A forecast you can finally trust.
It knows why last year looked like that. Forecaster builds a model of your business from your connected history, the rules your people have taught Metis, and the plan you give it: a growth goal, a what-if, or the deck you already have. Every assumption is written down and correctable. Then it tracks the business against the forecast and tells you the moment something drifts.
No credit card. The trial starts at your first answer, not the day you sign up.
Forecaster 2027
or upload your plan (deck, PDF or spreadsheet) · or try a what-if
Building your model
Assumptions the model is making
Confirm or change. Each one becomes a rule.
Forecaster 2027 · tracking since Jan 1
Drift · Apr 1
March is 3.1% under forecast.
Service is on plan. The gap is hardware: the campaign started Mar 22 in Google Ads, not Mar 1 as the plan assumed, so March bookings will land in April.
Type the goal. Confirm the assumptions. Watch reality tracked against the forecast, with the reason attached when it drifts.
How the forecast gets made today
Most companies forecast one of three ways. All three miss the same thing: the reasons the history looks the way it does, which live in your people's heads and not in the numbers.
A spreadsheet, a growth rate, twelve equal months. It ignores the seasonality, the renewal cycle and the campaign that starts in March.
Knows your business
No
Thirty tabs built by one person. Right when they built it, stale by the second month, and nobody else can change an assumption without breaking it.
Updated
Rarely
A projection of the last 24 months with a confidence cone. It cannot know that half your contracts renew every other year, so it forecasts the dip as a decline.
Explains the why
Never
A trend line doesn't know your business. Your people do.
Found in the history
The cycle
A training company's own forecast missed that half its customers renew every other year.
Forecaster found the pattern in two years of history and asked. The owner confirmed it in one click, it became a rule, and the forecast, the dashboards and every answer anyone asks now carry it. The model is built from your history, the rules your people have taught Metis, and the plan you give it, with every assumption written down where you can change it.
No probability theater. Assumptions you can read, correct and own.
How it works
Give it the plan
A goal in a sentence ("grow 20%"), a what-if ("close rate up three points"), or the plan you already have: upload the deck, PDF or spreadsheet and Forecaster reads it. Your connected systems supply the history.
It builds the model and shows its work
Forecaster builds a model from your history, the rules your people have already taught Metis, and the plan. Every assumption it makes is listed, including cycles it found in the history that nobody mentioned. Confirm or change each one; it becomes a rule.
It tracks reality against the forecast
Weekly, monthly, quarterly: the business against the forecast and the goal. When something drifts, Forecaster says which assumption broke and what it would take to get back on plan. Dashboard tiles compare to forecast, not just to last year.
Forecasts from the tools you already pay for
- QuickBooks
- Xero
- Stripe
- Square
- Google Ads
- Meta Ads
- Google Analytics
- Search Console
- HubSpot
- Salesforce
- Shopify
- ServiceTitan
- Jobber
- CallRail
- Google Sheets
- Anything else: drop an export and it counts
Every assumption written down. None of them hidden in a cell.
A forecast is only as good as its assumptions, and in most companies nobody can list them. Forecaster puts them on one page, with who confirmed each one and when.
Assumptions are rules. The same rules that drive every answer, dashboard and report drive the forecast, so the forecast never disagrees with the numbers finance quotes. Change a rule and the forecast updates; find a cycle in Forecaster and it becomes a rule everywhere else.
There are no confidence cones and no probabilities to argue about. There is a model, its assumptions, and the gap between the forecast and what happened, with the reason attached.
What people forecast first
By month, by line, by unit
Service and hardware, locations and brands, each on its own seasonality and rules, rolled up to the company.
When the money actually lands
Built on your real terms and payment timing, not invoice dates. Net 90 customers land in the quarter they land.
From leads to closed
Ad spend to leads to customers at your real conversion rates, so a marketing plan becomes a revenue forecast.
Goal, what-if, or your deck
Hand it the board plan and it tells you what the numbers would have to do, and whether the history says they can.
Questions Forecaster answers
Type next year's goal. See what the history says.
Connect the systems the numbers live in, give Forecaster a goal or your plan, and read the assumptions it makes about your business. Correct the ones that are wrong and you have a forecast your people actually believe. Included on every plan; free to start, and the trial clock runs from your first answer.
Questions people ask first
How is this different from a trend line or a spreadsheet forecast?
A trend line extends the past. A spreadsheet applies a growth rate. Neither knows why the past looked the way it did. Forecaster builds its model from your history, the rules your people have taught Metis about how the business works (seasonality, renewal cycles, payment terms, what counts as a sale), and the plan you give it. Every assumption is listed so you can correct it, and the model finds patterns in the history that nobody wrote down.
Does it use probabilities or confidence intervals?
No. A confidence cone tells you the model is unsure without telling you why. Forecaster gives you one forecast, the assumptions behind it, and the gap between forecast and actual as it happens, with the assumption that broke named. If an assumption is wrong, you change it and the forecast changes.
Can I upload the plan we already have?
Yes. Drop in the deck, PDF or spreadsheet and Forecaster reads the targets and the initiatives out of it. It then tells you what the numbers would have to do to hit them, and whether your history supports that.
How far ahead does it forecast?
As far as your history supports. Most companies forecast the coming year, which needs at least two years of history for seasonality and cycles to show. More history on your plan means a more confident model, and Forecaster tells you when it is working from too little.
What happens when reality drifts from the forecast?
Forecaster checks the business against the forecast on the schedule you choose, weekly, monthly or quarterly. When something drifts, it says which assumption broke, how much of the gap it explains, and what it would take to get back on plan. Dashboard tiles can compare to the forecast instead of to last year.
Do forecasts count as queries?
No. Building and updating a forecast does not count against your monthly queries on any plan. Asking Metis a question about the forecast counts like any other question.
Who owns the data?
You do. Connections are read-only, each business sits in its own isolated database, you can delete any source's data or export every rule and forecast in one click, and we never train models on your data.
What does it cost?
Forecaster is included on every plan. Plans start at $75 a month for the whole company, month to month, no contract, with a 14-day free trial that begins at your first answer. See pricing.