Your distributors ship your product. Then it disappears.
Windrose collects what your partners actually sold — from spreadsheets, PDFs, EDI, whatever they send — and turns it into records your finance team will sign off on. Then it finds the money leaking out of your claims.
There is no pilot number on this page. When there is one it will be the first thing you see, with the customer profile and the period it came from.
The number you have to defend
Quarter end. Somebody senior asks what actually sold through the channel, and what the claims cost you.
You have a figure. You know how it was built: eleven files, four of them late, two of them re-sent after you pointed out the missing column, one of them a PDF somebody re-keyed. You know which parts are solid and which parts are a best reading of a scanned page.
Nobody in the room is going to ask which parts. They are going to take the number and put it in a board pack, and it will be your number for the rest of the quarter.
That is what Windrose is built for. Not visibility, and not a dashboard — the narrower, more expensive problem of being able to stand behind a channel number, line by line, to the row it came from, without having spent the previous fortnight assembling it.
What it costs you between now and then
Eleven partners, eleven versions of the same file.
One sends an export, one sends a scan, one sends the same workbook with a column moved. Somebody on your team spends the first week of every month asking for files and the second week making them line up, and neither week is the job you hired them to do.
You are asked what sell-through was, and you answer carefully.
Not because you do not know, but because the number was assembled by hand from files that arrived late, and the parts you are least sure of are the parts nobody will ask about until the quarter that matters.
The claims come in, and they get paid.
Checking a chargeback claim properly means reading it against a contract term, a tier threshold and an eligibility rule, line by line, for every partner, every cycle. Nobody has that week. So the claims are sampled, or approved, and everybody knows what that means.
One month end, twice
None of this is a saving measured at a customer yet. It is what the product is designed to change, stated plainly enough that you can tell us where we are wrong.
How it goes now
- Chasing partners for files, in three languages
- Normalising eleven layouts by hand
- Sampling claims, because reading them all is a week you do not have
- A number assembled the night before the review
- “Let me come back to you on that”
How it is meant to go
- Partners chased on their own calendar, in their own language
- A changed layout is a ninety-second review, not a re-mapping
- Every claim read against its contract term, every cycle
- A number that was true on the second of the month
- One click from any figure to the row it came from
Four things stop being your problem
They ship in this order, because nothing downstream is worth anything until the data arrives — which is why the least interesting one goes first.
Stop chasing eleven partners for a spreadsheet.
Windrose asks for the file on each partner's own calendar, in their language, reads it in whatever shape it arrives, and involves you only where a person is genuinely needed. Partner compliance becomes something you can show rather than something you apologise for.
Know which partner runs out before they tell you.
Days of cover by partner and SKU against real lead times, with the accounts that will go short in front of the manager who covers them — while there is still time to do something about it. It drafts the replenishment. You send it.
Find the money going out of the door in claims.
Every chargeback and rebate claim read against contract entitlement, eligibility and end-customer classification, with the dispute pack assembled and the source documents cited. Under-claims come back too — money you were owed and did not take.
See a partner stop buying before it looks normal.
Order cadence measured against each sub-dealer's own baseline rather than a global rule, so the ones quietly drifting to a competitor surface while the relationship is still warm.
One claim, opened all the way
This is the difference between being handed a number and being able to defend it. The claim, the entitlement and the delta are three separate lines, never one computed figure — and the chip after the delta opens the evidence. It works on this page.
- Tier mismatch18
- Duplicate11
- Expired tier9
- Eligibility6
- Under-claim3
Claim 8841-B · Meridian Diagnostics
Tier mismatchHigh confidenceProvisional| Claimed credit | SGD 410.00 × 400 | SGD 164,000 |
| Entitled credit | SGD 379.00 × 400 | SGD 151,600 |
| Delta | estimated | SGD 12,400 |
Tier 3 requires 1,000 units per quarter. Attainment is 640 units, projected 890 at quarter close. Contract MD-2024-11 §4.2, confirmed by J. Tan on 12 Aug. Rule set channel_entitlement v7.
Not issuable yet. The delta depends on quarterly attainment, so it re-adjudicates at quarter close on 30 Sep and may resolve to nothing. Issuing it today would put a dispute in front of your partner that you would have to withdraw.
Every figure in this screen is a design fixture, not a customer result. It is here to show how a delta is separated from its claim and its entitlement, and what opens when you ask where a number came from.