Most businesses have a number in mind for next month's sales. Where that number comes from varies a great deal. Sometimes it's last year plus a bit. Sometimes it's what the owner needs to cover costs. Sometimes it's what the bank was told. And often one number is being used for three different jobs at once, and doing none of them well.
So let's start by pulling those jobs apart, and then build each one properly.
Forecast, target and budget are three different numbers
A forecast is your best honest estimate of what will happen. Its only job is to be accurate. A good forecast is one that turns out close to reality, whether reality is good news or bad.
A target is what you're asking the team to achieve. Its job is to direct effort. A good target is one that stretches people without breaking their belief that it's possible.
A budget is the plan the business commits to: the sales figure that the year's spending is built on. Its job is to be safe enough to spend against.
These three numbers should usually be different. The target is typically above the forecast, because it asks for extra effort. The budget is often at or a little below the forecast, because you don't want to commit spending to sales you're hoping for.
When one number does all three jobs, it gets pulled in conflicting directions. If the forecast is also the target, people shade it down to make it easy to hit. If the target is also the budget, the business spends money on sales that were never likely. Keeping them separate lets each one be honest.
Does that make sense? A lot of the frustration owners feel with forecasts comes from asking one number to be honest, motivating and safe at the same time. It's like using the same thermometer to measure the room, set the air conditioner and decide what to wear. Give each job its own number, and each one can do its job well.
Building a forecast from drivers
A driver is something that causes sales. A forecast built from drivers multiplies them together, instead of starting from a total and adding a percentage.
For a retail shop: footfall × the share of visitors who buy × average spend.
For a subscription or repeat business: customers at the start of the month − customers lost + new customers, × average monthly spend.
For a distributor: number of active accounts × average order per account per month.
The advantage is that each driver can be checked against your history, and each one can be reasoned about separately. "Sales will grow 20%" is a single guess. "We'll add four new accounts a month, average orders will stay flat, and we'll lose one account a quarter" is three things you can test, and when the forecast misses, you'll know which one was wrong.
Then add the shape of the year. Almost every business in Sri Lanka has a seasonal pattern: the April festival season, December and the year-end, school terms, tourist seasons, harvest cycles. Use the last two or three years to find your monthly pattern, and apply it to the drivers.
Forecasting from a pipeline
For businesses that sell through quotes, proposals or negotiations, B2B suppliers, contractors, agencies, there's a better source than history: the deals you're currently working on.
List every open opportunity with its value and the stage it's at. Then give each stage a probability, based on how often deals at that stage have closed for you in the past. Multiply each deal's value by its stage probability and add them up. That's a weighted pipeline.
For example (the probabilities here are an illustration; use your own history):
8 enquiries at LKR 500,000 each, 10% likely = LKR 400,000
4 proposals at LKR 600,000 each, 30% likely = LKR 720,000
2 deals in negotiation at LKR 750,000 each, 60% likely = LKR 900,000
1 verbal yes at LKR 900,000, 90% likely = LKR 810,000
Weighted pipeline: LKR 2,830,000
No single deal will close for exactly its weighted value. Each one either closes or doesn't. But across a reasonable number of deals, the weighted total is a far better forecast than adding up every deal that "should" close, which is a common habit, and tends to come out too high.
A weighted pipeline also tells you something a history-based forecast can't. If it's thin for the months ahead, you know now that sales will dip later, while there's still time to generate more enquiries.
Notice the chain this protects you from. An over-optimistic forecast leads to spending committed against sales that never come. That spending, with no sales behind it, shows up as a cash squeeze a few months later. And the squeeze gets blamed on bad luck, when it was set in motion the day the forecast was written.
Three scenarios
The future isn't a single number, and pretending it is produces false confidence. So forecast three versions: a low case, where the key drivers come in weaker than you expect; a base case, your honest best estimate; and a high case, where things go well.
The base case is what you plan around. The low case is what you check your cash against, because it tells you whether the business survives a bad quarter. The high case tells you what you'd need to be ready for, stock, staff, capacity, if things go well.
Setting targets that work
A target should come from the forecast, plus a stretch the team understands. "The forecast says 30 new accounts this quarter; the target is 36, because we're adding a second salesperson in month two." That's a target with a reason, and people can see the path to it.
Targets also work better when they're set on the things the team controls, as well as on the result. A salesperson controls how many calls they make, how many visits, how many proposals they send and how quickly they follow up. They don't fully control whether a customer buys this month. These activity measures are called leading indicators, because they move before sales do. Sales itself is a lagging indicator. Tracking both lets you see a problem in the activities weeks before it shows up in the sales.
When targets backfire
There's a well-known pattern with targets: when a number becomes the target, people find ways to hit the number that don't produce what the number was meant to measure. With sales targets this shows up in a few predictable ways.
Discounting at the end of the month, to close deals that would have closed anyway next month at full price. Pulling next month's sales into this month, which makes this month look good and next month look bad. Selling to customers who won't pay, or won't stay. Neglecting existing customers, because only new sales count towards the target.
This can happen with an entirely honest team. It's what happens when a target rewards one number and ignores the others. The fixes are to measure the things the target might be traded against: margin (the share of each sale left after paying for the goods) alongside revenue, payment alongside sales, and existing accounts alongside new ones. And to review the pattern of sales across months, so pull-forward shows up.
Review every month
A forecast is useful only if it's compared with what happened. At the end of each month, put the forecast next to the actuals, driver by driver. Where did it miss, and why? Was it footfall, conversion or spend? Was it a deal that slipped, or a stage probability that's wrong?
Over a few months, two things happen. Your forecasts get more accurate, because you learn which drivers you consistently over- or under-estimate. And your understanding of the business gets sharper, because forecasting forces you to say what causes your sales, and the misses show you where you were wrong.
The bigger point
Owners tend to relate to forecasts in one of three ways. The first treats the forecast as the number they want, which is a hope with a spreadsheet around it. The second, after a few forecasts have missed, concludes that forecasting is pointless because the future can't be known. The third uses forecasting as a way of learning. Every forecast is a set of beliefs about what causes your sales, written down so they can be checked. When it misses, you find out which belief was wrong. Over time, that's how an owner comes to understand their own business far better than intuition alone allows.
How Gridwerk helps
Building a driver-based forecast, a weighted pipeline, scenarios and targets that hold up is part of our Forecasting service, and the sales side, funnel management and team targets, is part of Sales & Marketing. If your sales numbers feel more like hope than prediction, start with the free diagnostic. It's eighteen questions, about seven minutes, and then we call you to go through the rest.
Gridwerk diagnostic