The Hindsight KPI: Scoring 2027 Forecasts After the Fact
Most UK SMEs enter a new year with a revenue target and a forecast. Fewer return to the forecast once the year is underway. The target gets reviewed; the forecast gets forgotten. That is a missed control. A forecast is not merely a planning artefact. It is a prediction that can be graded, and the grading is where useful management information begins.
For 2027, add a forecast-accuracy target to the planning pack. The aim is not to predict the future perfectly. It is to know how much weight to place on your own numbers, where bias enters the process, and how much cash headroom to hold when the next forecast is presented to a lender, a board, or an investor.
Why grade the forecast?
Revenue targets measure ambition. Forecast accuracy measures judgement. A business that consistently forecasts within a narrow band can plan hiring, stock, and cash with greater confidence. A business that misses its own forecasts by wide margins is not necessarily failing, but it is operating with a weaker control environment than its management accounts might suggest.
Scoring forecasts also creates a feedback loop. Without it, the same optimistic assumptions survive year after year: pipeline conversion rates that were never tested, revenue that arrives later than expected, costs that are understated because they are inconvenient. A simple scorecard makes those patterns visible without turning the exercise into a blame session.
Set an accuracy target, not a perfection target
A realistic 2027 target for a small UK business might be: at least eight of twelve monthly revenue forecasts within plus or minus 10 per cent; an average absolute revenue error no higher than 12 per cent; and monthly cash low-point forecasts within plus or minus 15 per cent. Quarterly revenue forecasts should be tighter, perhaps within plus or minus 7 per cent, because there is more time for small monthly errors to offset one another.
These are starting points, not universal standards. A project-based consultancy with lumpy revenue may need wider bands than a subscription business with recurring income. The important step is to choose a band, write it down, and apply it consistently. A target that is never missed is too loose; a target that is never met is not a control.
A one-page forecast scorecard
Keep the scorecard deliberately short. One page, one row per period, and a small number of fields are enough to expose the pattern. A workable format is:
| Period | Forecast revenue | Actual revenue | Signed variance | Absolute variance | Score | Forecast cash low point | Actual cash low point | Note |
|---|---|---|---|---|---|---|---|---|
| Jan 2027 | ||||||||
| Feb 2027 | ||||||||
| Q1 2027 |
Use a simple scoring rule. For example: five points where the absolute variance is within 5 per cent; three points within 10 per cent; one point within 15 per cent; zero beyond 15 per cent. Apply the same rule to revenue and cash, then report a monthly score and a rolling twelve-month score. The signed variance matters as much as the absolute variance: a run of positive signed variances points to systematic over-forecasting, while a mix of positive and negative values suggests volatility rather than bias.
Build a review cadence that survives the year
Forecast scoring fails when it is treated as an annual post-mortem. By the time the year ends, the lessons are too late to influence behaviour. A better cadence is monthly and quarterly.
Monthly, after the management accounts are available, hold a twenty-minute forecast review. Compare the month just closed with the forecast made for it. Record the variance, assign the score, and note one reason for the difference. Do not attempt to explain every line; focus on the two or three assumptions that drove the largest variance.
Quarterly, extend the review to forty-five minutes. Look at the rolling score, the direction of bias, and whether the forecast process itself needs adjustment. If sales forecasts are consistently late by a month, change the sales forecast rather than blaming the sales team. If cash forecasts are consistently too high, revisit the assumptions about debtor days and payment terms.
In October or November, use the accumulated scores to calibrate the following year's forecast. That is the point at which the 2027 scorecard becomes an input to the 2028 plan, not a retrospective curiosity.
What the scores reveal
A forecast-accuracy scorecard tends to surface three useful findings. The first is optimism bias. If revenue forecasts are routinely higher than actuals, the business may be counting pipeline too early, ignoring seasonality, or allowing commercial pressure to override evidence. The second is volatility. A business with a low average error but a wide range may be forecasting well on average but carrying more risk than the average suggests. The third is process drift. Accuracy often deteriorates when a forecast is built by one person without challenge, or when the sales and finance views are never reconciled.
These findings feed directly into cash planning. If the typical absolute cash error is 15 per cent, a cash buffer sized on a perfect forecast is too thin. If the signed error is consistently positive, the buffer should be sized on the lower end of the forecast range rather than the midpoint. Lenders respond to this kind of evidence. A forecast track record that shows explained variances and a calibrated process is more credible than a single optimistic projection.
Make the target honest
A forecast-accuracy target should be reviewed, not weaponised. Its purpose is to improve the quality of decisions, not to punish the person who prepared the spreadsheet. Set the target at board level, report it alongside revenue and margin, and treat a missed forecast as information rather than a failure. The objective for 2027 is a business that knows how wrong its forecasts are likely to be, and plans accordingly.
That is a sharper target than revenue alone. It turns hindsight into a KPI, and it gives every future forecast a measured basis for confidence.
