Ask three people for last month's wholesale number and you'll get three answers.
Picture a brand where sales says it booked $4.2M, operations says $3.1M shipped, and finance reports $2.9M in net revenue. (These figures are illustrative, not from a customer.) Nobody is wrong. Each team counted a different event, from a different system, on a different date.
The trouble starts when those numbers meet in the same meeting. One team thinks another is padding or sandbagging. An hour goes to reconciling instead of deciding. By the next review, the gap is back.
This post is about closing that gap without a data warehouse project. You need four things: a shared list of order events, one order record that the ERP backs up, a one-page metric dictionary, and a weekly view that bridges the numbers instead of arguing about them.
Here's an illustration. It's not customer data, and your accounting policy governs the details.
A retailer pre-books $100,000 of spring product. Before anything ships, they cancel $10,000 of styles. The brand ships $60,000 in March. Another $25,000 is scheduled for April. The last $5,000 is waiting on a late factory receipt. Later, the brand agrees a $3,000 markdown allowance on the March shipment.
At the end of March, each team looks at the same order and reports something different.
| Team | Question they're answering | Number for this order |
|---|---|---|
| Sales | What did we book? | $100,000 booked, or $90,000 after the cancellation |
| Operations | What went out, and what's still open? | $60,000 shipped, $30,000 open |
| Finance | What did we earn? | $57,000 net: $60,000 invoiced, less the $3,000 allowance |
All three numbers are correct. None of them is "the" number. Each answers a different question, and each team is going to keep answering its own question unless you give them a shared way to connect them.
Most of the disagreement comes from five places.
An order isn't revenue. Accounting standards draw this line clearly. Under IFRS 15, an entity recognizes revenue when it satisfies a performance obligation by transferring goods to the customer, and the standard treats discounts, rebates, refunds, credits and price concessions as variable consideration. US GAAP's ASC 606 works the same way, and PwC's guidance on rights of return explains how revenue is limited to the amount the entity expects to be entitled to. Sales is measuring a commitment. Finance is measuring a transfer.
Wholesale orders move. Brands say so in their own filings. Columbia Sportswear's 10-K for fiscal 2025 says sales to its wholesale customers, other than international distributors, are generally order by order and subject to cancellation and rescheduling before shipment. G-III Apparel's 10-K describes sales as occurring order by order, subject to cancellation. Wolverine World Wide's 10-K for fiscal 2018 goes further: it says a period-to-period comparison of backlog is not necessarily meaningful and may not predict eventual actual shipments. So even "open orders" is a number each company has to define for itself. Deloitte's roadmap notes that remaining performance obligations are commonly referred to as the "backlog" disclosure, and a Deloitte Heads Up from 2018 notes that contracts with an original expected duration of one year or less are exempt from that disclosure. For a seasonal order book, that leaves much of the definition up to you.
Pre-book and at-once run on different clocks. A pre-book order may be placed months before it ships. An at-once order ships now. If one team reports by order date and another by start date, the same order lands in different months. RepSpark's Order Summary Report lets you run booked, canceled, shipped and returned dollars by either Start Date or Order Date. That's useful, and it's exactly why the team has to pick one.
Channels get counted differently. An order a retailer places on their own isn't the same as one a rep keys in. RepSpark order records carry the order's source and the creator's role, and the Orders page can filter by order source. Some orders never start in the platform at all, such as EDI or phone orders that land in your ERP directly. If one team counts them and another doesn't, the totals drift.
Net isn't gross. Concessions such as markdown allowances and chargebacks typically reduce revenue, because they fall under the price concessions and credits that IFRS 15 treats as variable consideration. Check the treatment with your accounting team. Sales rarely sees those numbers when it closes the book on an order. Finance rarely sees why a style was cancelled.
Independent research shows the pattern more broadly. A 2024 Gartner survey of 303 sales leaders found that 84% agreed sales analytics had less influence on performance than leadership expected. Respondents named poor data quality and limited cross-functional collaboration as barriers, each at 44% (Gartner). That's sales analytics across industries, not wholesale specifically. But it matches what enterprise wholesale teams describe.
Don't start by arguing about which number is right. Start by listing the events an order passes through, then give each one a name everybody uses.
For most wholesale brands, the list looks like this:
RepSpark's order flow already follows this shape. Per RepSpark's order flow documentation, an order moves from Draft to Submitted, may require brand approval, is sent to the ERP, and becomes Valid once the ERP confirms it. Shipped, open and cancelled are tracked at the line level, so you can see them by style. Your ERP owns what happens after that.
Write the list on one page. Then add a date basis to each. Booked by order date. Shipped by ship date. Invoiced by invoice date. A shared date basis removes a lot of month-end arguments before they start.
A number is only as trustworthy as the system it comes from. Shipped, invoiced and net are facts your ERP creates. The ordering platform should show them, not recompute them.
That's the pattern behind Full Turn's setup. Its case study describes AIMS360 as the system of record for inventory and reporting, connected to RepSpark through a direct API integration. It says what gets sold and what gets shipped now work from the same source of truth.
Be realistic about what flows back. RepSpark can display ERP shipment and invoice data when your integration provides it, and the Orders page brings Orders, External Orders, Invoices and Payments into one place. External Orders are orders synced back from the ERP, including RepSpark orders processed downstream and orders from other sales platforms. What your finance team sees depends on what your ERP integration sends, so ask your implementation team to confirm which events flow back for your ERP.
The same logic applies to inventory. If you want the full argument, our post on what available inventory data actually means covers why the ERP should own the logic and the ordering platform should show it.
This is the part most teams skip, and the part that makes the biggest difference.
For each metric, write down:
You don't need to invent this format. The SEC's 2020 guidance on key performance indicators generally expects public companies to give a clear definition of each metric and how it's calculated, why it's useful, how management uses it, and to consider disclosing what changed in the calculation (KPMG summary). That's a sensible checklist even if you're private. A change log matters most. When someone quietly shifts "open orders" from start date to order date, trend lines break and trust goes with them.
MIT's Center for Information Systems Research found the same problem in a case. At FEMSA's OXXO, each functional area used its own profitability metric. The company replaced them with a single metric, and the change needed top-management sponsorship to stick (MIT CISR). It's one case, not proof. But the lesson fits: a shared definition needs a sponsor.
Keep the dictionary short. Ten to fifteen metrics is plenty. If it needs more than one page, you're documenting things nobody argues about.
Give each metric to the team that creates the underlying event. Sales owns booked. Operations owns shipped, open and fill. Finance owns invoiced and net. "Owns" means owning the definition and answering questions about it, not winning the argument.
Then pick two or three metrics all three teams care about, and make them shared. Two good candidates:
Shared metrics give the teams a reason to talk. They also expose problems that live in the handoffs. For example, RepSpark's results with L*Space show invalid orders falling from about 20% to near zero. Invalid orders are a booked number nobody can ship against, and they're the kind of problem that shows up as a gap between sales and ops. If you want the ownership side in more depth, see how brands structure a team to own their wholesale stack.
You don't need one number. You need every number tied to its definition, and a bridge between them.
Gartner has made a version of this point to finance leaders. In an April 2021 release based on a December survey of 127 finance leaders, Gartner advised finance leaders to promote a "sufficient versions of the truth" strategy rather than a "single source of the truth" approach, and warned that scattered analytics lead to duplication and less confidence in which results are "correct" (Gartner). Several versions are fine when each is named, owned and reconciled.
So build the bridge. One page, every week, with the same rows:
When the numbers don't match, the page tells you which step changed. A late factory receipt moves dollars from shipped to open. A cancellation moves them out of booked. An allowance moves them from invoiced to net. Nobody has to guess.
A weekly cadence is our suggestion, not a research finding. Use weekly in peak booking and shipping season, and monthly when volume is slow. What matters is that it's regular, short, and run from one page.
Plan for edge cases from the start. Decide whether EDI and phone orders appear in the bridge. Decide how to treat orders split by ship date. Decide whether a split order counts once or twice. Write the answers in the dictionary.
You can start without buying anything.
If your orders live in several places, expect step five to be the hardest. Our posts on why forecasting breaks down when wholesale data lives in too many places and how to get one version of the truth go deeper on the data side. If the disagreement is really about changing platforms, read why sales, ops and finance rarely agree on when to switch.
RepSpark connects to your ERP and treats it as the system of record, with 20+ ERPs and 35+ integrations supported. Retailers, reps and your own sales team place orders in one platform. Orders flow to your ERP, and the ERP confirms them back.
Inside RepSpark, the Order Summary Report shows booked, canceled, shipped and returned dollars by date range, and Top Accounts shows booked, shipped, open, canceled and returned dollars by account. The Orders page brings orders, external orders, invoices and payments into one place. Order records carry source and creator-role fields, so you can see how an order entered the system.
To go deeper, see how RepSpark works with enterprise brands, how ERP integrations work, and what the platform offers operations teams and sales teams. If you want to see your own order events mapped end to end, schedule a demo.
Each team counts a different event. Sales usually counts bookings, operations counts shipments and open orders, and finance counts invoiced revenue net of allowances, chargebacks and credits. The teams also pull from different systems and use different date bases, such as order date or ship date. Each number answers a different question, so they rarely match without a bridge between them.
Bookings are orders accepted. Shipments are goods that have left the warehouse. Invoices are bills issued in the ERP. Revenue is recognized under accounting standards such as IFRS 15 and ASC 606 when goods transfer to the customer, in the amount the company expects to be entitled to. That can be lower than the invoice after allowances and returns. A booked order is a commitment, not revenue.
You need one system of record for each fact, not one number for everything. Your ERP should own shipped, invoiced and net. Gartner has advised finance leaders to aim for "sufficient versions of the truth" instead of a single source, because each version can be valid if it's defined and reconciled. The goal is that every number has a named definition, an owner and a bridge to the others.
For each metric, list the name, the event it counts, the date basis, what's included and excluded, the system where it's calculated, an owner, and a change log. Keep it to one page and ten to fifteen metrics. The SEC's guidance on key performance indicators uses a similar checklist: a clear definition, how it's calculated, why it's useful, and how management uses it.
Give each metric to the team that creates the underlying event. Sales owns booked orders. Operations owns shipped, open and fill. Finance owns invoiced and net revenue. Ownership means owning the definition and answering questions about it. Then choose two or three shared metrics, such as perfect order, that all three teams review together.
They run on different clocks. A pre-book order can be placed months before it ships, so reporting by order date and reporting by start date put the same order in different months. At-once orders ship almost immediately, so the two dates are close. Pick a date basis for each metric, write it down, and report pre-book and at-once separately where it matters.
It gives every team the same order record and keeps downstream facts in the ERP. Orders flow from the platform to the ERP, and the ERP confirms them back. Where your integration provides shipment and invoice data, the platform can display it next to the order. What flows back depends on your ERP and integration, so confirm which events sync before you design your reporting.
We suggest weekly during peak booking and shipping periods and monthly otherwise. This is practical advice, not a research finding. What matters more is that the review runs from one page, uses the same rows each time, and explains movement between booked, shipped, invoiced and net instead of debating which number is right.
Sources: