Revenue forecasting for SMEs

For an SME, a revenue forecast works best as a simple structure in three layers: fixed recurring revenue from contracts and the accounts, recurring purchasing by existing customers based on history, and new revenue from signed orders plus a cautious estimate of the pipeline. That fits in a single spreadsheet, takes an hour a month to maintain and is often more accurate than an elaborate model built on a half-completed CRM. What matters most is not the method, but comparing it every month with what actually happened.

Why forecasting is different in an SME

A business with EUR 2 to 20 million in revenue rarely has a finance team updating the forecast every week. The managing director or the person responsible for finance does it alongside everything else, usually in Excel, often just before a meeting with the bank or the shareholders. A few other things also weigh more heavily than in larger companies.

• Few deals, big differences. Twenty new customers a year is too few for statistics. One deal more or less makes the quarter.
• Concentration. The largest customer is often 10 to 20 percent of revenue, sometimes more. What that customer does determines the year.
• Sales lives in a few heads. The managing director or two account managers know which deals are real. Only part of that is in the CRM.
• The accounts lag behind. If the month-end close takes three weeks, the reality you forecast against is already old by the time you see it.
• Much revenue sits outside the CRM. Regular customers order by email or phone, directly into the ERP or accounting system.

A forecasting method imported from a large software company, with stage percentages and pipeline coverage per salesperson, is a poor fit for that. What does fit is a structure that starts with what is certain.

What are the layers of an SME revenue forecast?

Layer 1: fixed recurring revenue

Maintenance contracts, subscriptions, licences, rent, service contracts. This sits in your contracts and your invoicing, in Xero, Exact, AFAS, Twinfield, Moneybird, Dynamics or whatever you use. For each contract: amount, frequency, end date, indexation clause. You can forecast this layer almost exactly, apart from two risks: contracts that expire and indexations that are not applied.

Layer 2: recurring purchasing by existing customers

Customers without a contract who nonetheless order regularly. Here you work from history: what did this customer or customer group buy per quarter over the past two to three years, and is there a seasonal pattern? How to build that is covered in forecasting from historical revenue. For an SME it is often enough to take the average of the same period over the past two years, corrected for customers who have left or clearly grown.

Layer 3: new revenue

Here you split into two:

• Signed, not yet delivered. Orders and assignments that are agreed. This is almost certain revenue; only the timing can move.
• Not yet signed. The pipeline. Include only deals for which you can name the customer and a realistic date, and apply a cautious percentage. If you have little history, use the average new revenue per quarter over the past two years as a ceiling.

If your CRM is incomplete, and in an SME it almost always is, the approach in forecasting with incomplete CRM data helps.

Worked example

Worked example: suppose you run a technical services business with EUR 4 million in revenue. You are building the forecast for the fourth quarter.

At a glance you see that almost 80 percent of the quarter comes from the first two layers. The discussion about the five new deals matters, but concerns one euro in seven. If the forecast misses, it is more likely down to a departed customer in layer 2 or an unapplied indexation in layer 1 than to the pipeline.

The one-hour monthly routine

A forecast is only useful if it is maintained. This rhythm is achievable for most SMEs.

1. Fill in actuals (15 minutes). Put last month's invoiced revenue per layer next to the forecast. If the books have not been closed yet, use the invoices sent.
2. Explain the differences (15 minutes). Per layer: why were you off? One customer fewer, a project slipped, a deal came in early? Write it down in one sentence.
3. Review contracts (10 minutes). Which contracts expire in the next six months? Has this year's indexation been applied everywhere?
4. Walk through the pipeline (15 minutes). Together with whoever does sales: which deals are still real, which are slipping, which are dead?
5. Update the forecast (5 minutes). Adjust the coming months and keep the old version.

That last point matters. By keeping the previous version each month, after a year you see how good your forecast was and in which direction you were off. That is the basis for an honest range.

Where it goes wrong in SMEs

• The forecast is only made when the bank asks for it. Then it is a presentation, not a management tool.
• Only the total, no layers. When the forecast misses, you do not know where.
• Too much faith in the pipeline. In small businesses the salespeople are often also the owner, and owners are optimistic about their own deals.
• The largest customer as a fixed value. Treat that customer as a scenario, not a certainty. How to model that is covered in scenario forecasting for B2B.
• Leakage in the base. Forgotten indexations, uninvoiced additional work and contracts that have quietly expired make layer 1 smaller than it should be. The forecast may then be accurate, but the amount is too low.

When a spreadsheet is no longer enough

For most SMEs a spreadsheet is a perfectly good forecasting tool. It starts to pinch when:

• you have more than a few hundred contracts or customers per layer and want to forecast per customer;
• the forecast has to be updated every week, not every month;
• you have to combine data from three or more systems by hand and errors creep in every time;
• several people work on the same file at the same time.

At that point a link between your CRM, your accounts and your forecast pays off, or a tool that does this for you. Whether that makes sense for your business is worked out in can an SME use Revenue Intelligence. Whether AI then delivers more than a good spreadsheet depends mainly on the amount of data; see AI forecasting vs traditional forecasting.

For the full framework around forecasting, see what is revenue forecasting.

Checklist

• Do you have a list of all recurring contracts with amount, end date and indexation clause?
• Has this year's indexation been applied for every customer?
• Do you know the revenue per quarter over the past two years for each customer or customer group?
• Are signed orders kept separate from deals that are not yet agreed?
• Do you keep last month's forecast every month?
• Do you know what happens to the year if your largest customer halves?

Frequently asked questions

How far ahead should an SME forecast?

Twelve months rolling is common: each month the horizon moves on by a month. Detail the first quarter by month and the rest by quarter. Further ahead is mainly useful for investment decisions, with scenarios rather than a single number.

Do you need a CRM for a good forecast?

No. For the recurring layers you only need your invoicing and contracts. For new revenue a CRM helps, but for small teams a list of the deals that are genuinely in progress, maintained by whoever does sales, also works.

How accurate can an SME forecast be?

That depends mainly on how much revenue is recurring. A business with many contracts can often get close to reality per quarter. A business that does a handful of large projects a year cannot, and is better served by scenarios.

Is a revenue forecast the same as a cash flow forecast?

No. A revenue forecast predicts when revenue arises. A cash flow forecast predicts when money comes in and goes out. You need the revenue forecast to build the cash flow forecast, with your actual payment periods in between.