A diagnostic for reconciling — and why running it more than once means you need something else
By Steele Consulting
Every business hits this moment. Two reports come back on the same question. The numbers don’t match.
Sales pulled Q3 at $2.4M. Finance pulled it at $2.1M. The COO pulls up shipments — 145 orders in operations, 132 in the warehouse system. Nobody is lying. All four numbers came out of real systems, run by real people, with real reason to trust them.
Everybody in the room now has the same three thoughts. Which one is correct? Why do the systems disagree? And how do we stop having this meeting?
The reconciliation itself is a solvable problem — five steps we’ll walk through in this post. But the deeper answer is that if you’re running the reconciliation more than once a quarter, you don’t have a report problem. You have an integration problem. And the fix for that is different, more permanent, and worth understanding before you spend another year doing weekly reconciliations by hand.
Why two reports drift
Reports on the “same” business metric drift for five specific reasons. Recognizing which one is at play is 80% of the diagnostic work.
1. Definition drift
The most common cause and the easiest to miss. Sales counts revenue when the order is booked. Finance counts it when the invoice is issued. Marketing counts it when the campaign attributes the click. All three are “revenue” — but they’re not the same revenue. Two reports produced by two teams with two different definitions will always drift, no matter how good the underlying systems are.
2. Source drift
The two reports are pulling from different systems entirely. The CRM has one version of a customer record. The accounting system has another. The ERP has a third. Every system was set up at a different time by different people for different reasons, and there’s no source of truth — just three sources of near-truth that mostly agree.
3. Timing drift
Report A was generated at 9 a.m. Report B was generated at 3 p.m. Between them, ten transactions posted. Both reports were correct at the moment they ran, and they will never agree because they weren’t looking at the same moment.
4. Filter drift
Report A includes test accounts. Report B excludes them. Report A includes returns as negative revenue. Report B excludes returns entirely. Report A converts foreign currencies at yesterday’s rate. Report B converts at today’s. Each choice was sensible at the time; the reports diverge because the choices were different — and usually undocumented.
5. Manual-entry drift
Somewhere in the chain, a human copied a number from one system into another, and typed it wrong. Or updated one field and forgot the other. Or ran a report and typed the results into a slide deck. The more human hands the data passes through, the more of this you get.
Almost every real-world discrepancy is one of these five, or a combination.
The 5-Step Reconciliation
When two reports disagree, this is the sequence that gets you to the truth fastest.
Step 1: Definition check
Before anything else, get everyone in the room aligned on what each report claims to measure. Write the definition down. Compare them side by side. Half of all discrepancies resolve at this step because the two reports weren’t actually measuring the same thing.
Step 2: Source check
Ask each report’s author: what system did the data come from, and how was it queried? If the two reports came from different systems, that’s your discrepancy. Reconcile the two systems, or agree which one is the source of truth for this metric.
Step 3: Timing check
When was each report generated? What time window did it cover? Two reports on “Q3 revenue” run a day apart can legitimately disagree if a big transaction posted overnight. Note the run times and adjust for anything that changed between them.
Step 4: Filter check
What did each report include and exclude? Currencies, subsidiaries, test accounts, refunds, promotional adjustments, in-transit orders. This is where the tedious work lives. It’s also where most discrepancies actually resolve, because the filters were always different and nobody wrote them down.
Step 5: Line-by-line
If the discrepancy is still unresolved after the first four steps, drop to the smallest unit — individual transactions, individual orders, individual line items — and reconcile them one by one until you find the delta. This is slow and expensive, but it always works.
Run these five steps and you will find the answer. What you will not fix is why you have to run them again next quarter.
The bigger point: reconciliation is a symptom
Every hour someone in your business spends reconciling two conflicting reports is an hour that shouldn’t need to happen.
The reason it happens is that your systems are producing separate versions of the same truth, and the reconciliation is what humans do to bridge that gap. It’s the exact pattern we called out as a Friction Cost in our DRIFT framework — labor spent working around a system boundary that shouldn’t exist. And it’s the same underlying problem we wrote about in When a Spreadsheet Becomes a System, when the “spreadsheet” is often literally the tool someone built to reconcile two other systems.
Reconciliation once a quarter is annoying. Reconciliation once a month is expensive. Reconciliation once a week means you’ve built a full-time job out of paying humans to do what a properly integrated system would do automatically. Most growing businesses cross this threshold without noticing.
The real fix: automation and integration
At Steele Consulting, this is one of the most common problem shapes we solve — because it’s an exact fit for what we build. When two reports disagree because two systems don’t talk to each other, the durable answer isn’t better reports. It’s an integration layer that makes the systems talk.
Concretely, this usually looks like one of three approaches.
Direct integrations
We build a connection between the two systems so that data flows automatically from one to the other, in real time or on a defined schedule. The CRM and the accounting system push to each other. The ordering system and the warehouse system share the same order record. No human touches the handoff, so no human introduces drift. This is the right pattern when you have two or three systems that need to stay in sync and the definitions between them are already clean.
A unified data layer
When you have more than two systems in play — and most growing businesses do — the right architecture is often a central data warehouse that pulls from all the sources, applies consistent definitions and filters, and serves every downstream report from the same underlying numbers. Every dashboard, every board slide, every audit report queries the same source of truth. Reconciliation goes away as a category of work. This is the pattern we recommend most often for clients with four or more systems producing overlapping data.
Automated reconciliation with alerts
For cases where systems genuinely have to remain separate — regulatory reasons, vendor limitations, cost — we build a reconciliation service that runs the five-step sequence automatically, flags discrepancies the moment they emerge, and routes them to the right owner before they become a Monday-morning problem. The human still resolves the discrepancy, but they’re doing it on real-time data rather than at the end of the quarter.
None of these is exotic. All of them are the kind of work that has extraordinary ROI when the alternative is a full-time-equivalent of manual reconciliation labor. If the humans in your finance, ops, or sales teams are spending measurable hours every month keeping two systems in sync, the integration pays for itself before the year is out.
How this connects to the other decisions you’re making
Two-report discrepancies almost always intersect with the other frameworks in this cluster. The DRIFT framework quantifies the cost of leaving the problem alone — the Friction Cost bucket usually holds most of it. Build, Buy, or Bend is the decision about which fix to pursue — a middleware tool, a data warehouse, or a custom integration. And Process Problem or Software Problem is the diagnostic for whether the drift is really about broken definitions (process) or broken plumbing (software). Usually it’s both, and both need to move.
How we approach this at Steele Consulting
Our practice was built for this exact category of problem. We build the integrations, the data pipelines, and the reconciliation services that turn “which report is right?” into a question your business stops having to ask. Every one of those projects begins with the same conversation: what’s the actual source of truth, what should be automated, and what’s the honest ROI of stopping the manual work.
If your team has been reconciling two reports every month, every week, or — worst case — every day, that’s an integration problem waiting for a solution. Reach out and we’ll walk through what the fix looks like for your systems.