housing associations and retrofit contractors.