A sales forecast you can defend: chance to win, stages and honest numbers

Ask most sales managers where their monthly number comes from and, if they are honest, the answer is: they looked at the pipeline, thought about it, and said a number. Then they adjusted it down slightly because they have been wrong in that direction before.

That is not forecasting. It is estimation with extra steps, and its main weakness is that it cannot be argued with — there is nothing in it to check, so when it is wrong nobody learns anything.

A pipeline forecast is not more accurate than an experienced manager’s instinct. What it is, is inspectable. Every part of it can be questioned, and when it misses you can find out which part was wrong. That is the whole value.

The mechanics, which are simpler than they sound

Every stage in your pipeline carries a chance to win: a rough percentage saying how often a lead sitting in that stage eventually closes. Multiply the leads in each stage by that stage’s percentage, add it up, and you have a forecast.

LumaKol states it plainly in Reports: about so many of your open leads should close, based on each stage’s chance to win. Ten leads in a stage you rated 20% contribute two. Four in a stage rated 60% contribute two and a bit. Nothing more sophisticated is happening, and nothing more sophisticated needs to.

Two details matter. Stages are typed as open, won or lost, and won and lost count as closed — so they are not in the forecast, they are the thing the forecast is trying to predict. And the number is a count of leads, not money, which we will come back to.

Setting the percentages: use your own history, not a template

The default instinct is to spread percentages evenly up the pipeline — 20, 40, 60, 80 — because it looks tidy. Real pipelines are almost never tidy.

Get the real numbers instead. For each stage, look back over a quarter and ask: of the leads that reached this stage, what share eventually reached a won stage? That is the chance to win. It takes an afternoon once and it is worth more than every refinement you will make afterwards.

When you do this, expect to be surprised in a specific way: there is almost always one stage where the number collapses. A team that assumed a smooth ramp finds that “quotation sent” converts at 15% while “site visit done” converts at 55%. That single fact usually changes where the manager spends their week, and it is worth more than the forecast it was calculated for.

That table is an example of the shape, not a benchmark — your numbers will look nothing like it, and if you copy them you have simply replaced a guess with a stranger’s guess. The pattern to look for is the row where the count falls off a cliff: between Interested and Site visit booked, this team loses more than half of everyone who said yes. That is a follow-up problem, and no amount of forecasting fixes it.

Four ways this forecast will lie to you

1. It has no sense of time

A stage probability says whether a lead will close, not when. A lead that has sat in “Interested” for four months still contributes its full 25% to the forecast, exactly like one that entered yesterday.

This is the biggest single distortion in any stage-based forecast, and the fix is not mathematical, it is hygiene: a lead nobody has touched in sixty days is not a 25% chance, it is a lost lead that has not been marked lost. The by-agent report shows you overdue follow-ups and how many leads each person touched in the last seven days; a pipeline where those numbers are bad is a pipeline whose forecast is fiction.

2. It counts leads, not money

The forecast tells you how many leads should close. It does not know that one of them is worth forty times another, because it does not know what any of them are worth.

If deal sizes in your business vary a lot, a lead count is close to meaningless on its own. The honest workaround is to keep a currency custom field — budget, or deal value — on every lead, and read the forecast alongside the median value of your open pipeline rather than multiplying by an average. Averages are dragged upwards by the one enormous deal everybody is already thinking about.

If you send quotes from LumaKol, part of this is done for you: when a customer accepts a quote, or it is marked paid, its total becomes the lead’s deal value. Reports then shows the open pipeline value of those leads, a money forecast weighted by each stage’s chance to win, and what was collected in the week. It only covers leads with a quote, so treat it as a second view beside the lead-count forecast rather than a replacement.

3. Agents move leads for reasons that are not the customer’s

The moment a stage has a probability attached, moving a lead into it becomes a claim about the future, and people respond to that. Some agents park everything in a middle stage to look busy; others refuse to advance anything until it is nearly signed, and their forecast is permanently pessimistic.

The defence is definitional, not managerial. Each stage carries a one-line description that agents see when they move a lead into it, and that line should describe an observable event, not a feeling. “Site visit booked, awaiting the customer’s decision” can be checked. “Warm” cannot. When the stage meaning is an event, the forecast stops being a mood.

4. It is only as current as your call outcomes

All of this rests on leads being in the right stage, which rests on agents logging what happened after calls. If outcomes are skipped, the pipeline is a photograph of some earlier week.

This is the argument for wiring stage movement into the outcome itself: an outcome can move the lead to a stage and schedule the next follow-up automatically, so the pipeline updates as a side effect of the agent doing the one thing they were always going to do. Where feedback matters most, you can require it before the next call can start.

Reading the forecast like a manager

The forecast number itself is the least interesting output. Three things around it are worth more:

• The shape of the funnel. Leads sitting in each stage right now, with the stage you can click into. A pipeline that is enormous at the top and empty in the middle is a contact problem; one that is fat in the middle and thin at the end is a closing problem. These need opposite interventions.
• Overdue follow-ups per agent. This is the leading indicator. It moves weeks before the forecast does, and unlike the forecast it names the person and the day.
• The gap between agents on the same stage. When one agent converts site visits at twice the rate of another, you have found something a forecast can never tell you, and it is usually fixable in a fortnight.

Then do the only thing that makes a forecast better: write down this month’s number and compare it with what actually happened. Not to grade yourself — to find out which stage’s percentage is wrong. One quarter of that and your forecast will beat instinct. Skip it and you have built a spreadsheet that produces a plausible number every month and never learns.