Forecasting a pipeline: spreadsheet, CRM report, or software?
Most sales teams should stay on a spreadsheet considerably longer than they do. A sheet does almost everything a forecast needs, costs nothing, and keeps the one thing that actually improves accuracy. It fails at exactly three things, and until you are hitting those, buying software mostly replaces something that was working with a subscription that does the same job.
I ran a pipeline forecast in Google Sheets for a year before building anything, so the sections below are not a pretext. The spreadsheet case is the strongest case here, and I will make it properly before getting to where it broke.
What a spreadsheet does well
Five columns carry a forecast: owner, deal, value, the owner's own odds, and close month. Add one computed column for value × odds and you have weighted pipeline — the same number your CRM prints, from data you trust more.
Then add the column almost nobody has: odds × (1 − odds) × value², sorted descending. That is each deal's share of the uncertainty in the quarter. In most pipelines three or four deals out of thirty carry most of it, and those are the only deals whose movement changes the answer. This is the single most useful thing on this page and it is one formula.
A spreadsheet is also the only option with no vendor's opinion baked in. Every tool encodes assumptions about what a stage means and what a probability is. In a sheet, the model is yours and you can see all of it.
And the habit that matters most works perfectly in a sheet: on the first of each month, duplicate the tab and freeze it. Ten seconds. That frozen copy is what lets you ask, later, whether you forecast high or low — and that question improves accuracy more than any tool does.
Where it genuinely breaks
1. It cannot give you a range. This is the real one. Weighted pipeline is a point estimate, and with few deals of uneven size the point sits in the gap between outcomes that can actually happen. Getting the distribution means simulating the quarter thousands of times, which a spreadsheet does badly and slowly.
Against a 300.000 € target. The first row is what a sheet gives you. The rest needs simulation.
The weighted figure says 366.000 €, comfortably above target. The odds of actually reaching target are 63%. Both come from the same six deals. The sheet can only tell you the first.
2. Analysing the frozen tabs is tedious enough that people stop. Keeping twelve monthly snapshots is easy. Answering "which rep's 70% actually closes 70% of the time" across those twelve tabs is an afternoon of manual work, every time you want to know. In practice this is where the discipline dies — not at the recording, at the reading.
3. It degrades past about fifty rows and with more than one editor. Not because the maths breaks. Because nobody scrolls, so the long tail stops being reviewed, and because two people editing during a meeting produces a version nobody trusts afterwards.
Why the CRM report is not the answer either
The obvious alternative is the forecast report you are already paying for. It has two problems that no amount of configuration fixes.
Its probabilities come from the stage, not the deal. "Negotiation = 70%" is a statement about your company's history applied flat to an opportunity it knows nothing about — and in many CRMs it is a number an administrator typed in at implementation and nobody has revisited.
And it reports a point estimate, so it inherits problem one above. A cleaner CRM gives you a more accurate version of a number that was the wrong shape to begin with. That is why teams who clean their CRM religiously still keep a shadow spreadsheet: the hygiene was never the thing that was wrong.
What the market actually offers
Three categories, which get mixed together in every comparison article and are not substitutes for each other. Prices below are each vendor's own published figures as of November 2026.
Revenue intelligence platforms — Clari, Gong and similar. They capture activity: calls, emails, meetings, CRM changes, and infer deal health from it. Genuinely valuable when you have sixty reps and cannot sit in every call. None of them publish pricing, which tells you the evaluation starts with a sales conversation. Clari is now part of Salesloft; its pricing page redirects there.
Forecasting and pipeline analytics — Pipemetry at $29–59 per seat per month, aimed at teams of roughly 10 to 100 reps, with CRM connection optional; Forecastio, HubSpot-only, pricing unpublished, offering weighted, time-series and AI methods. These analyse the pipeline data you already have.
Probabilistic forecasting — the small category that reports a range and odds rather than a number. Funnelcast is the most established, at $1,500–2,000 a month for unlimited users, requiring a Salesforce or HubSpot connection. The Pipe Review, which I build, is in this category: free for the forecast, 29 € a month for the accuracy history, no CRM connection, deals typed in during the review.
How to decide
- Under ~50 deals, no range needed, one person maintaining it: stay on the sheet. Add the uncertainty column and start freezing monthly tabs. This is the right answer far more often than the market would like.
- You need the range and the odds: a sheet cannot do it. That is the honest threshold for buying something in the third category.
- You have a year of frozen tabs and have never analysed them: this is the most common real trigger. The record exists and reading it by hand is the barrier.
- Your problem is that reps do not update anything: no arithmetic fixes missing data. You need activity capture, which is category one, and it will not be cheap.
- Over ~100 similar-sized deals: weighted pipeline is already a decent estimate. The range adds less, and analytics tooling will serve you better than a simulator.
Where I am biased
I build one of the products named above, so discount this accordingly. Two things I would say anyway.
The sheet is a better first step than any of us, including me. If you have not yet got the five columns and the monthly freeze, do that before evaluating anything — it costs nothing, and if you skip it you will buy a tool to solve a discipline problem and still not know whether you forecast high.
And the specific thing that should make you look past a sheet is not dissatisfaction or scale. It is one question you cannot currently answer: what are the odds we hit the number, and which deals decide it? If you do not need that answer, you do not need any of this.