Sales Forecasting: How to Project Revenue Without Guessing

Most sales forecasts are wrong. That’s not a failure—it’s the nature of projecting the future. The real problem isn’t inaccuracy. It’s building forecasts that look precise but are actually just optimism dressed up in a spreadsheet.

If your forecast always shows hockey-stick growth and then reality consistently falls short, that’s not forecasting. That’s wishful thinking with a decimal point.

This article is about doing it differently: building projections that are honest about uncertainty, grounded in real data, and actually useful for making decisions—not just for calming investors or satisfying an annual planning ritual.

What forecasting is (and what it isn’t)

A sales forecast is not a prediction of what will happen. It’s a structured estimate of what could happen, based on what you know right now, under specific assumptions.

The moment you treat a forecast as fact—“we will hit €2.3M this year”—you’ve already made the first mistake.

A good forecast does three things:

It forces you to make your assumptions explicit.
It shows the range of realistic outcomes, not just the target.
It gives you an early warning system: when reality diverges from projection, you know something needs attention.

Think of it less like a GPS telling you where you’ll end up, and more like a navigator who says: “Given current heading, speed, and winds, we’ll arrive at these coordinates—but these are the variables that could change our arrival.”

The forecasting disease: top-down fantasies

The most common approach to sales forecasting goes something like this:

“Last year we did €1.8M. We’re growing, our market is growing, and we’ve hired two more salespeople. So we’ll say… €2.5M. That’s a 38% increase. Sounds ambitious but achievable.”

Then this number flows into the budget, headcount decisions, and client promises—all built on a foundation that has zero connection to how actual sales happen.

This is top-down forecasting: starting from a desired or estimated total, then working backward to justify it.

The problem isn’t the methodology in principle—top-down has a place when you have reliable industry data and a long track record of how your business scales. The problem is using it as a substitute for thinking when you don’t actually know where the number comes from.

The result is forecasts that:

Are systematically optimistic (because nobody wants to present a flat year).
Can’t be decomposed into the actions that would actually create them.
Create no accountability, because “the market surprised us” always works as an excuse when things go wrong.

The method that earns its keep: bottom-up forecasting

Bottom-up forecasting starts from the opposite end. Instead of deciding the total and working backward, you build the total from its components.

The core question is: where, specifically, is this revenue going to come from?

That means breaking revenue down to its actual drivers:

How many clients, deals, or units?
What’s the average revenue per client / deal / unit?
What’s the conversion rate at each stage of your pipeline?
What’s the typical sales cycle length?
What’s the renewal or repeat purchase rate?

When you can answer these questions with real data—or at least with honest estimates tied to specific assumptions—your forecast stops being a number you invented and starts being a model of how your business actually works.

An example

Say you run a flight training school. Instead of declaring “we’ll do €900K in flight training this year,” a bottom-up approach looks like this:

Current enrollment: 8 active PPL students. Pipeline: 5 serious leads at various stages.
Historical conversion from inquiry to enrollment: approximately 40%.
Average dropout rate before course completion: 20%.
Revenue per completed PPL: ~€8,500. Per incomplete course: ~€3,200 average.
Capacity ceiling: 12 simultaneous active PPL students.

Now you can model each cohort through its stages, apply your real conversion and completion rates, and arrive at a number with actual structure behind it. And critically: if you want to hit €900K, you can see immediately what has to change—enrollment rate, capacity, pricing, completion rate, or a combination.

That’s the difference. The number isn’t a wish. It’s the output of a model you can stress-test.

Build your forecast in layers

Once the bottom-up logic is in place, organize it into layers that reflect the certainty of each revenue source.

Layer 1: Existing business. Revenue you have reasonable confidence in—recurring contracts, confirmed renewals, ongoing relationships with stable patterns. This is your floor. In most established businesses, 50–70% of next year’s revenue is actually visible if you look carefully at what’s already committed or nearly certain.

Layer 2: High-probability pipeline. Deals or clients that are advanced in your sales process—proposals out, substantive conversations held, realistic timelines in place. Apply an honest probability discount: if you have €400K in this stage and your historical close rate is 55%, book roughly €220K, not €400K.

Layer 3: New business development. Revenue from prospects not yet in your pipeline—new markets, campaign results, new products, referrals not yet initiated. This is the most uncertain layer. It belongs in the forecast, but sized conservatively and clearly labeled as such.

The discipline is to not let Layer 3 rescue shortfalls in Layers 1 and 2. This is the classic mistake: undercounting existing business and overweighting vague future growth in order to hit a target you’ve already committed to.

The three scenarios you must have

One number is never enough. Every honest forecast requires at least three versions:

Conservative (base downside): What happens if things go reasonably wrong? Key deals slip. A client doesn’t renew. A product launch is delayed. This scenario should still be operationally viable—you need to be able to survive it and keep running the business.

Base case: The most likely outcome given what you know today. Not optimistic, not pessimistic. What do you honestly expect? This is the scenario you plan operations against.

Optimistic: What happens if the pipeline converts better than history suggests, you land that large account in late-stage discussions, and market conditions remain favorable. Useful for understanding upside potential—not for committing to spending.

Why does this matter? Because the most important decisions should be stress-tested against the conservative case, not the optimistic one.

If your conservative scenario requires cutting costs or accessing liquidity, you need to know that in January—not in August when the numbers have already stopped lying for you.

Mistakes that quietly destroy forecast quality

Knowing the method is necessary but not sufficient. Several habits will corrupt even a well-structured forecast.

Anchoring on last year. Last year is data, not destiny. If your market has shifted, your team has changed, or your competitive position is different, last year’s number is context—not a baseline to apply a growth percentage to without examination.

Counting revenue before it’s real. “They said yes in principle” is not a signed contract. “We’re very close” is not a closed deal. Unconfirmed commitments should carry an appropriate probability discount, always.

Mixing forecast and target. A forecast is your best estimate of what will happen. A target is what you want to happen. Confusing these produces numbers that look like data but are really goals. When you manage to goals dressed as forecasts, you lose your early warning system—the one thing the forecast is actually for.

Not updating as you go. A forecast written in December that’s untouched until the following December is theater. Forecasts must be live: updated monthly with actuals, revised when the pipeline changes, explicitly amended when key assumptions break.

Forecasting revenue without forecasting the inputs. If your growth plan requires 40% more sales activity, but you haven’t hired anyone or changed your process, the revenue number is floating. Good forecasting always connects outputs—revenue—to the inputs that will generate them: hours, headcount, conversion rates, marketing spend.

How to actually use a forecast

The forecast is not a filing exercise. Used properly, it should drive real decisions:

Investment decisions. Can you afford the equipment, hire, or expansion you’re considering? Under which scenario does it work, and under which does it create a cash problem?

Cash flow management. When do you need liquidity, and how much runway do you have before revenues have to arrive to avoid a problem?

Hiring timing. When do you need to add capacity? If the conservative scenario materializes, which hires stay on plan and which get postponed?

Pricing decisions. Are you being tempted to discount because you’re behind forecast? Having the forecast in advance makes the cost of that decision visible before you make it.

Team clarity. What does the team need to close, convert, or develop in the next 90 days to stay on the base-case trajectory?

The most valuable moment in a forecast cycle is when actual results diverge from projection. When you’re tracking 20% below the base case by Q2, that should trigger a diagnosis: Is the pipeline soft? Is conversion worse than expected? Is a market assumption breaking down? What has to change before the gap compounds?

A forecast that doesn’t generate questions when reality diverges is useless—and a forecast that never gets compared to reality is decoration.

A few things that matter more in aviation businesses

For ATOs, aeroclubs, charter operators, aerial work companies, and anyone running an aircraft-intensive operation, forecasting has a few specific dimensions worth naming.

Utilization drives everything. Revenue per aircraft per year is almost entirely a function of how many hours it flies. Forecasting revenue without forecasting utilization—and the factors that drive or limit it—is building on sand.

Seasonality is real and routinely underweighted. Most aviation operations have strong seasonal patterns: summer peaks, exam cycles, weather windows. These aren’t surprises. But they create systematic forecast errors when modeled as smooth annual trends.

Regulatory events create step-changes. A change in exam requirements, a new certification category, an AD that grounds part of a fleet, a change in operating rules—these affect revenue in discrete jumps that market-trend forecasting misses entirely. They belong in your scenario planning as specific risk events.

Revenue concentration creates fragility. If 45% of your training revenue depends on two airline pipeline agreements, and those don’t renew, your conservative scenario is not a bad quarter—it’s a structural problem. Good forecasting makes concentration visible before it becomes a crisis.

Revenue and cost forecasts must share the same assumptions. In aviation, many costs are hour-driven. If your revenue model assumes 1,100 aircraft hours this year, your maintenance budget should also be built on 1,100 hours of flying—not a different assumption made in a separate spreadsheet by a different person.

How to tell if your forecast is honest

Stand back from any forecast you’ve built and ask:

Would you bet your own money on the conservative case materializing? If not, it isn’t conservative enough.
Can you trace every major revenue line back to specific clients, deals, or conversion rates? If a line can’t be decomposed, it’s a guess.
Is the growth in this forecast consistent with the resources you actually have? Or does it silently assume productivity or capacity that doesn’t exist yet?
If you showed this to a sharp, neutral person, would you be uncomfortable with their questions? If yes, the forecast probably isn’t honest.
If you miss by 20%, what breaks? If the answer is “everything,” the business depends on forecast precision—which is a fragile foundation.

Forecasting as a tool for better conversations

The real payoff of good forecasting isn’t that you get closer to the right number. It’s that you have better conversations—with partners, advisors, lenders, and your own management team—because those conversations are anchored in explicit assumptions and honest ranges rather than headline targets.

When your advisor can say “last quarter we assumed 42% conversion on the training pipeline—actual came in at 34%. Let’s figure out if this is a temporary variation or a structural shift,” you’re having a conversation that produces value.

When the conversation is “we expected €1.4M and got €1.1M, nobody saw it coming,” you’re having a conversation that produces blame distribution and nothing else.

Good forecasting doesn’t make the future easier to predict. It makes it easier to navigate—because you’ve already thought through what happens when reality diverges from the plan, and you’re not surprised by having to adapt.

Building an honest forecast is often harder than it sounds—not technically, but emotionally. It requires resisting the pressure to project what you want, and being willing to show numbers that might lead to uncomfortable conversations. But those conversations, had early enough, are exactly what separates businesses that adapt from businesses that get blindsided.

If you’re working on a forecast that needs stress-testing—whether it’s for an aircraft acquisition, a fleet expansion, or a new business model in aviation—that’s a process where an independent perspective often changes the outcome. Not by adding complexity, but by separating what you know from what you’re hoping for.

Sales Forecasting: How to Project the Future Without Making It Up