The run rate formula, the choice of window and why it matters, a worked example against a seasonal and a lumpy business, the four ways run rate misleads, seasonality, one-off orders, a short window and a trend, and the rule for when to annualise and when not to. With the account-level use that makes run rate a ranking rather than a forecast.
Run rate is the number that gets quoted when the year is not over. It is easy to compute and easy to compute wrongly, and the wrongness is invisible because the formula is so simple. This guide gives the formula, the window, a worked example, the four ways it misleads, and the per-account use that avoids most of them.
Run rate = revenue in the window ÷ window length × target period length
Three months of $1.2m, annualised: $4.8m. One month of $400,000, annualised: $4.8m. Both are run rates. Only one of them is stable.
| Window | Stable? | Current? | Use |
|---|---|---|---|
| 1 month | No | Very | Rarely |
| 3 months | Usually | Yes | Monthly-ordering accounts |
| 6 months | Yes | Somewhat | Lumpy accounts |
| 12 months | Very | No, it is the trailing year | Seasonal businesses; baselines |
Stated on every figure. A run rate with no window is a number without a meaning.
A company with a strong fourth quarter.
| Quarter | Revenue | Annualised run rate from that quarter |
|---|---|---|
| Q1 | $3.0m | $12.0m |
| Q2 | $3.4m | $13.6m |
| Q3 | $3.6m | $14.4m |
| Q4 | $6.0m | $24.0m |
| Year | $16.0m |
The Q4 run rate overstates the year by half. The Q1 run rate understates it by a quarter. The trailing twelve months says $16m, and the growth from Q1 to Q3 says the next year will be higher than that. None of the four quarterly run rates is a forecast.
Seasonality. The quarter above. Use a trailing twelve, or compare the window to the same window last year and scale the prior year.
One-off orders. A $900,000 project order in a three-month window adds $3.6m to the annualised figure and will not repeat. Exclude orders flagged as one-off, or use a longer window, and state which.
A short window. One month of a lumpy account is noise. The window has to hold several of the account's normal orders.
A trend. A business growing 5 percent a quarter has a run rate that is already behind. Run rate is flat by construction; if the trend matters, show it beside the run rate rather than baking it in.
Per account, with the same window for all, run rate is a value:
| Use | Ranked by |
|---|---|
| Dormant list | Run rate in the window before the account went quiet |
| Growth list | Change in run rate, this window against the prior |
| Concentration | Run rate, so one large month does not distort the share |
| Gap valuation | Norm minus run rate, so the gap is against a stable actual |
Ranking is robust to the four problems in a way that a total is not: an account's seasonality affects its rank only against accounts with different seasonality, and the window is the same for all.
Σ accounts' window revenue = ledger revenue in the window
Run rate per account rolls up to run rate for the company only if the account revenue sums to the ledger. Check it first.
Window unstated. Two people quote two run rates and both are right.
Q4 annualised in January. The budget that follows is a third too high.
Run rate presented as forecast. The pipeline is never asked about.
Per-account windows differ. The ranking compares a one-month figure to a six-month one.
Mapped once, the ledger produces run rate per account, per rep and per company at a stated window every month, with the one-off flags and the trend beside it. Covirage builds this from the export as it is. The run rate term has the short definition, and the dormant threshold guide covers the list that run rate most often ranks.
Long enough to be stable, short enough to be current. Three months is common for monthly-ordering accounts; for a lumpy business, six or twelve. The window is stated with the figure and the same for every account in a comparison.
No. It is the present scaled up. A forecast adds pipeline, seasonality and known changes. Run rate is the floor that a forecast should be compared to, and the difference is what the forecast is claiming.
As a value for ranking. The dormant list ranked by run rate before the account went quiet; the growth list by the change in run rate; the concentration figure by run rate rather than a single month. It is more useful as a ranking than as a total.