It is the last week of the quarter. The forecast said four big deals would close. One closes. The other three slide into next quarter, just like last quarter. Management asks what went wrong. The honest answer: the forecast mostly measured how optimistic the sales reps were. In this article you read how to build a sales forecast that is right, with an honest pipeline, a fixed pipeline review and win probabilities from your own numbers.
Why a sales forecast is wrong
Most forecasts are built on two things: the phase in the CRM and a win probability per phase. Both are filled in by the sales rep. Without a fixed definition of when a deal is in a phase, that is an estimate. And whoever has a target rarely estimates low.
On top of that, deals rarely drop out in one go. They stay put. A four-month-old quote still counts, because nobody dares to mark it as lost.
That is not unwillingness. A sales rep who marks a deal as lost also loses a bit of hope. And in many teams an empty pipeline feels like a bad conversation with the manager. So the deal stays.
Three rules for an honest pipeline
- A phase only counts when the exit criterion has been met. A deal is in "proposal" when the customer has discussed the quote with you, not when it has been sent.
- Every deal has a next step with a date. A deal without an agreed next step is standing still, however good the conversation felt.
- A deal that sits in a phase too long gets discussed. Set per phase what too long means. For example: a quote with no response after six weeks goes on the list for the pipeline review.
The pipeline review
A forecast only becomes reliable when you walk through the pipeline together regularly. Not all deals, but the deals that matter. A fixed setup of half an hour per sales rep, once every two weeks:
- The deals for this quarter. One question per deal: what does the customer still have to do before he signs? If the sales rep cannot answer that, the deal does not belong in this quarter.
- The deals standing still. What is the next step, and who takes it? If there is none, mark the deal as lost with a reason. That is not failure, it is information.
- What is shifting. Which deals moved from this quarter to the next, and why? If the same reason keeps coming back, that says something about your process.
In the review, ask questions about what the customer does, not about what the sales rep thinks. "How sure are you?" produces a feeling. "When did the customer say he will decide?" produces a fact.
A few questions that work in every review:
- Who decides? And has the sales rep spoken to that person himself?
- What happens if the customer does nothing? If there is no reason to choose now, the deal usually slides.
- What is the agreed next step? With a date, and who takes it.
- What could still stop this? Budget, a colleague who still has to agree, a competitor.
Win probabilities from your own numbers
Do not use standard percentages like 20, 50 and 80 percent. Look back at the past year: how many deals that entered "proposal" were eventually won? That is your own percentage. It is often lower than you thought, and exactly for that reason more reliable.
A worked example: of a hundred deals that entered the proposal phase last year, thirty were won. If there are now twenty deals in proposal worth 400,000 euros together, an expectation of 120,000 euros is more realistic than the sum of what sales reps hope.
If you have little history, start with what you have. Six months of numbers beats a standard percentage from a book. Recalculate it every quarter. After a year you have a number you can build on.
Building a forecast in five steps
- Set the phases and exit criteria. What has to have happened before a deal may move to the next phase? Write it down in terms of what the customer does.
- Clean up the pipeline. Go through all open deals. No next step and no response in months? Mark them as lost, with a reason.
- Calculate your conversion per phase. Look back over the past year. How many deals went from each phase to won?
- Set a maximum duration per phase. A deal that sits longer goes on the review list.
- Schedule the pipeline review. Fixed moment, fixed setup, once every two weeks. Keep it short.
How the forecast connects to your sales process
A forecast is only as good as the process underneath it. The phases and exit criteria come from your sales process. The time a sales rep spends on follow-up comes from your activity matrix, for example the agreements on quote follow-up. The pipeline review belongs in your meeting structure, with a fixed frequency and a fixed agenda.
That is why it helps not to improve the forecast in isolation. If a phase is wrong, the conversion is wrong too. And if many deals keep dropping out in one phase, that is a question for the process, not only for the sales rep. Sometimes the answer is a better exit criterion, sometimes an extra contact moment.
Common mistakes in forecasting
- Mixing up forecast and target. The target is what you want. The forecast is what you expect. Whoever sets the forecast equal to the target only sees a shortfall when it is too late.
- Looking only at the total amount. A big pipeline says little if half of it has been standing still for months. Also look at age and next steps.
- The review as an interrogation. Sales reps who have to defend themselves give socially desirable answers. Ask about the customer, not about the person.
- Not analyzing lost deals. The reason for a loss is the best source for improving your process. Record it, every time.
- Treating the CRM as the truth. The CRM is only as good as what is in it. Without exit criteria it mostly measures how the team interprets the phases.
What it delivers
A forecast that is right brings calm. You know earlier when a quarter is going to fall short and can still adjust. And the pipeline review becomes a coaching moment instead of an interrogation, because the conversation is about concrete steps by the customer.
The conversation with management changes too. You can explain what a number is based on. And if it falls short, you can show where it went wrong: too little inflow, too long a cycle or a phase where many deals drop out.
Frequently asked questions
How do you build a reliable sales forecast?
With fixed exit criteria per phase, a next step with a date for every deal, win probabilities from your own history and a fixed pipeline review. Together they make the difference between hope and expectation.
How often do you run a pipeline review?
Once every two weeks per sales rep works for many teams. Half an hour is enough if you limit yourself to the deals for this quarter and the deals standing still.
Which win probabilities do you use per phase?
Your own. Look at how many deals from each phase were won in the past year. Standard percentages say nothing about your customers and your process.
When do you mark a deal as lost?
When there is no agreed next step left and the customer does not respond within the period you have set per phase. Always record the reason.
Capture the conversion per phase and the norm for how long a deal may sit in a phase in KPIs per phase of building layer 7, with the CRM as data source and an owner per KPI. The exit criteria belong with Phases in building layer 2, the pipeline review itself in the Meeting structure of building layer 5. Because the KPIs build on your phases, a new exit criterion shows you straight away which norms you need to look at again.