Subscribe now

The problem with an unreliable forecast is that it can look convincing until reality proves it wrong.

There is a familiar sequence that plays out in B2B companies at the end of a bad quarter.
A few weeks earlier, the forecast looked reasonable. There was enough pipeline to support the target. The largest opportunities were progressing. The sales team felt confident about the deals expected to close. Leadership might even have made decisions on the back of that number: hiring, spending, cash planning or targets for the following quarter.

Then the quarter ends and the number is wrong.

A $1.2M forecast becomes $800K in actual revenue. Deals that were supposed to close in September suddenly have October dates. An opportunity everyone had been talking about for months quietly disappears. Another closes at half the expected value.

The post-mortem normally starts with the forecast itself.

Were the probabilities wrong? Were the salespeople too optimistic? Should managers have challenged the deals more aggressively? Do we need a better forecasting tool? Should we introduce another category between “best case” and “commit”?

Those can all be reasonable questions. But they start too late.

If you are asking why your sales forecast is always wrong, the more useful place to look is usually the pipeline that produced it.

A forecast does not exist independently from your revenue operation. It is the final output of dozens of earlier decisions: what your company considers an opportunity, when a deal moves stage, how its value is calculated, why a close date is entered, when an inactive opportunity is removed, and what evidence is required before a salesperson can say that something is likely to close.

If those inputs are inconsistent, the forecast can be mathematically perfect and commercially useless.

That is why improving forecast accuracy often has less to do with becoming better at predicting the future and more to do with becoming better at describing what is happening now.

Your forecast is only as good as your pipeline data

It is useful to think of the forecast as the last step in a much longer process. Your pipeline data shows the opportunities the business believes are active and progressing. The forecast then uses that information to estimate how much revenue is likely to close within a particular period.

So if the pipeline data is unreliable, the forecast will be too.

Imagine a business with $2M of open opportunities against a $600K quarterly target. On the surface, that looks healthy. But the number only matters if the opportunities behind it are credible. They need to be genuinely qualified, sitting in the right stages, carrying realistic values, attached to defensible close dates and, most importantly, still connected to an active buying process.

This is where forecasting can create a false sense of precision. A $100K opportunity sitting in a stage with a 60% probability contributes $60K to the weighted pipeline. The maths is perfectly correct, but that does not make the output reliable.

If the opportunity should not be in that stage, the probability is wrong. If the buyer has not agreed the scope, the deal value may be wrong. If the close date exists only because the CRM required one, the timing may be wrong too.

The calculation can be perfectly sound while the data feeding it is weak.

This is also why adding sophistication at the forecasting layer rarely fixes a weak revenue system. More categories, AI scoring, stricter calls and more elaborate spreadsheets can improve the way information is analysed, but they cannot compensate for pipeline data that gives an inconsistent picture of what is actually happening.

Before changing the forecasting methodology, you need confidence that your pipeline data reflects commercial reality: real opportunities, meaningful progress, credible timing and active buying processes.

Five pipeline problems that make your forecast less reliable

Most forecast problems can be traced back to decisions made much earlier in the sales process. The exact symptoms vary from company to company, but five issues appear repeatedly: when a prospect becomes an opportunity, how deals move through stages, how probabilities are assigned, how close dates are set, and how long inactive opportunities are allowed to remain open.

1. Prospects become opportunities too early

A prospect replies to an email, agrees to a call or says the product sounds interesting. One salesperson creates an opportunity immediately. Another salesperson on the same team waits until the buyer has confirmed a real problem, discussed a credible project and agreed to a next step.

Both eventually show up in the CRM as opportunities, even though the level of buying intent behind them can be completely different.

At a small scale, this may not cause much damage. Across an entire sales team, however, it can significantly overstate how much genuine pipeline the business has. The problem becomes particularly visible when leadership starts using pipeline coverage as reassurance.

Suppose the quarterly target is $500K and the CRM contains $2M of pipeline. Four times coverage sounds comfortable. But if a meaningful portion of that $2M came from prospects that became opportunities before they were properly qualified, the ratio tells you very little about how much revenue is realistically available to close.

There is no universal definition of a qualified opportunity. A $250K enterprise software deal and a £10K advisory project should not necessarily follow identical qualification criteria. What matters is that your business has a clear definition and that two salespeople faced with the same situation would reach roughly the same conclusion.

The question to answer is simple: what needs to be true before we are willing to call this a real opportunity?

If that definition is loose, forecast error enters the system from the moment those opportunities start contributing to the pipeline.

2. Your stages are based on seller actions

The next problem usually appears in the way opportunities progress.

Many CRM pipelines are built around sales activities: Discovery Completed, Demo Delivered, Proposal Sent, Negotiation, Contract Sent. These stages are convenient because they are easy for the seller to identify. The rep knows whether they sent a proposal or completed a demo.

The weakness is that seller activity often tells you very little about whether the buyer has genuinely progressed.

Consider two opportunities sitting in “Proposal Sent.” In the first, the buyer has confirmed the scope and budget, involved the economic decision-maker and booked a meeting to review the proposal on Friday. In the second, the prospect finished a call by saying, “Send me something and I’ll take a look.”

The CRM can show both deals in exactly the same stage. If that stage carries a 60% probability, the forecasting model may treat them almost identically. In reality, they are at very different points in the buying process.

A stronger pipeline uses observable buyer evidence to define progression. The buyer disclosed budget. The right stakeholder joined the process. Scope was agreed. A proposal review was booked. Procurement started. A decision date was confirmed.

The goal is to make stage movement depend on evidence the team can see and verify. Complex B2B sales will always require judgement, but clear criteria give that judgement boundaries.

When stages are tied to specific buyer actions, similar deals are much more likely to be classified in the same way across the team. Leadership can then look at a stage and have a clearer idea of what has actually happened in the buying process, which makes that pipeline data far more useful for forecasting.

3. Your probabilities are based on rep judgement

Even with well-defined stages, the forecast can still go wrong if the probabilities attached to opportunities have no connection to what has historically happened.

A salesperson looks at a deal and says they are 80% confident. Another rep might look at a similar deal and call it 50%. Both may have perfectly reasonable arguments, but the number is still based largely on personal judgement.

This becomes especially problematic when those percentages are used to calculate weighted pipeline. A $200K opportunity at 80% contributes $160K to the forecast. Change the rep’s confidence to 50% and that contribution drops to $100K. Sixty thousand dollars has disappeared from the forecast without anything changing on the buyer’s side.

Historical conversion rates give you a much stronger starting point.

If only 35% of opportunities that reach a certain stage historically become Closed Won, that tells you something important about what reaching that stage has actually meant in your business. The same applies when you break conversion down by segment, deal size, product or sales motion where there is enough data to do so.

That does not mean every deal in the same stage has exactly the same chance of closing. Individual opportunities will always have different circumstances. Historical conversion gives you a baseline grounded in what has actually happened, while deal-specific evidence helps you understand where an opportunity may sit relative to that baseline.

The key is that probability should have a clear basis. When the number comes mainly from a rep’s confidence, the forecast starts measuring optimism alongside commercial reality.

4. Your close dates are often based on assumptions

The expected close date has enormous influence over a forecast, yet it is often one of the least defensible fields in the CRM.

Take the largest deals expected to close this quarter and ask why each one carries that specific date. Sometimes there is a clear answer. The customer’s existing contract expires at the end of October. Their investment committee meets on the 18th. Procurement needs the agreement completed before the new financial year. The project needs to launch in January and implementation takes six weeks.

Those dates are tied to the buyer’s process. They have commercial meaning.

In other cases, the answer is much weaker. The rep expects the deal to close this month. They want it in the quarter. Or they selected 30 September when the opportunity was created because the CRM required a date.

Then September arrives and the deal does not close, so the date moves to October. If October passes, it moves again to November. Nothing meaningful may have changed in the buying process, yet the same opportunity continues to appear in successive forecasts as revenue that is expected to arrive soon.

Over time, this creates a misleading picture of timing. Leadership keeps seeing the same revenue one month ahead, while the evidence behind the close date gets weaker with every move.

Repeated close-date movement is a useful revenue signal. It tells you either that the organisation does not understand the buyer’s decision process or that something in that process is not progressing as expected.

A date entered in a CRM only becomes useful when there is a commercial reason behind it.

5. Dead opportunities remain alive for too long

Another common problem is stale pipeline: opportunities that remain open in the CRM even though there is little evidence that the buyer is still actively moving towards a purchase.

Imagine a business showing 40 open opportunities worth $1.2M against a $400K target. At first glance, that looks like plenty of pipeline.

Then you review the deals properly. Fifteen have recent buyer activity and clear next steps. Eight have received no response for three weeks. Six were supposed to make a decision last month, but nobody knows what happened. Four have no future meeting scheduled. Several others remain open because the salesperson hopes the account will eventually re-engage.

Once you separate the active buying processes from the stale opportunities, the picture changes completely. The CRM may show $1.2M of pipeline, while only $450K is attached to buyers who are currently engaged and moving forward.

That gap changes how you should read your pipeline coverage. Against a $400K target, $1.2M looks comfortable. But if only $450K is genuinely active, the business is much closer to the target than the headline number suggests, with very little room for deals to slip or be lost.

The danger is that stale opportunities can hide that gap for weeks or even months. As long as those deals remain open, leadership may believe there is enough pipeline and hold off on taking action. When it finally becomes clear that those opportunities are unlikely to close, there may be much less time left in the quarter to generate new pipeline or recover the gap.

This is why every sales process needs clear criteria for keeping an opportunity open. A deal might remain active because the buyer has engaged recently, there is an agreed next step, a future meeting is booked, or there is a clear reason why the process has temporarily paused. If none of those conditions is true for a period that makes sense for your sales cycle, the deal should be reviewed and either taken out of the active pipeline or marked as lost.

Your forecast review should challenge the evidence behind each deal

Even with better pipeline data, the forecast still needs to be reviewed regularly. Deals change. Timelines move, stakeholders disappear, budgets get delayed and buying processes lose momentum. The purpose of the forecast review is to catch those changes early and make sure the number still reflects what is happening in the deals today.

That requires more than asking the rep whether they still feel confident. A salesperson can genuinely believe a deal will close because the conversations have been positive, the solution fits and the customer has shown interest. But confidence alone does not tell you whether the buying process has actually moved forward since the last review.

A useful forecast review should test the assumptions behind the deal. If it is expected to close this month, what has the buyer done that supports that timing? If it remains in a late stage, has the buyer completed the actions required to be there? If the deal is still considered likely to close, what has happened since the last review that supports that view?

For example, imagine a $150K opportunity that has been in the forecast for three weeks. The rep remains confident, but the procurement process that was supposed to start on Monday has not started, the next meeting has not been booked, and the buyer has not replied to the last two emails.

At that point, the forecast should change. Depending on the evidence, the deal may need to move out of the current forecast category, have its expected close date pushed back, or be removed from the forecast altogether until the buying process starts moving again. The rep may still believe the deal will eventually close, but the forecast should reflect what the buyer is doing now.

This is where a good forecast review adds value. It forces the team to compare what the CRM currently says with the latest evidence from the buyer. Deals with stronger evidence can stay where they are. Deals where the evidence has weakened should be challenged, moved or removed from the forecast before the gap appears at the end of the month.

Over time, this also creates a more consistent standard across the sales team. Reps know what evidence they are expected to bring when they call a deal likely to close, and managers have something concrete to challenge when that evidence is missing.

The result is a forecast that changes when the buying process changes, rather than weeks later when the expected revenue fails to arrive.

How to make your sales forecast more reliable

When forecasts miss repeatedly, companies often respond by tightening the forecasting process: more review meetings, more categories, more manager overrides and more scrutiny on individual deals. That can make the process feel more controlled without changing the quality of the pipeline behind it.

The more useful question is what would have needed to be different earlier in the sales process for the forecast to have been more accurate in the first place. That is where the biggest improvements usually begin.

Start with the definition of an opportunity. If prospects are entering the pipeline before there is a genuine buying process, every number built on top of that pipeline will already be overstated.

Then look at stage progression. The business needs clear criteria for what has to happen before an opportunity moves forward, ideally based on observable buyer actions. Without that consistency, two similar deals can sit in different stages depending on who owns them.

Once those foundations are clear, probability becomes much more useful. Historical conversion rates can tell you how often opportunities at each stage actually become customers, giving the forecast a baseline grounded in your own performance rather than rep confidence alone.

Close dates need the same discipline. Later-stage opportunities should have a clear reason for why they are expected to close when they are, based on the buyer’s process rather than the seller’s target or an arbitrary CRM date.

The pipeline also needs to stay current. Opportunities that no longer have buyer engagement, a credible next step or a clear reason to remain active should be reviewed so they do not continue inflating pipeline coverage.

Finally, the forecast review should keep testing whether those assumptions are still true. Deals change, so the stage, probability, close date and forecast category should change when the buyer evidence changes.

The point is to avoid spending time refining the forecast while the information feeding it is still unreliable. Once the pipeline gives you a more accurate picture of the opportunities in play, the forecast has a much stronger foundation to work from.

Some forecast misses are unavoidable

No forecast will be perfectly accurate. Buyers change their minds, budgets get frozen and internal decisions get delayed. Some uncertainty will always remain.

The important question is whether the miss came from something genuinely unexpected or from information that was already visible. A sudden budget freeze is difficult to anticipate. A deal that has already slipped twice, has no next meeting and has gone quiet should not surprise anyone when it misses again.

A more reliable forecast will not eliminate misses. It will reduce the ones the business should have seen coming.

How Revenue House looks at forecasting

At Revenue House, we treat forecast accuracy as an output of the wider revenue system rather than an isolated sales-management problem.

During a Revenue Engine Audit, we look upstream at how opportunities are qualified, what pipeline stages actually mean, how probabilities are set, what evidence moves a buyer forward, how close dates are established, when stale opportunities are removed, and whether leadership can trace the forecast back to something observable in the buying process.

Sometimes that work reveals a forecasting problem. More often, it reveals something earlier in the revenue system that was making an accurate forecast almost impossible in the first place.

The objective is to make sure that when the future does surprise you, it happens because something genuinely unpredictable changed — rather than because the pipeline was carrying assumptions that should have been challenged much earlier.