How verification works

What each verification tier actually proves, and exactly how a mediator's reimbursement rate is calculated.

Two different things, and they are not the same

A listing carries two independent signals, and it matters that you read them separately:

  • Verification is what TrustDrive has checked about the mediator — who they are, whether the business exists, and whether anyone has looked at their briefs. It is earned by producing evidence, and it cannot be bought.
  • The reimbursement record is what influencers have reported about being paid. It is arithmetic over declared campaigns. TrustDrive does not decide it and nobody's opinion goes into it.

A mediator can be fully verified and still pay badly. A mediator with a good record can be completely unverified — most listings are, because most were created by an influencer rather than by the mediator themselves, and nothing about the mediator can be checked until they claim the listing.

The verification ladder

Four rungs. Each one names what was actually checked, and each answers a question you might reasonably ask before handing over your own money. There is deliberately no rung for "we audited their bank records against what influencers declared" — that check is not performed, so there is no badge that could claim it.

Unverified

Nothing has been checked. This is the floor, and it is the permanent state of every listing nobody has claimed. It is not an accusation and it does not mean anything is wrong. It means exactly what it says: nobody at TrustDrive has confirmed anything about this mediator.

Identity verified

If they stop replying, is there a real named person on file, and is the number the campaigns come from actually theirs?

The listing has been claimed; a one-time code was confirmed on the WhatsApp number campaigns are actually sent from, not a secondary line; a government ID or business PAN was uploaded; and a person at TrustDrive checked that the document matches the name on the listing and has not been altered. Valid for 730 days.

What it does not prove: nothing about whether they pay, and nothing about their briefs.

Business verified

Is there a traceable entity and a traceable money channel — so if they do not reimburse me, there is something to name in a complaint?

Identity verification still valid, plus business documents — a company or GST registration, or a PAN and bank proof for a sole operator — read and checked by a person against the name and address on the listing. We read the documents; we do not query a government registry, and we do not confirm that money actually moves through any account named on them. Valid for 365 days.

What it does not prove: still nothing about whether they pay, and nothing about whether their briefs put your account at risk.

Compliance audited

Will working with them put my own marketplace or social account at risk?

Business verification still valid, plus a written policy describing how they instruct influencers to disclose paid relationships, plus a sample of actual campaign briefs they sent to influencers and of deliverables that were actually published, read for whether the disclosure was really there and for the account-risk practices listed in the listing policy. Their stated compliance position must already be a compliant one — the audit checks whether the claim holds, so it is a precondition rather than an outcome. Valid for 270 days — the shortest window on the ladder, because cash-flow practices change faster than a business's identity or registration does.

It is revoked immediately if a brief is substantiated as instructing non-disclosure or requiring an account-risk practice.

Rules that apply to the whole ladder

  • No tier is ever self-declared. A mediator can request an upgrade; they cannot set one. There is no form anywhere in the product that writes a verification tier.
  • No tier can be bought. Verification is a different thing from the paid plans entirely, and no payment can write a verification field. Where a mediator pays for visibility, that is disclosed as what it is.
  • Every tier expires, and every tier depends on the one below it. Losing one drops a mediator to the highest rung whose evidence is still valid, never silently to nothing. Losing the claim on a listing takes the whole ladder with it.
  • Being listed is free of any check; being recommended is not. Being listed, reviewed and read requires nothing, because a directory that only covers businesses who opted in covers the wrong half of the market. Putting a mediator on a curated shelf is different, and TrustDrive's position is that it should require a named person behind the listing and, where the work is marketplace reviews, briefs that somebody has actually read. Neither can be bought: there is nothing on this site for a mediator to pay for.

Not in force yet. Both halves of that last rule. Identity verification is what will gate recruiting status and recruiting placement, and the compliance audit is what will gate any surface promoting marketplace-review work. Neither gate is wired to a tier yet, so read the tier badges on a listing rather than assuming a placement implies one.

How the reimbursement rate is worked out

This is the number on the listing that says something like "reimbursed 94% in full across 212 settled campaigns reported by 130 influencers". If you are a mediator disputing your own figure, this section is how to reproduce it from your own ledger.

Which campaigns count

A campaign is in the calculation if both of these are true:

  • The mediator confirmed it, or their confirmation window passed. A campaign the mediator is currently contesting is excluded while it is being decided, and a campaign staff removed is excluded permanently. A campaign the mediator simply ignored still counts, labelled as unconfirmed — otherwise ignoring TrustDrive would be a one-click way to erase your own record.
  • The money question has settled — reimbursed in full, reimbursed in part, or not reimbursed. Campaigns still waiting on a deliverable or still inside the payment window are live exposure, not record, and they are reported separately.

Two ratios, and we publish the worse one

Over that set, two things are counted:

  • By count — campaigns reimbursed in full, divided by campaigns counted.
  • By value — total rupees reimbursed, divided by total rupees ordered.

The published rate is the lower of the two. Not the average of them — the lower one. Here is why that matters. A mediator who reimbursed 190 of 200 campaigns has a 95% rate by count. If the ten they did not pay were the ten largest orders, the rate by value might be 61% — and 61% is what you actually stand to lose. Taking the lower number means a large failure cannot be diluted by a hundred small successes. Read the other way, a mediator who short-pays many small orders while settling the big ones is caught by the count instead. Averaging the two would cancel out exactly the signal each one contributes.

Nothing is smoothed, weighted, or adjusted. There is no prior, no recency curve and no hidden coefficient. A rate a mediator cannot recompute from their own records is a rate they cannot argue with, and that would forfeit the one advantage this number has over an invented "trust score".

There is no overall trust score on this site. A single blended number would hide the thing that actually distinguishes two mediators with identical payment records: one runs ₹400 phone-case campaigns and one runs ₹18,000 smartphone campaigns, and those are completely different risks to you. So the exposure figures sit beside the payment figures, and neither is folded into the other.

When no rate is shown at all

Below 5 settled campaigns from at least 3 different influencers, there is no rate. Not a zero, not a provisional figure — the listing says there is not enough data yet.

A zero would state a measurement nobody took, and it would rank a brand new honest listing below a mediator with a genuinely bad record. The second half of the floor — different influencers, not just different campaigns — exists because anybody can create a listing about anybody here. Counting campaigns alone would let one person with a few accounts manufacture a reimbursement rate for a mediator they have never met.

At launch, most of the directory will have no rate. That is the truth, so that is what it says.

The two speed numbers

They are measured on different clocks and they answer different questions:

  • Days to reimbursement — from when you submitted the deliverable to when you were paid. This is the wait after you have done your part.
  • Days your money was held — from when you placed the order to when you were paid. This is the one you actually feel, because the money leaves your account at the order, not at the deliverable.

Both are medians, not averages, so one extreme case does not distort them.

What counts as late

Against the mediator's own stated payment window, not against a schedule we invented. A mediator who publishes "45 days" is not overdue on day 30. Only where a mediator has stated nothing do we apply a default of 30 days, and there is a 3-day grace period on top before anything is counted as overdue, so a long weekend does not flip a listing's signal.

Stating a very long window buys a mediator no position in the directory. It only puts the stated window on the page next to the observed median, where you can compare them.

What the numbers cannot tell you

  • They describe what influencers reported. They are not a guarantee about what will happen to you.
  • Verification of identity or of a business says nothing about whether that business pays. Only the top tier speaks to that, and only for the period it was audited.
  • An audit examines a sample, on a date. It is not continuous supervision.
  • Nothing here can make marketplace-review work safe for your account. TrustDrive can make the risk legible, make a mediator's terms explicit, refuse the worst operators, and record who paid. That is worth having and it is not safety. The listing policy sets out the practices that put your account at risk and what we do about them.

Related: what a non-payment flag means, the review policy, and the listing policy. Any change to how these numbers are calculated is published in the policy changelog.