What our ISP data shows us from Phase 1 of pensions dashboards testing

As pensions dashboards testing continues, we’re keen to share a picture of what our ISP data is showing us about possible matches, and why schemes and providers need a plan now, not at launch.

Last year, our Pension Pulse research projected that 18.52% of requests to pensions dashboards could return a possible match rather than a confirmed one. It was a projection, built on analysis of nearly three million member records.

Following the conclusion of Phase 1 of testing, we now have real data from 2,152 find requests processed through our ISP. It is a relatively small dataset, gathered in a controlled phase made up largely of industry testers and there are important things to bear in mind when reading it. But it is the first real signal that the projection holds. And it points to an operational challenge the industry needs to confront before dashboards go live at scale.

What we saw during phase 1 of dashboards testing

 The findings are based on c2,500 live find requests processed through Heywood's ISP during the first phase of pensions dashboards testing. Across those requests, 376 pensions were successfully located, with the ISP returning a further 62 possible matches. 

That figure sits within the corridor of what existing projections suggested we should expect. Our own Pension Pulse research in 2025 projected 18.52% of dashboards requests could return a possible match. MaPS data is closer to 10%. Our early live data falls between the two.

The scale implication is what makes this matter. With around 16 million savers expected to use pensions dashboards once fully operational, even conservative estimates suggest schemes and providers could face substantial volumes of member enquiries requiring investigation and resolution.

What to bear in mind

We are sharing this data because we believe the industry benefits from more transparency on early findings. Three things are worth keeping in mind.

• This is early-stage data. The sample is relatively small and drawn from a single, controlled phase of testing. It is a starting point, not a conclusion. There is also a wide range of results – for example testers with relatively common surnames receviced a larger number of false-postitive possible matches – i.e. pensions they knew weren’t theirs.
• Some find requests are experimental. In a testing environment, the same person may run multiple searches with different information to see how the system behaves. That inflates request volumes in a way that will not reflect real-world usage.
• This is a pension-engaged audience. Phase 1 testing was made up largely of pensions industry personnel, who are far more likely to have accurate, recent and complete records than the general public.
• The early testers were not asked to follow up their possible matches, so data providers will not have borne the brunt of enquiries yet.

The phase 1 data lines up closely with what has been projected. Possible matches are a structural feature of how matching works at scale, and they are going to be a recurring part of the dashboards experience. And that stands true even if you’re confident that your data is perfect.

Why this matters and why it gets harder, not easier

The testing cohort was largely made up of pensions industry professionals who typically maintain more accurate and complete records than the wider population. Phase 2 of consumer testing began in March 2026, scaling up user volumes across far broader demographics, including people with accessibility needs and lower digital skills. The records being searched will be more varied, older and less complete. Real members at full rollout will bring historic, fragmented and incomplete data with them.

The figures presented in Phase 1 may represent the floor, not the ceiling. With phase 2 now underway, we are keen to see what a broader, less pension-engaged audience reveals, because it will be a much closer proxy for live conditions.

What follows will be the full-scale rollout – when all the members turn up. Even a low possible match rate, applied to a number that large, translates into a very high volume of member interactions that need resolving. Each one is a moment where the member experience is ambiguous, and where the scheme or provider is on the hook for follow-up.

A possible match is not a found pension. It is an asterisk against the central promise of dashboards: reconnecting people with their savings.

If anticipated projections run true, something in the region of 16 million requests to pensions dashboards when they launch, will be made per year. With a massive peak at the start. This means possible matches are going to be an issue for schemes and providers immediately, and for some time to come. This is not a problem that resolves itself after go-live. This is not a problem that you can think, “let’s wait and see how it goes”. It will be a big wave, followed by a much higher tide of enquiry than usual. It is an ongoing operational reality that needs planning for now. Don’t get swamped.

What schemes and providers need to do

Pensions dashboards are coming. The single most important thing schemes and providers can do is treat their data as the foundation everything else rests on.

Find data, value data and member data all need to be accurate, complete and current. Accurate data is what turns a possible match into a confirmed one, and what stops a saver who really is a member from being turned away with a no match. Remember, once connected, schemes stay connected and data decay does not stop, so data readiness has to be treated as an ongoing discipline rather than a one off project to fix.

It is an inconvenient truth, however, that schemes with perfect data will still get possible matches. This will occur when others’ members with similar data to your members use dashboards. This means that in the short term, your procedures for resolving these possible matches needs to be ready and your teams trained. In the longer term, it will also mean that you will need to manage and fine tune your matching criteria in the face of real world experience.

Getting all of this this right is what protects the member experience and keeps administration workloads manageable, with the success of the Pensions Dashboards Programme contingent on it. People need to connect with their pensions and a possible match that pushes a member offline to chase the scheme creates friction at the worst possible moment. They found you online, so don’t push them offline.

Strong data reduces how often that happens in the first place, and where a possible match still occurs, resolving it while the member is still online, through a secure, scheme-owned journey such as our Possible Match Resolver, keeps the experience digital and controlled. Without that groundwork, the risk once dashboards are live is an administrative scramble: surges in queries, manual resolution case by case, and service levels under pressure at exactly the wrong time.

What we are doing next

We will keep monitoring possible match rates as testing progresses and sharing what we see. The industry is already preparing for dashboards, and data readiness is rightly part of that work. The more visible those efforts become across schemes, providers and ISPs, the better the experience will be when members arrive.