It’s working day four of the close, and the controller is staring at two revenue numbers. One came from the ERP, one came from the FP&A team’s workbook. They differ by $40,000, and nobody in the room can say which is right.
That gap is not a math error. It’s the absence of a single source of truth for financial close, and it’s the reason your month-end keeps slipping past the deadline you promised the board.
This article lays out what a single source of truth actually is, why most finance teams never reach one, and how to build one for the close without an enterprise budget or a year-long data project.
A single source of truth is a governance outcome, not a system
Ask ten finance leaders to define a single source of truth and most will point at a system. The ERP. The data warehouse. The shiny new dashboard. The belief that buying one platform creates one truth is the single most expensive misconception in the close.
A single source of truth exists when any number in your close can be traced to one agreed definition, one agreed source, and one agreed calculation. It’s a governance outcome. Not a database, not a screen.
You can run three systems and still have a single source of truth, as long as the data flowing between them is defined, reconciled, and owned. You can also run everything inside one ERP and have five competing truths, because the data inside it was never governed.
Gartner research has long found that the average organization maintains several competing “sources of truth” for the same financial data. One system does not fix that. Governance does.
What this means for the close specifically
The close is where every ungoverned definition finally collides. Sales booked it, AP recorded it, the GL posted it, and FP&A re-cut it. Four versions, one deadline.
A single source of truth for financial close means that when a reconciliation, a workpaper, and the board pack all reference the same balance, they show the same number, or the difference is documented and explainable on the spot. That’s the whole game.
Why most close teams never get there
Competing truths don’t appear because anyone is careless. They appear because smart people work around a system that doesn’t answer their question. Here’s how it happens during the close, step by step.
1. The shadow spreadsheet
A budget holder needs revenue by product line. The ERP gives them revenue by account. So they build a spreadsheet that re-slices the data using a CRM export.
That spreadsheet becomes “the product revenue report.” It uses different timing and different definitions from finance’s version, and now two truths exist, both built by competent people solving a real problem.
2. Definition drift
Sales calls it revenue when the deal is signed. Finance calls it revenue when it’s recognized. Operations calls it revenue when it’s delivered.
Every definition is defensible in its own context. Without one written rule for which definition the close uses, every team closes against its own private version of reality.
3. The timing gap
Entity A closes on day three. Entity B closes on day eight. The group report on day five blends A’s final numbers with B’s estimates, then the board pack on day twelve uses finals for both.
The two reports disagree. Neither is wrong. But with no documentation of the timing, the board concludes the numbers can’t be trusted, and trust is the one thing the close is supposed to produce.
4. Dirty data flowing in from upstream
This is the failure no one talks about. Most close problems are actually AP problems that arrived a month early.
If invoices were miscoded, vendors were duplicated, or accruals were guessed at during the month, the close doesn’t create those errors, it inherits them. You can govern the close perfectly and still get a bad answer if the data feeding it was never clean. A single source of truth has to start upstream, before the close ever begins.
The old way versus the single-source way
Most teams run the close the first way and wonder why it never gets faster. The contrast makes the cost obvious.
| Dimension | Fragmented close | Single source of truth |
| Where numbers live | Scattered across ERP, email, and dozens of spreadsheets | One governed environment with defined sources |
| Reconciliations | Rebuilt manually in Excel each period | Run against clean, validated data with audit trails |
| When discrepancies surface | At review, on the last day | Continuously, as data flows in |
| Definitions | Implicit, in people’s heads | Written, agreed, enforced |
| Versions of a number | As many as there are reports | One, or a documented exception |
| Workpapers | Recreated from scratch monthly | Generated automatically from the close process |
| Audit prep | A scramble for evidence | A standing trail auditors can self-serve |
| Upstream data | Cleaned reactively during close | Structured at the AP stage, before close |
| Close visibility | “Ask the controller” | A live dashboard everyone can see |
| Knowledge | Lives with one person | Lives in the system |
The right column isn’t a fantasy. It’s what disciplined governance plus the right tooling produces, and you can reach it incrementally.
Build it in five steps, starting with five numbers
You don’t need a comprehensive data program. You need to start with the handful of numbers your board actually sees, govern those, and expand outward.
Trying to govern everything at once is how these projects die.
Step 1: Pick the five numbers that matter
For most mid-market companies the list is short: total revenue, gross margin, EBITDA or operating profit, net cash position, and one operational metric the board always asks about. Start with the five figures on every board slide, nothing more.
Step 2: Define each one in writing
For each number, document the exact calculation, the source system, the timing of extraction, the inclusion and exclusion rules, and the person accountable for it. For most teams this is the first time those definitions have ever been written down, and writing them down is half the battle.
Step 3: Trace and reconcile across sources
Find every place each number appears: the board pack, the management report, the dashboard, the statutory accounts. Compare the figures and document every difference and its cause. An unexplained difference isn’t a rounding quirk, it’s a governance gap with your name on it.
Step 4: Eliminate the competing truths
For each unexplained gap, decide which source is authoritative and align the rest to it. This doesn’t mean forcing every report to show an identical figure, because timing and scope differences are legitimate.
It means every difference is now documented and explainable.
Step 5: Put it on a cadence
Once the five numbers are governed, reconcile them every period on a fixed schedule with clear owners and tolerance thresholds. When the first five are stable, add the next five, and the single source of truth grows one reconciliation at a time.
Where the single source of truth should actually live
Governance is the principle. But the close still needs a place where reconciliations, journal entries, workpapers, and the audit trail physically come together, or the governance lives in a binder no one opens.
For mid-market finance teams, that place should do four things:
- Reconcile high volumes of transactions automatically and surface only the genuine exceptions
- Keep the flexibility of Excel-style work while wrapping it in versioning, change tracking, and an audit trail that standalone spreadsheets can’t reliably provide
- Explain period-over-period movements with drill-down to the underlying transactions in one screen
- Post clean journal entries to the ERP through a controlled approval workflow
This is the gap purpose-built financial close software is designed to fill.
The point isn’t to replace your ERP. It’s to give the close a governed home that the ERP alone was never built to be.
How DOKKA Close builds the single source of truth, upstream and down
Most close tools sit downstream and react to whatever data shows up. DOKKA is built around a different idea: clean the data upstream so the close has less to fix in the first place.
That’s the upstream-downstream model. DOKKA AP structures and validates invoice data at the accounts payable stage, before it ever reaches the close. Fewer coding errors and duplicate vendors going in means fewer reconciliation breaks coming out.
DOKKA Close then automates reconciliations, flux analysis, and close workflows against that already-clean data. Together they form one record-to-report layer that prevents close problems instead of just reacting to them.
Inside DOKKA Close, every closing task is an activity card with a status, an owner, unlimited reviewers, and built-in chat, so the close stops living in email threads and offline trackers.
Reconciliations run on a split screen: the Excel-like grid on one side, the chart-of-accounts mapping on the other, with the delta calculated automatically and posted to the ERP with a single click through a journal template. Existing Excel reconciliation templates upload directly, so nobody rebuilds work they already trust.
Flux analysis is where the single-source idea pays off most visibly. You set your own materiality thresholds, then drill from any flagged variance straight into the underlying transactions in the same screen, instead of toggling between sheets and losing your place.
Reporting delivers Balance Sheet and P&L out of the box, embedded inside DOKKA so versioning and audit trails are preserved, with Excel export whenever you need to share externally.
The result is one governed environment where the number on the dashboard, the number in the reconciliation, and the number in the report are the same number, with a trail behind it. DOKKA is ISO 27001 certified and has completed an independent SOC 2 examination, and it integrates natively with NetSuite, QuickBooks, SAP Business One, Priority, and more, with API connectivity for ERPs not on the list.
It’s also built for the team that actually has this problem. DOKKA is used by 3,500+ finance teams and is designed for mid-market groups of roughly two to ten people, the segment that needs enterprise-grade control without enterprise complexity, cost, or a months-long rollout.
Curious what the numbers look like for your team? The close automation ROI calculator gives you an estimate in a couple of minutes.
Three mistakes that quietly rebuild the chaos
Even teams that buy into the idea tend to fall into the same traps. Each one feels like progress and quietly recreates the problem it was meant to solve.
Declaring a source of truth without eliminating the others
A team names the ERP “the single source of truth” in a kickoff deck and moves on. Meanwhile three departments still maintain their own revenue cuts.
Naming a system is branding, not governance. A single source of truth exists only when the competing numbers are reconciled and resolved, not when one system gets a label.
Buying tools before fixing definitions
The instinct is to solve a data problem with a purchase. But software pointed at ungoverned data just produces conflicting numbers faster and spreads them wider.
Define the numbers first, then automate the governed process. Technology is the accelerator, not the engine. Tooling on top of clean definitions is leverage; tooling on top of chaos is a megaphone for it.
Treating every difference as an error
Two reports showing different revenue can both be correct, one pulled on day three with preliminary figures, the other on day ten with finals. Without timing documentation, that looks like a mistake and erodes trust.
The fix isn’t to force the numbers to match. It’s to document why they differ so the difference is expected, not alarming.
Frequently asked questions
What is a single source of truth in financial close?
It’s the state where any number in your close traces back to one agreed definition, one source, and one calculation, so a reconciliation, a workpaper, and the board pack all show the same figure or a documented exception. It’s a governance outcome, not a single piece of software.
Does a single source of truth mean replacing all our systems with one platform?
No. You can keep multiple systems and still have a single source of truth, as long as the data between them is defined, reconciled, and owned. Consolidation isn’t the goal, governance is, and a purpose-built close platform helps by giving that governed data one place to live.
Why does our close produce different numbers in different reports?
Usually because of definition drift (teams define a metric differently), timing gaps (reports pulled on different days with different data completeness), or shadow spreadsheets that re-slice data outside the governed process. Each is fixable by defining the number, documenting the timing, and reconciling on a cadence.
How long does it take to build a single source of truth for the close?
For the five board-pack numbers, a focused team can get there in roughly 60 to 90 days, and the full close suite takes longer. The trick is to start narrow with five numbers and expand, rather than attempting everything at once.
How does clean AP data help the financial close?
Most close errors are inherited from upstream: miscoded invoices, duplicate vendors, and guessed accruals all create reconciliation breaks at month-end. Structuring and validating that data at the AP stage means the close starts from clean data and has far less to correct, which is the basis of DOKKA’s upstream-downstream model.
The close is a trust exercise, and trust needs one number
The job of the close isn’t to produce numbers. It’s to produce numbers the board can act on without second-guessing them, and that only happens when there’s one number, one definition, and a visible trail behind it.
Start with five numbers. Define them, reconcile them, govern them, and then give the work a home that keeps it clean from the AP stage through to the report.
Do that, and working day four stops being an argument about which figure is right.
Want to see how a single source of truth for the close works in practice? Book a free demo and we’ll walk through it with your numbers.
