Before the dashboard: agreeing what the numbers mean
Before a company can manage by its numbers, it has to agree what they mean. That work is mostly about definitions, owners, and a reconciliation to the ledger, and very little about tools.
- Management reporting usually fails on definitions, not technology. The same metric name means different things in sales, operations, finance, and treasury, and each meaning is valid for its own purpose.
- A business intelligence tool shows whatever definitions it is given, so buying one before the definitions are agreed only makes the disagreement faster.
- A metric dictionary, with a definition, formula, owner, source, and reconciliation for each measure, does more than any new platform. The executive set should be small: ten to fifteen measures, each tied to a decision.
- Operational metrics should reconcile to the ledger every month. When the management pack and the accounts disagree, confidence in both wears away quickly.
The monthly business review opens with the revenue slide. Sales reports that March was up 12 percent on last year. Finance says 4 percent. Treasury points out that collections fell. The next forty minutes go on working out who is right, and the meeting ends without a decision about the business, because in fact nobody was wrong.
Some version of this happens in companies of every size, and the usual response is to buy reporting technology: a data warehouse, a dashboard platform, a consultant to connect the two. Months later the dashboards look good and the arguments carry on, now with better charts. The tools were never the problem. The organisation had not agreed what its numbers meant.
One word, five numbers
Revenue is the clearest example, because every function has a legitimate need for its own version of it.
Five legitimate answers to “What was revenue in March?”
Illustrative company, the same month measured five ways, ₹ crore
Illustrative figures. Each measure is valid for its own purpose; the problem is presenting them interchangeably without a definition.
Source: SA&A analysis
Soni Aditya & Associates
Sales leadership manages bookings, because bookings show the pipeline turning into commitments. Operations and billing track invoices raised. A regional manager paid on gross sales wants discounts and returns kept out of the measure used to judge performance. Finance reports revenue under IFRS 15, recognised as performance obligations are satisfied and net of variable consideration. Treasury cares about cash collected.
Each of those figures is correct. The trouble starts when they appear under a single label, in different meetings, with nobody saying which is which. After that, every variance discussion begins with a reconciliation, and people learn to distrust their own reports.
Why tools do not solve it
Reporting platforms are very good at pulling data together and displaying it. They have no view on what it means. When a dashboard project begins, as most do, with data connections and chart design, the definitions get decided implicitly by whoever builds the data model. That is usually an analyst making reasonable choices without the authority to make them stick. The dashboard ends up reflecting one function’s view, and the others reject it.
A dashboard built before its definitions are agreed does not settle the disagreement. It automates it.
The sequence has to run the other way. Agreement on meaning comes first, and it is an organisational decision rather than a technical one. Technology then implements a settled definition, which is a much smaller project and a much more successful one.
What a usable metric looks like
A metric is ready for management reporting when it can be written down so precisely that two people, working separately, would arrive at the same number. In practice each measure needs a small set of attributes.
| Attribute | Example: days sales outstanding |
|---|---|
| Question it answers | How quickly are we turning credit sales into cash? |
| Definition | Trade receivables at month end, expressed as days of recent credit revenue |
| Formula | Trade receivables excluding GST, divided by credit revenue for the last 90 days, multiplied by 90 |
| Included and excluded | Gross of the credit loss allowance; excludes unbilled revenue, advances, and related party balances |
| Source | Receivables sub-ledger and the revenue accounts in the general ledger |
| Owner | A named person: the head of credit and collections |
| Frequency and cut-off | Monthly, on the ledger after close |
| Reconciliation | Receivables agree to the balance sheet; revenue agrees to the income statement |
The detail matters more than it looks. Trade receivables include the GST billed to customers, and revenue does not. A company that calculates the ratio from raw balances overstates collection days by roughly the rate of tax, and one that switches between methods from month to month reports an improvement that never happened. These are exactly the choices a dictionary exists to settle once.
Leaving GST in receivables adds nine days to DSO without a single customer paying later
Days sales outstanding for one company, calculated two ways, days
Illustrative: revenue of ₹300 crore over the trailing 90 days, and receivables of ₹166.7 crore before GST at 18 percent (₹196.7 crore including GST).
Source: SA&A analysis
Soni Aditya & Associates
Fewer measures, clearer owners
Executive reporting should contain the ten to fifteen measures that actually drive decisions, and each should have exactly one owner, answerable for the number and for explaining how it moved. Everything else is analysis. It is useful, available on request, and kept out of the pack leadership reviews.
Two habits help keep the list short. The first is to ask every proposed metric to name the decision it informs, and to drop the ones that cannot. The second is to balance lagging measures, which describe outcomes, with a smaller number of leading ones that predict them: order intake ahead of revenue, overdue receivables ahead of write-offs, attrition in critical roles ahead of delivery problems.
Reconcile to the ledger every month
Operational reports and financial statements come from different systems, are cut off at different times, and are adjusted for different things. They will not agree on their own. They should agree explicitly, through a standing monthly bridge that explains the difference between the operational figure and the ledger: timing and cut-off, returns and credit notes, discount and rebate accruals, unbilled revenue, intercompany sales, and taxes.
Once the bridge exists, its reconciling items become predictable, and anything unexpected stands out at once as either a data problem or a real event. More importantly, leadership learns that the pack and the accounts tell one story, which is the condition for trusting either of them.
The sequence that works
- Start with decisions. List the decisions leadership takes each month and quarter, and the questions each one depends on.
- Collect the measures already in use. Gather every metric currently reported, including each function’s version of it.
- Write and agree the dictionary. Draft the definitions, resolve the conflicts between functions, and have the leadership team sign it off, not the reporting team.
- Map sources and build the bridge. Identify the system and field behind every input, and the reconciliation to the ledger.
- Build the reports last. Visualise the agreed measures in whatever tool the company already owns, if it is good enough.
- Control changes. Treat any change to a definition as a controlled change: dated, communicated, and either restating prior periods or flagging them.
Companies that work in this order usually find the technology question shrinks, and the monthly review changes character. The forty minutes once spent reconciling numbers go on the question the numbers were meant to raise: what are we going to do about March?
