Accountability

How to Forecast Revenue When Your Sales Are Lumpy

Friday, 20 November, 2026

Stop forecasting one smooth number. I think the monthly average is the most dangerous number in a lumpy business. To forecast revenue when sales are lumpy, build the forecast in layers. Start with contracted revenue. Add weighted pipeline. Add a small stretch band. Tie every dollar to the date the cash should land.

Lumpy revenue is not a forecasting error. It is the shape of your business. Your forecast has to match that shape.

Here is a simple example. A flat forecast says $80,000 a month. Real life says $20,000, then $30,000, then $250,000, then $20,000. The average is true. It is also useless. Payroll is due in month two, not month three.

A monthly average hides the months that can sink you

Most forecasting advice was written for companies with hundreds of small sales. Volume smooths things out for them. You do not get that help. When ten deals make your year, one slipped deal moves the whole year.

Big companies with full finance teams miss too. CFO.com reported on a survey of 405 finance, sales, and revenue operations professionals. In that survey, 81% of finance and sales leaders said their forecasts are usually at least 5% off. And 43% said at least 10% off. Yet 95% of finance and revenue operations teams said they were confident planning with those same forecasts.

That gap is the danger. Confidence is not accuracy. In my view, confidence without accuracy is how owners hire ahead of revenue, skip a tax reserve, or sign a lease on a month that never comes.

Here is what I see break in a lumpy business:

  • Averages hide timing. Revenue booked in March and cash collected in May are two different events. Cash pays the bills.
  • One deal skews everything. A single large job makes last quarter look great and next quarter look broken.
  • Gut feel wins by default. Owners round up the hopeful deals because they remember the last big win.
  • Seasons get mistaken for failure. A roofer in January is not failing. He is on schedule.

Find out why your sales are lumpy before you build anything

The cause tells you which method will work. Lumps come from a short list of sources. Name yours first.

  1. Large deal size. A few big contracts drive most of the revenue.
  2. Long sales cycles. Deals take months, then several close in the same week.
  3. Seasons. Demand rises and falls on a calendar you can predict.
  4. Project billing. You bill on milestones, so revenue lands in chunks.
  5. Referral dependence. Leads come in bursts, not a stream.
A chart lists five common causes of lumpy revenue and the forecasting method that fits each cause.
Five common causes of lumpy revenue. Most businesses have two or three at once.

Most lumpy businesses have two or three of these stacked together. A commercial HVAC company can have big deals, long cycles, and seasons all at once. That is why a generic template fails.

The last cause is the one I push on hardest. Our position at Accountability Now is that referrals alone are not a plan. That is why our S.C.O.R.E. business coaching starts with sales and marketing. A repeatable way to win customers will not remove every lump. It does make the gaps between lumps smaller.

Now pull your last 24 months of invoices and sort them by size. If a few large invoices make up most of your revenue, forecast those large deals one at a time. Do not treat them as a rate.

How to forecast revenue when sales are lumpy: use three layers

The layered method sorts your forecast by how sure the money is. You plan spending against the sure money. You treat the rest as upside. It works because it refuses to blend certain money with hopeful money.

Layer 1: contracted and committed revenue

This is signed work, scheduled invoices, retainers, and recurring revenue. Put each item in the month you will invoice it. Then move it to the month you will collect it. Use your real payment history, not the terms on paper.

Say your terms are net 30 but customers really pay in 38 days. Your forecast uses 38. Hopeful collection dates are one of the quietest ways to wreck a cash plan.

Layer 2: weighted pipeline

This is your open deals, each one multiplied by its chance of closing. HubSpot describes the weighted pipeline forecast the same way: take each deal amount, multiply it by the win probability, and add up the results.

The math is simple. A $60,000 deal at a 25% chance counts as $15,000. Ten deals like that count as $150,000 on paper. You will not close exactly that. Read the weighted number as an expected value over many periods. It is not a promise for one month.

The trap is the probability. Most small businesses guess. I think guessing is how forecasts die. Use your own history. Of the proposals you sent in the last 12 months, what share closed? How long did they take? If you do not track this, start today.

Layer 3: stretch revenue

This is leads that are not real deals yet, plus walk-in and referral work you cannot predict. Put a small number here based on your history. Or put zero. Never pay fixed costs from this layer.

The rule that makes it work

Cover your fixed costs from Layers 1 and 2 only. Look at the next 90 days. If contracted revenue plus weighted pipeline does not cover payroll, rent, and debt payments, you have a sales problem today. You do not have a forecast problem. That is a useful alarm. Most owners hear it too late.

Put a date on every deal and treat the big ones as yes or no

Odds are only half of the forecast. Timing is the other half. A deal that is likely to close in 90 days does nothing for next month’s payroll.

We recommend four plain stages. Each stage needs a test that must be true before a deal moves up:

  • Lead qualified: the budget and the decision maker are confirmed.
  • Proposal sent: the scope and price are delivered in writing.
  • Verbal yes: the terms are agreed and the paperwork is pending.
  • Signed: the contract or deposit is in hand.

I am not going to hand you a close rate for each stage. Any number I gave you would be a guess about your business. Pull your own win rate for each stage from your last 12 months and use that.

Then add the date. Put each weighted deal in the month you expect it to close. Add your delivery time and your real payment time to find when the money lands. This is where most spreadsheets fail. They total the pipeline and call it a forecast. A pipeline total is a size. It is not a schedule.

One more rule. Any deal that is a large share of your year gets forecast as a yes or a no. Do not let a very large deal show up as a comfortable weighted slice. Build the plan twice, once with the deal and once without it. If you cannot survive the plan without it, that deal is not reliable revenue. It is a bet.

Show three numbers, because one number claims a precision you do not have

A single number says you know more than you do. Three numbers tell the truth.

  • Floor: Layer 1 plus only your verbal-yes deals. Plan fixed costs against this.
  • Expected: Layer 1 plus the full weighted pipeline. Use this for hiring and marketing choices.
  • Stretch: expected plus part of your stretch layer. This can trigger growth moves, but only after the money lands.
A diagram shows a three-layer revenue forecast stacked above a rolling 13-week cash view.
The three layers feed a 13-week cash view. The floor is what you plan fixed costs against.

Check the gap between floor and expected each month. A wide gap means you depend on deals you do not control. A narrow gap means a stable base. Both are useful to know. A forecast that never shows you the gap is hiding the one thing you need to see.

Update the three numbers every time a deal changes stage. Write down why you moved it. That one habit exposes wishful thinking faster than any software.

Smooth the cash, because you cannot smooth the revenue

Trying to make revenue arrive evenly is mostly a fantasy. You can add retainers or sell smaller deals. Both help. Neither one fixes this quarter. So make your spending respond to the cash instead. I have written before about why cash flow beats growth, and a lumpy business is the clearest case.

  1. Build a cash floor. Hold enough to cover fixed costs through your usual dry spell. If your longest drought is 75 days, hold 75 days of fixed costs. Not a vague three months.
  2. Pay yourself a fixed draw. Set owner pay at your floor case. Sweep the extra into reserve after big months.
  3. Reserve for taxes right away. A big month is not all yours. Move the tax share the day the cash lands.
  4. Keep a rolling 13-week cash view. Weekly detail catches the timing gaps a monthly view buries.
  5. Ask for deposits and milestone billing. Pulling cash forward shrinks the gap between the work and the money.

The fixed draw comes straight from how we teach Operational Data, the O in S.C.O.R.E. We tell owners to plan their pay instead of treating it as whatever is left. We also tell them to run the business on profit, cash flow, and three or four numbers that drive it. I cover that in more depth in our post on operational data as a single source of truth.

Here is a view many owners resist. When people hear that revenue should be smoother, they chase more small clients to even things out. I think that is often a trap in a service business. Small clients can take nearly as much work to sell and serve. They also pull attention away from the larger deals that build the company. Our position is that revenue is not the whole scoreboard. Margin, cash, and capacity count too. Keep the big deals where they fit your model. Fix the cash plumbing around them.

Pick the method that fits the data you really have

There are three common ways to forecast. Each one fits a different business. Here is how I read them for a small company with lumpy sales.

Method Works best when Weak spot for lumpy sales
Pipeline-weighted You track deals with stages and close dates Guessed probabilities give you guessed totals
Historical run-rate Revenue repeats in a pattern by season It misses one-off big deals and market shifts
Driver-based You can tie revenue to leads, close rate, and deal size It needs clean data on every driver
A graphic compares pipeline-weighted, run-rate, and driver-based forecasting for a small business with lumpy sales.
Three forecasting methods. Start with the one your current data can support.

My opinion: most owner-led businesses should use pipeline-weighted for the next 90 days and a seasonal run-rate for the months after that. That beats a fancy model. Driver-based forecasting is excellent once you have a long stretch of clean data. Before that, you are modeling noise.

Do not buy software to fix this first. Gong surveyed 2,015 business leaders at private companies in the US and UK. It reported that 81% of those companies missed a sales forecast in at least one quarter between the first quarter of 2021 and the third quarter of 2023. Gong points to outdated technology and practices as one cause. Keep in mind that Gong sells forecasting software, and half of the people it surveyed work at tech firms.

That may be true for a large team. For a small company, I think the bottleneck is simpler. Nobody updates the pipeline. Our position is that software needs a defined job. A spreadsheet that gets updated every Friday beats a CRM that never gets updated.

Split repeat revenue from new revenue

Repeat work is closer to a rate. New work is closer to a lottery. Forecast them apart. Sales teams do not always do this well. Outreach surveyed 500 sales professionals in the US, Canada, and the UK in early 2024. It found that 74% forecast expansion revenue, but only 58% forecast new-customer revenue.

Do the split in your own model:

  • Repeat and recurring: maintenance agreements, retainers, service plans, and renewals. Forecast these like a utility bill, then adjust for customers who leave.
  • Expansion: add-on projects from current customers. Forecast from how often your customers have really bought again.
  • New business: everything else. Use the weighted pipeline and treat it as the layer that swings.

If new business is nearly all of your revenue, that is your real finding. You do not have a forecasting problem. You have a business that resets to zero every quarter. I think a small recurring layer, like service contracts or retainers, does more for forecast accuracy than any spreadsheet trick.

The forecast fails when nobody owns it

In the owner-led businesses our team works with, the spreadsheet is rarely the problem. The owner updates it on good weeks. The owner skips it on bad weeks. Nobody compares it to what really happened. The forecast drifts into fiction.

We saw the fix at Gardeners Center, a garden and landscape retailer. The owners were the bottleneck. There were no shared numbers. Our coach BenJoe installed a weekly Scorecard Meeting. It had a short list of numbers, and each number had an owner. He coached the owners to run it. Then he coached the team to run it alone. The engagement record says the owners stepped out of daily operations. That result is reported by the client and our team. It is not audited.

[…] He coached us through how to implement them and gave us the tools to continue facilitating them on our own.

Gardeners Center

A forecast review is the same kind of meeting. I have argued before that meetings are not accountability unless they end in a decision and an owner. These rules get you there:

  • One owner for the pipeline. Not the team. One person updates it every week.
  • A fixed review time. Same day, same hour, 30 minutes. Cover stage changes, new deals, lost deals, and why.
  • A log of every lost deal with the reason. Patterns show up that no model would show you.
  • Reps own their numbers. Each salesperson forecasts their own deals and answers for the miss.
  • Forecast against actual, every month. Track the miss in dollars and in months of timing.

That last rule is the one owners avoid. It is the most useful one. It is also the 4Cs of Accountability applied to a forecast. Critique Success: when you hit the number, ask why, so you can repeat it. Correct Failure: when you miss, name it fast and fix the cause. If your last three forecasts ran 30% high, cut the next one by 30%. That is calibration. Calibration is what separates a forecast from a wish.

The real numbers often tell a different story than the owner’s gut. At 518 Floor Care, the owner stayed busy but the business did not make money. Our coaches John Brunty and Kristy Faulk reviewed the books in QuickBooks. They found the pricing model was wrong. They built a pricing spreadsheet, and John coached the owner through the changes. In the client’s words, the business is now “profitable and free of debt.” That is one client’s report. It is not an average or a promise. I share it because the answer was in numbers the owner already had.

Sometimes the gap is bigger than a meeting. Before Accountability Now, I was Chief Revenue Officer at MoneySolver. I have also served as a fractional CRO, which is a part-time revenue leader who works inside the team. In that role at The Pearl Collective, I rebuilt the sales process, marketing process, leadership team, and offer structure. The engagement record reports average 35% annual growth over three years. That figure is reported, not audited, and it is one client. My point is narrow. If your sales process, your pipeline habits, and your revenue plan are all weak at once, you have a leadership gap. A better spreadsheet will not close it.

A working forecast answers five questions in under a minute

You do not need a complex model. You need one page, updated weekly, with three numbers, a date on every deal, and a cash floor. When it works, you can answer these fast:

  • What cash is contracted for the next 90 days?
  • What is the weighted pipeline, and when does it land?
  • What happens to payroll if the largest deal slips a quarter?
  • How many days of fixed costs does the reserve cover?
  • How far off was last quarter’s forecast, and why?

Two limits are worth respecting. First, forecast past the next quarter only in broad strokes. Nobody can predict a lumpy business 12 months out, deal by deal. Use seasons and your recurring base for the long view. Keep deal detail for 90 days. Second, recheck your inputs on a schedule. Close rates, cycle length, and collection time all move.

My stance is simple. Stop trying to make a lumpy business look smooth on paper. Plan around the lumps. Pull your last 12 months of invoices today and rank them by size. Build the three layers for the next 90 days before you make another hiring or spending decision.

Frequently asked questions

How do you forecast revenue when sales are lumpy?

To forecast revenue when sales are lumpy, split the forecast into three layers: contracted revenue, weighted pipeline, and stretch revenue. Put each deal in the month you expect the cash to land. Report the result as a floor, an expected case, and a stretch case instead of one number.

What is a weighted pipeline forecast?

A weighted pipeline forecast multiplies the value of each open deal by its chance of closing, then adds the results. A $60,000 deal at a 25% chance of closing counts as $15,000. The weighted total is an expected value over many periods, not a promise for one month.

How often should a lumpy revenue forecast be updated?

A lumpy revenue forecast should be updated every week at the pipeline level. We recommend one named owner and one fixed review time. Close rates, sales cycle length, and collection time should be checked against real results on a set schedule.

What should you do when one deal is most of your revenue?

When one deal is most of your revenue, forecast that deal as a yes or a no, not as a weighted slice. Build the plan twice, once with the deal and once without it. If the business cannot survive the plan without the deal, reducing that dependence is the first priority.

How big should a cash reserve be for a business with lumpy sales?

A cash reserve for a business with lumpy sales should match the longest dry spell the business has actually had. Measure the reserve in days of fixed costs. If the longest drought was 75 days, hold 75 days of fixed costs.

Should a small business buy forecasting software to fix a lumpy forecast?

Forecasting software will not fix a lumpy forecast if nobody updates the pipeline. Our position is that software needs a defined job. A simple spreadsheet that one person updates every week is a better starting point for most small companies.

What is Operational Data in S.C.O.R.E.?

Operational Data is the third stage of S.C.O.R.E., the business operating system Accountability Now teaches and installs. Operational Data means managing the business with a few numbers that change decisions, including profit, cash flow, and three or four business-driving indicators. Operational Data also means planning owner pay instead of treating it as whatever is left.

This article was drafted with AI research and writing tools, then rewritten and fact-checked by the Accountability Now team against our own client records and Google’s published guidance. Client results are reported by clients or our team and are not independently audited unless stated.

Recent Blog

Should You Offer Payment Plans to Clients? A Practical Guide

Should You Offer Payment Plans to Clients? A Practical Guide

Thursday, November 19, 2026

HomeBlogShould You Offer Payment Plans to Clients? A Practical Guide By Don Markland Key highlightsA payment plan is...

Read More
How to Raise Prices with Existing Clients Without Losing Them

How to Raise Prices with Existing Clients Without Losing Them

Wednesday, November 18, 2026

HomeBlogHow to Raise Prices with Existing Clients Without Losing Them By Don Markland Key highlightsIn the Q3 2025...

Read More
Discounting Is Quietly Killing Your Margins: Here’s the Fix

Discounting Is Quietly Killing Your Margins: Here’s the Fix

Tuesday, November 17, 2026

HomeBlogDiscounting Is Quietly Killing Your Margins: Here's the Fix By Don Markland Key highlightsOn a job with a...

Read More

Let's Get Started.

Big journeys start with small steps—or in our case, giant leaps without the space gear. You have everything to gain and nothing to lose.

I’m ready to start now.