The short answer
Where deals take months, this month's closed revenue reflects advertising from months ago. Optimising against it means reacting to old decisions.
Use qualified enquiries as the working signal and validate against closed work over longer periods.
Measure in stages
- Enquiries, immediately.
- Qualified enquiries, within days.
- Proposals or quotes sent, within weeks.
- Closed work, over months.
- Check that the earlier stages predict the later ones.
Point five is the important discipline. If qualified enquiries stop predicting closed work, the qualification definition has drifted and needs revisiting.
Do not react to short periods
| Cycle length | Review period |
|---|---|
| Weeks | Monthly is reasonable |
| A few months | Quarterly |
| Longer | Twice yearly, with interim signals |
| Highly seasonal | Same period year on year |
| Very low volume | Longer still, expect noise |
Low volume is the harder constraint. With a small number of deals, most month to month variation is noise and acting on it is worse than doing nothing.
Keep the record long enough
Attribution has to survive the length of the cycle. If the source is lost when an enquiry moves into a sales system, you cannot connect closed work to advertising at all.
Carrying the source through to the outcome record is what makes long cycle measurement possible, and it has to be in place before the cycle starts.
Budget with the cycle in mind
- Expect a gap before revenue appears
- Do not cut during the gap, it is the design of the thing
- Plan the commitment for at least one full cycle
- Agree in advance what would justify stopping
- Keep interim signals visible so the gap is not silent
The last point is what keeps the budget. A period with no revenue and no reporting looks like failure; the same period with visible enquiry and proposal activity does not.