How to Build a Revenue Forecast That Doesn’t Start With Last Year + 10%

I have a complicated relationship with revenue growth percentages.

Not because they’re useless.

Because they have a tendency to arrive before the explanation.

“Revenue grows 10% next year.”

Okay.

Why?

“Well, we grew about 10% this year.”

That is certainly information.

It is not yet a revenue forecast.

One of the easiest ways to build next year’s forecast is to take this year’s revenue, add a percentage, spread it across 12 months and move on with your life.

I’ve seen versions of this plenty of times.

Sometimes the percentage comes from historical growth.

Sometimes it comes from the budget target.

Sometimes Sales supplied it.

Sometimes 10% simply feels like the sort of number a respectable company should grow.

Then someone changes it to 12%.

EBITDA improves.

Everyone feels better.

Finance has had a productive afternoon.

The problem is that nothing about the business actually changed.

That’s why when I review a revenue forecast, I don’t start with the growth rate.

I want to know what has to happen in the business to produce it.

What Is a Revenue Forecast?

A revenue forecast estimates how much revenue a company expects to generate over a future period based on the underlying drivers of the business.

Those drivers vary considerably by company.

For a SaaS business, they might include:

• existing customers,
• new customers,
• pricing,
• churn,
• expansion,
• contraction,
• and timing.

For a services company:

• billable employees,
• utilization,
• billing rates,
• projects,
• backlog,
• and capacity.

For a product company:

• units,
• price,
• product mix,
• channels,
• returns,
• and seasonality.

The mechanics differ.

The principle doesn’t.

Revenue should come from something happening in the business.

If I can’t explain what creates the revenue, I don’t have much confidence in the number.

Why Last Year + 10% Is So Tempting

Because it’s easy.

Last year we did $50 million.

Next year we’ll grow 10%.

Revenue forecast: $55 million.

Done.

Nobody had to discuss customer retention.

Nobody had to challenge the pipeline.

Nobody had to determine whether the sales team has enough capacity.

Nobody had to ask whether pricing can really increase 6%.

This is an extremely efficient process if your objective is finishing the spreadsheet.

It is less useful if your objective is understanding next year’s revenue.

Historical growth absolutely belongs in the conversation.

If a company has grown between 8% and 12% for five years, I want to know that.

But history is a reference point.

It isn’t the mechanism.

The business still has to produce the next dollar.

I want to know how.

Start With the Revenue Equation

Before I build a revenue forecast, I want the simplest possible explanation of how the company makes money.

For some businesses:

Customers × Average Revenue per Customer = Revenue

For others:

Units × Average Selling Price = Revenue

Or:

Billable Headcount × Utilization × Billing Rate = Revenue

A SaaS model might look more like:

Beginning ARR + New ARR + Expansion − Contraction − Churn = Ending ARR

Then revenue recognition adds another layer.

Your equation may be more complicated.

That’s fine.

But I want to be able to describe the economic engine without opening 14 tabs.

If we can’t explain how revenue happens in plain English, I’m not ready to model it.

Step 1: Understand the Existing Revenue Base

Before forecasting growth, understand what you’re starting with.

How much revenue is already reasonably visible?

For a recurring-revenue business, this may mean the existing customer base.

For services, it might include backlog and contracted projects.

For other businesses, purchase orders, subscriptions or recurring customer behavior may provide visibility.

I usually separate existing revenue from revenue that still has to be created.

That’s important.

If the forecast says $60 million and $45 million is already supported by existing customers or contracted work, I think differently about the remaining $15 million.

That $15 million has a job to do.

Where does it come from?

New customers?

Price?

Expansion?

New products?

Volume?

The gap between what already exists and what must still happen is one of the first things I want to see.

Step 2: Separate Volume From Price

Revenue growth can look identical financially and completely different operationally.

Suppose revenue increases 10%.

Maybe volume increased 10%.

Maybe price increased 10%.

Maybe volume increased 4% and price increased 6%.

Maybe mix changed.

Those aren’t the same story.

This distinction becomes especially important when management starts planning resources.

If growth is coming from more customers, Operations may need more capacity.

If it’s mostly pricing, maybe not.

If the model simply says:

Revenue +10%

we’ve hidden information management may actually need.

I don’t like hiding useful things behind percentages.

Percentages already have enough power in Finance.

Step 3: Model Existing Customers Separately From New Customers

For many businesses, this is one of the most useful distinctions you can make.

Existing customers behave differently from customers you haven’t acquired yet.

For the existing base, I want to understand:

renewals,

churn,

expansion,

contraction,

pricing,

and expected usage or volume.

Then model new customers separately.

How many?

When do they arrive?

What’s the average deal size?

How quickly does revenue begin?

This prevents a common problem where all revenue growth is blended together and nobody can explain what’s actually driving it.

A forecast should tell me whether we’re depending on customers we already have or customers Sales still needs to find.

Those are very different levels of certainty.

Step 4: Be Careful With the Sales Pipeline

This is where Finance and Sales get to strengthen their relationship.

Sales has a pipeline.

FP&A has questions.

It’s a beautiful tradition.

Pipeline can be enormously useful for forecasting revenue.

But I don’t automatically treat pipeline value as forecast revenue.

I want to understand:

What stage is the opportunity?

What’s the historical conversion rate?

How long does each stage usually take?

How large are the deals?

How often do close dates move?

How quickly does revenue begin after signing?

And perhaps most importantly:

Who submitted the forecast?

After enough forecast cycles, you learn that different people have different relationships with probability.

Some sales leaders won’t call something committed until the contract is practically framed on the wall.

Others have already mentally spent the revenue from a customer who hasn’t answered an email in three weeks.

Neither makes them bad at Sales.

It does mean Finance should understand the pattern.

The forecast needs the best estimate of what will happen, not the most emotionally compelling CRM field.

Step 5: Use Conversion Rates Carefully

Historical conversion rates can help translate pipeline into expected revenue.

Suppose historically:

40% of qualified opportunities close,

the average deal is $100,000,

and the average sales cycle is 90 days.

Useful.

But averages can become dangerous when Finance stops looking underneath them.

Did conversion change by segment?

Product?

Salesperson?

Deal size?

Lead source?

Geography?

Has the sales process changed?

Did the company recently increase pricing?

Is the current pipeline actually comparable to the historical pipeline?

I like historical averages.

I just don’t like them enough to stop asking questions.

Step 6: Get the Timing Right

A deal closing is not always the same thing as revenue appearing.

This sounds obvious until you look at enough models.

Suppose Sales expects a $1 million contract to close in September.

Great.

Does the company recognize $1 million in September?

Maybe.

Maybe not.

Implementation could take time.

Revenue might be recognized monthly.

There may be usage assumptions.

There may be milestones.

There may be seasonality.

There may be accounting considerations.

A revenue forecast needs to reflect when revenue is actually recognized, not simply when everyone celebrates the deal.

This is also why Sales forecasts and revenue forecasts aren’t necessarily interchangeable.

Sales is forecasting sales.

Finance is forecasting financial results.

They’re related.

They are not twins.

Step 7: Model Churn Before You Get Excited About Growth

New business is more fun.

I understand.

Nobody wants to begin the planning meeting with a long discussion about customers leaving.

Unfortunately, customers sometimes insist on participating in the forecast anyway.

For recurring-revenue businesses, churn can completely change the growth story.

Suppose the company expects $10 million of new business.

Wonderful.

If $7 million of existing revenue leaves, we have a different conversation.

This is why I like revenue bridges.

Beginning revenue.

Plus new business.

Plus expansion.

Minus contraction.

Minus churn.

Equals ending revenue.

Now management can see what’s actually producing growth.

I’ve seen companies celebrate impressive new sales while the existing customer base was quietly leaking out the other side.

Finance should probably mention that.

Preferably before the celebration gets expensive.

Step 8: Understand Capacity Constraints

This matters especially in services and operationally constrained businesses.

A forecast may show demand for $30 million of revenue.

Can the company actually deliver $30 million?

Do we have enough people?

Equipment?

Inventory?

Production capacity?

Implementation resources?

Sometimes Sales can sell faster than Operations can deliver.

This is a good problem until it becomes a bad one.

For services businesses, I often want to understand:

billable headcount,

available hours,

utilization,

billing rates,

hiring timing,

and employee turnover.

If your revenue forecast requires 50 consultants and the hiring plan produces 38, Excel does not get the deciding vote.

Capacity does.

Step 9: Don’t Let the Target Become the Forecast

We covered this in the budget-versus-forecast discussion, but revenue is where I see the problem become particularly tempting.

The company wants to grow 20%.

That’s the target.

Finance builds the driver-based forecast.

It says 13%.

Nobody particularly enjoys 13%.

Someone asks:

“What would we have to believe to get to 20%?”

That is a perfectly legitimate question.

Build the bridge.

Maybe we need more pipeline.

Higher conversion.

More Sales capacity.

Better retention.

Faster implementation.

More pricing.

Now we have a useful conversation.

The dangerous version is simply changing the revenue growth assumption to 20%.

That isn’t a plan.

It’s a request.

Excel is very polite about requests.

Step 10: Build the Revenue Bridge

One of my favorite ways to review a revenue forecast is to build a bridge from current performance to forecast performance.

For example:

Current Revenue

• Existing customer growth
• Price increases
• New customers
• New products
  − Churn
  − Contraction
  ± Mix / timing

= Forecast Revenue

Now I can see where growth comes from.

Suppose the company plans to grow from $50 million to $60 million.

I don’t want to hear only:

“We’re growing 20%.”

I want to see the $10 million.

Where is it?

Maybe:

$2 million comes from pricing.

$3 million from existing customer expansion.

$7 million from new customers.

Minus $2 million of churn.

Now we’re talking about a business.

And I can start annoying everyone properly.

How confident are we in pricing?

Which customers expand?

How much pipeline supports the $7 million?

Why is churn improving?

Suddenly the growth rate has somewhere to hide less.

That’s progress.

Step 11: Compare the Forecast With History

Now I bring history back.

How fast have we grown before?

What happened to pricing?

How much new business did Sales actually close?

What was conversion?

What was churn?

How quickly did new hires become productive?

What was the seasonal pattern?

If the new forecast looks dramatically different from history, that’s not automatically wrong.

Businesses change.

But I want the explanation.

If Sales productivity has averaged $800,000 per rep and the forecast assumes $1.3 million, what changed?

New territory?

Better leads?

Higher pricing?

Different product?

AI?

Caffeine?

Something needs to explain the improvement.

“That’s the plan” isn’t the explanation.

Step 12: Look for Leading Indicators

A good revenue forecast shouldn’t wait for revenue to tell you revenue is changing.

Depending on the business, leading indicators might include:

pipeline,

bookings,

website traffic,

leads,

conversion,

renewals,

customer usage,

backlog,

sales activity,

implementation volume,

or order intake.

The useful indicator is the one that moves before the financial result.

This is where FP&A can become much more valuable than monthly reporting.

By the time revenue misses, the interesting event may have happened three months ago.

Pipeline slowed.

Conversion deteriorated.

Customers reduced usage.

Renewals weakened.

If Finance only looks at revenue, we’re reading the ending of the story.

I want the earlier chapters.

Step 13: Build Scenarios Around the Variables That Matter

I don’t need every possible combination of revenue outcomes.

I want the uncertainties management might need to respond to.

What if conversion falls?

What if pricing holds better than expected?

What if churn increases?

What if hiring is delayed?

What if a major customer doesn’t renew?

What if the product launch slips?

Then quantify the impact.

This is where a revenue model becomes useful for decisions.

Maybe slower Sales hiring barely changes this quarter but creates a large gap six months from now.

Maybe improving retention by two points is worth more than adding several million dollars of pipeline.

Maybe the plan depends almost entirely on one assumption nobody has been discussing.

Those are useful discoveries.

They occasionally ruin otherwise pleasant meetings.

That’s part of the service Finance provides.

Step 14: Compare Forecast With Actual Results

Once the forecast is running, don’t just replace the old forecast with a new one and forget what you previously believed.

Keep score.

What did we forecast?

What happened?

Why were we wrong?

Was it:

volume?

price?

conversion?

churn?

timing?

mix?

sales capacity?

customer behavior?

The goal isn’t to punish whoever supplied the assumption.

If people learn that every miss results in an interrogation, they become very creative about the assumptions they provide.

The goal is to learn.

Maybe a particular Sales region consistently forecasts conservatively.

Maybe close dates move an average of 30 days.

Maybe expansion is more predictable than new logos.

Maybe implementation capacity is the real constraint.

Over time, those patterns should make the forecast better.

This is where experience starts doing some of the work.

You remember which assumptions behaved badly last time.

I consider grudges against bad assumptions perfectly healthy.

How Often Should You Update a Revenue Forecast?

It depends on how quickly the business changes.

Monthly works well for many companies.

Some high-volume businesses may need more frequent updates.

Others can operate effectively with quarterly reforecasting.

The important part is that the forecast changes when meaningful information changes.

I don’t want Finance rebuilding the entire revenue model every Friday because one opportunity moved stages.

I also don’t want everyone staring at a forecast built four months ago while Sales already knows the quarter has changed.

The cadence should match the speed of the business.

Not Finance’s desire to spend more time forecasting.

We already have hobbies.

Presumably.

Revenue Forecasting Methods

There are several ways to approach revenue forecasting.

Historical Growth

Apply historical growth rates to prior performance.

Simple and sometimes useful for stable businesses.

Weak when business conditions are changing.

Bottom-Up Forecasting

Build revenue from underlying operational drivers such as customers, units, price, pipeline or capacity.

This is generally the approach I prefer when those drivers are measurable.

Top-Down Forecasting

Start with market size, market share or broader growth assumptions.

Useful for strategic planning.

Usually less useful for near-term operating forecasts.

Pipeline-Based Forecasting

Use Sales opportunities, stages, probabilities and timing to estimate future revenue.

Very useful when the underlying CRM data and Sales process are reasonably reliable.

Driver-Based Forecasting

Connect revenue to the few variables that economically produce it.

This is often where FP&A can add the most insight because management can see not only the revenue outcome but what needs to happen to create it.

In practice, I often like combining methods.

History gives context.

Bottom-up drivers explain mechanics.

Pipeline gives current information.

Management judgment fills gaps.

No single method gets custody of the truth.

A Simple Revenue Forecasting Checklist

Before I trust a revenue forecast, I want to be able to answer:

What creates revenue?

How much revenue is already visible?

How much still has to be won?

What’s coming from existing customers?

What’s coming from new customers?

What assumptions are we making about price?

Volume?

Churn?

Conversion?

Timing?

Capacity?

What does history tell us?

What does the current pipeline tell us?

What leading indicators are moving?

Who owns the assumptions?

What happens if we’re wrong?

And perhaps most importantly:

Can someone explain the forecast without opening the model?

If the answer is no, I worry that we’ve modeled the spreadsheet more carefully than we’ve modeled the business.

The Revenue Forecast Should Tell a Story You Can Check

A good revenue forecast doesn’t just tell me:

Revenue will be $60 million.

It tells me:

We start with this customer base.

We expect this much retention.

This much expansion.

Sales needs this much new business.

That requires this pipeline.

At these conversion rates.

With this many salespeople.

At these prices.

And this timing.

Now I have something I can challenge.

More importantly, I have something I can monitor.

If pipeline weakens, I don’t need to wait three months for the revenue miss to learn that the forecast is in trouble.

If retention improves, I can see upside developing.

If hiring slips, I can understand the future impact.

That’s what I want from FP&A.

Not a number that looks reasonable.

A view of the business that tells management what has to be true.

Last Year + 10% Isn’t Always Wrong

I’ll give the percentage its defense.

Sometimes last year’s growth rate is a perfectly reasonable starting point.

Stable business.

Stable customers.

Stable pricing.

Stable market.

Not much operational change.

Fine.

I’m not going to build a 14-tab revenue model just to prove Finance knows multiplication.

Complexity isn’t the goal.

Understanding is.

If 10% is the right answer, great.

I just want to know why it’s the right answer.

Because there is a very large difference between:

We expect 10% growth because these business drivers support it.

and:

We expect 10% growth because I typed 1.10 into the formula.

Excel will happily calculate both.

Management deserves to know which one it’s looking at.