What to put in front of the person deciding

A review is the one outcome of the comparison that is decided by a human, which makes it the only one a reader can influence. This page sets out what that influence consists of - separating evidence, in writing, against the specific question asked - and what does not work at all.

Open the partner account Start with what the clause says

Joins the two: one artefact, one person expectedKeeps them apart: a verified fact that differsA person decides: an overlap a household also produces

Direct answerWhat moves a review is separating evidence: a verified document in each person's own name, a payout account in each person's own name, a second payment instrument, and a written explanation of any genuinely shared one. What does not move it is asserting that the accounts are different, restating that the money is one's own, or producing evidence in someone else's name. Submissions should be attached inside the account rather than sent by email, must answer the exact question the operator asked, and should be dated.

What counts as separating evidence

evidence 1

strongest A verified identity per person. Two different documents that have each passed verification against two different faces. Where this already exists the pair should never have linked, which is why a submission begins by establishing whether both records were verified at all.

evidence 2

strong A payout destination per person. A bank account or wallet in each holder's own name, shown as belonging to that person. This answers the funding question at the same time as the identity one, and it is the cheapest strong evidence to produce.

evidence 3

explanatory A truthful account of a shared instrument. A joint account, a household card, money transferred between partners. This is evidence about the money rather than about the people, and it resolves the overlap instead of denying it exists.

evidence 4

supporting Context, kept short. Two household members, an address explained, a duplicate created by accident and already known about. Useful, secondary, and never a substitute for the three above.

Worked example - a review and what each stage costs (illustrative) Day 0: withdrawal requested, comparison returns two records, account held. Nothing moves. Day 2: operator asks for identity and funding evidence on both records. Reader has supplied nothing so far. Day 4: reader submits two verified identities and two payout destinations in the correct names. Day 7: the ordinary internal target for a decision on a fully evidenced review. Day 21: if no decision has arrived on a complete file, the escalation clock in most licensed markets is measured in weeks rather than days, and the route is a written complaint rather than a chat. The arithmetic that matters is not the dates, it is the gap: evidence supplied on day 4 rather than day 2 moves the decision three days, and evidence supplied after day 7 arrives after the file has been closed on the facts then available. The window is short and it is entirely a reader's to use.

The shape of a submission

Answer the exact question asked. A review asks something specific - which instrument funded the account, whose name the document is in, whether a second person exists. A submission that does not address that question is a delay, however much it contains.
Attach it where it can be filed. Inside the account upload, not by email and never through a link somebody sent first. Evidence an operator cannot attach to a file is evidence that does not exist for the decision. The identity desk covers the document mechanics in full.
Send the least that answers it. Whole pages where the period matters, the name and address visible, unrelated lines masked. A document that has been edited raises doubt and usually generates a fresh request rather than closing one.
Keep a dated copy of what you sent and what came back. A version stamp and a message log are the whole record if the matter is escalated, and they cost nothing to take at the time.

What does not work

Asserting the difference. A statement that the accounts belong to different people is not evidence of it, and the comparison was built because statements are cheap.
Documents in someone else's name. Submitting another person's identity to keep two accounts apart converts a linkage review into an identity problem and usually loses the balance as well as the argument.
Removing the overlap after the fact. Changing a card, a phone number or an address once a review is open does not remove the overlap the review is about, and a pattern of changes during a review is itself recorded.
Escalating first. A complaint to a regulator or an alternative-dispute body before the operator's own route has been used is normally sent back to the operator, and it costs the weeks it took to reach it.

When the finding is right

Some reviews end in a finding of linkage, and this desk does not treat every one of them as an error. Where two accounts really are one customer, the useful work is not to contest the finding but to establish the resolution: which record is recognised, which instrument funded the other, and what the net position actually is. The consequences page sets out that arithmetic, and asking for it in writing - rather than for a reversal - is the request an operator can answer.

The complaint routes themselves. A linked-account decision is an account decision, so the ordinary complaint machinery applies: the operator's own complaints address first, then the alternative-dispute or ombudsman scheme the licence requires, then the regulator for conduct rather than for the money. The series' escalation desk covers those routes in full; this page stops at the evidence, because evidence is the only part of a linkage case that a reader controls.