Six weeks to connect. What the dashboard deadline is really testing.

The engineering is nearly done. That is the good news and it is also the misleading part, because the deadline measures connection and the thing that will be judged is matching.

The judging does not start on 31 October. Members get a dashboard from a date the government sets, with at least six months’ notice, and the Money and Pensions Service expects its own to open in the 2027/28 financial year. Between the two dates, a scheme is connected and nobody is looking yet. That gap is the useful part of this note.

A dashboard works like this: a member proves who they are, the ecosystem asks every connected scheme "do you hold a pension for this person", and each scheme answers. The answer is produced by comparing what the member supplied — name, date of birth, National Insurance number, addresses they have lived at — with what the scheme holds. That comparison is a matching rule, each scheme writes its own, and nobody has ever had to show their working before.

The three outcomes, and only one is good

For any member query, a scheme returns one of three things.

A match. The record is found and the member sees their pension. This is the point of the exercise.

No match. Nothing is returned. If the scheme genuinely holds nothing, correct. If the scheme holds a record under a maiden name, an old address, or a National Insurance number with a transposed digit, the member is told they have no pension with you — which is a statement your administration has just made, on your behalf, to a member, at scale, without anyone reading it.

A possible match. The rules say maybe. The member is asked to contact the scheme. Every one of these is a phone call, and they will arrive in a cluster, and the cluster will arrive in the weeks after the dashboards open when the press coverage lands. A firm that cannot handle a call volume spike will find out then. What an unanswered call actually costs is a calculation worth doing before that rather than after, and the evidence on missed enquiries is not comforting.

What to measure before 31 October

1. Your match rate against a realistic test set. Not against your own clean records — against records with maiden names, historic addresses and typos in them, because that is what members will supply.
2. Your possible-match volume, converted into expected contacts per week, converted into hours. This is the number that decides whether you need help, and it is arithmetic rather than opinion.
3. Your data gaps by cohort. Deferred members who left before a system migration are the ones who will not match, and they are concentrated in particular years. The fix is targeted, not general.
4. The audit trail on every match decision. When a member says "you told me I had nothing", you need to be able to show what was asked, what was compared, and what rule produced the answer. That is a record, kept at the moment, not a reconstruction afterwards.

What this does not tell you

Connection is not compliance with anything else. A scheme can be connected, matching well, and still be failing its members in ways a dashboard cannot show — charges, transfer times, the quality of what happens after the member calls. The dashboard makes one thing visible and leaves everything else exactly as it was.

It also does not tell you what the member will do next. Visibility of a small deferred pot has been assumed, in every business case written for this programme, to produce consolidation. It might instead produce a decade of members discovering pots and doing nothing. Nobody knows yet, and anyone who tells you they do is selling something.

What is knowable is the operational bit: more inbound contact, arriving in a spike, about records you have not looked at in years. That is a handling problem and a chasing problem, and both are easier to solve while nobody is looking than after. Firms in this position are the ones we work with in financial services.