Take a number into the budget meeting that finance already trusts

Platform-reported conversions do not reconcile to the general ledger, and finance notices eventually. Once that credibility goes, every marketing number in the room is discounted — including the accurate ones.

Starting from the company's own closed-sales export means your reporting and finance's reporting begin at the same figure. You are no longer asking anyone to accept a number from a system they do not audit.

Spend is cut, held or increased every planning cycle whether or not anyone can prove which channels worked. The absence of evidence is not neutral; it favors whoever argues most confidently.

A reconciliation against closed revenue moves the decision from advocacy to arithmetic. It is as likely to defend a channel you believe in as to end one you do not.

The oldest argument in any commercial organization is whether the leads were any good or the follow-up was. Marketing points at volume delivered; sales points at the quality of what landed; both are partly right and neither can prove it, because the two teams read different systems with different definitions of the same event. The argument is rarely resolved and it is expensive, not because anybody is being unreasonable but because the evidence required has never been assembled anywhere. It is also what decides how the budget is treated when times get harder.

A reconciliation between what marketing delivered and what sales closed is that evidence, and it is deliberately unflattering to whichever side is wrong. Close rate by source ends the general argument and starts a specific one: this source converts at a tenth of that one, so either it should be bought differently or it should be worked differently. Both of those are actionable and neither requires anybody to concede a position in the abstract. Run monthly, it gives the argument something to be about other than blame.

The same file answers a question sales asks and marketing usually cannot: which sources produce the deals worth the sales team's time. A source generating a large volume of small deals consumes exactly as much pipeline attention as one generating a few large ones and returns a fraction of it. Ranking on closed revenue rather than on opportunities created makes that visible in one column, and it tends to change qualification criteria before it changes media spend. That is the correct order of operations and rarely the one followed.

Marketing slides get accepted in board meetings and discounted immediately afterwards, and the mechanism is usually a single follow-up question nobody could answer. Where did that figure come from, and does it reconcile to the revenue in the accounts? A number that cannot survive that question does not merely fail; it costs credibility on the next slide and the one after. Platform-reported conversions cannot survive it, because they were never designed to reconcile to a ledger and do not. Everybody in the room knows this and the conversation moves on politely.

A figure built from the company's own closed-sales export starts from the number finance already agrees with, which changes what a follow-up question can do. It can ask how much of the total was traceable, and the answer is a stated percentage with the remainder shown rather than hidden. It can ask how a specific deal was credited, and the answer traces back to two rows somebody exported. It cannot ask whether the underlying revenue is real, because that revenue is what the board already accepted elsewhere in the pack.

The unattributed share is the part to present deliberately rather than apologetically. A board that sees forty per cent of revenue in Direct / Unknown and understands why — existing customers, referral, brand, trade — is better informed than one shown a complete attribution it has no reason to believe. It also protects against the failure mode where a suspiciously precise number is challenged once, found to be modeled, and drags everything around it down with it. Presenting the limits of a measurement alongside its findings is what makes the findings usable next quarter.

Most in-house teams receive reporting from at least two agencies and several platforms, each counting in its own way, and then produce an internal summary that reconciles none of them. The totals do not agree, everybody knows the totals do not agree, and the internal number becomes whichever version was easiest to defend. That is not a criticism of the agencies, who are reporting what their tools measure. It is a structural consequence of asking several interested parties to count one outcome without a shared denominator that somebody owns.

Running the reconciliation in-house gives every supplier the same basis. An agency's campaign report and the company's revenue view can be read together, differences explained rather than argued about, and a supplier whose numbers consistently exceed the reconciled figure becomes a specific conversation rather than a vague suspicion. Most agencies welcome this once they understand it applies to everybody equally, and the better ones use it to make their own case more forcefully. The ones who object usually object for reasons worth investigating on their own merits.

The practical arrangement is unremarkable. One workspace, the company's closed-sales export, whichever lead exports the agencies and platforms can supply, combined into one lead file with each row tagged by source and campaign, and a mapping confirmed once. Seats and roles control who can import and who can only read, so an agency can be given a report without being given the ability to change anything. Nothing is installed on the website, no permissions are granted to the CRM, and the whole arrangement can be tested on a finished year before anybody commits to it.