How to create a comprehensive plan
What actually drives each line - and how to build a plan you can defend when it misses.
# The short version
A plan made of typed-in numbers cannot be argued with, cannot be updated, and falls apart the moment reality diverges. A plan made of drivers can be discussed, stress-tested and re-run in minutes. Planning is genuinely hard, and it is iterative - nobody builds a working plan in one pass.
- Plan drivers, not numbers. For every line ask what has to be true for this number to happen.
- Follow your own business logic. What moves your revenue is specific to you - the patterns are borrowable, the judgment is not.
- Start with revenue and work down. Volume drives direct costs, revenue drives receivables, headcount drives personnel, capex drives depreciation - and cash comes last, because everything else lands in it.
- Spend your effort on revenue and personnel, the two biggest numbers you control. Drive the 20% of lines that move the business and hold the rest flat.
- Plan monthly and integrated - P&L, balance sheet and cash flow together. A yearly figure hides the months you run short.
- Step back and check the headline numbers, then go round again. Do the margins, the growth rate and the headcount look like a business you recognize, or a hopeful one? If not, find the assumption that produced the number you do not believe and change that - not the total.
- Re-forecast as reality moves. A plan you keep defending twelve months after it stopped being true is a budget, not a plan.
A plan you can narrate is a plan you can defend. If you cannot explain a line to yourself, it is a number you typed rather than drove.
# A plan follows your business logic
A plan is a model of how your business actually makes money. That model is yours. The lever that moves a machine shop is not the one that moves an agency, and borrowing someone else's structure gives you numbers you cannot defend when they miss.
- Which lever actually moves revenue - price, volume, customers, utilization, occupancy.
- What limits you: capacity, headcount, cash, demand. Every business is constrained by something.
- Where costs step rather than scale - the next hire, the second shift, the bigger warehouse.
The good news is that plenty of the plan is not specific at all. Certain positions behave the same way in every business, and certain industries share the same revenue shape - so you can borrow the pattern and spend your judgment where it counts.
- Working capital is universal: receivables follow revenue, inventory and payables follow direct costs. Days-based drivers work for anyone who has them.
- Below EBITDA is mostly contractual - loan schedules, depreciation on assets you already own, tax rules. Almost none of it is business-specific.
- Personnel scales with headcount everywhere; only the ratio differs.
- Revenue is where the specificity lives - which is why it gets the most attention in this guide.
Borrow the pattern for the lines that follow patterns. Spend your judgment on revenue and on the two or three costs that actually move.
# Build it in this order
Most lines depend on something above them, so the order you build in decides how much rework you do. This is the sequence.
- Revenue first. It is the only line with nothing upstream of it, and almost everything else keys off it.
- Direct costs next, off the same volume - that gives you contribution margin.
- Personnel, from headcount. In a services business this loops back: people are also your revenue capacity.
- The rest of operating expenditure, mostly held flat or at inflation.
- Capex, which produces both depreciation and fixed assets.
- Debt schedules, which produce interest and repayments.
- Working capital days on the three lines that move, which decide when the cash actually arrives.
- Cash last. Never planned - it is what falls out of all of the above.
If you find yourself planning a line before the thing that drives it, you will end up typing a number. Go back one step.
# Planning by department
Do we plan once for the company, or once per department and add them up?
If you have departments, you can plan bottom-up: each owner plans their own lines and the company total is the sum. The company view then becomes read-only - you change it by changing a department, which is the point.
The real benefit is not arithmetic, it is ownership. A number the sales lead built is a number the sales lead will defend and explain; a number finance built for them is one they will dispute the first time it is missed.
Department plans are optimistic. Every owner believes their own case, and the sum is more optimistic than any individual part - so apply a buffer at group level rather than arguing each department down.
That is the honest way to do it. Negotiating each department toward a number you already have in mind wastes everyone's time and teaches owners to pad their next submission. Take the plans as given, hold a central buffer against the total, and be open that it is there.
- Plan revenue where it is owned, and personnel in the department that does the hiring.
- Shared costs - rent, insurance, group functions - belong at company level, not spread across departments by a key nobody agreed to.
- Keep the buffer as one visible line at group level. A buffer hidden inside department numbers is indistinguishable from a bad plan.
- Balance sheet and financing stay central. Departments plan what they control, which is above EBITDA.
- Rolling up department plans and presenting the sum as the company case without a buffer.
- Allocating central costs into departments so the totals tie. It makes every department look worse and none of them can act on it.
- Asking for a department plan and then overwriting it. Do that once and you will not get an honest submission again.
# Eight checks before you call it done
Run these before anyone else sees the plan. Each one catches a specific, common, embarrassing error.
- Contribution margin percentage - plot it monthly across the plan. A step change you cannot explain means a direct cost is not actually moving with volume: it is fixed, mis-mapped, or driven off the wrong base.
- Headcount - does the implied number of people match what anyone has agreed to hire?
- Cash - does it ever go negative? If so you have found a financing requirement, not a modeling error to smooth away.
- Working capital - do DSO, DIO and DPO stay at plausible levels, or has an improvement crept in that nobody has committed to delivering?
- Growth - say your revenue growth rate out loud. If it is 60% and you have added two salespeople, one of those numbers is wrong.
- Seasonality - does the monthly shape resemble last year's? A perfectly flat plan is almost always an unfinished plan.
- The joint - put the last three actual months and the first three plan months side by side. If the plan starts at a level the business has never reached, you have written a wish, not a forecast.
- The reverse test - pick one month and explain every material line to yourself. Anything you cannot explain, you typed.
The most valuable check is the last one. A plan you can narrate is one you can defend, update, and learn from when it turns out to be wrong.
# Re-forecast, do not defend
A plan is worth something for as long as it describes a business you recognize. The moment reality has clearly moved - a quarter well behind, or well ahead - a plan that has not moved with it stops informing decisions and starts being defended.
Update regularly rather than protecting an annual budget that stopped being true in March. The discipline is re-forecasting, not accuracy at the first attempt.
- Keep the original as a frozen version. You need something stable to measure against, and re-forecasting on top of the original destroys exactly that.
- Review against actuals every month at close - that is the rhythm the plan is built on.
- Re-forecast the remaining months only. Actuals are actuals; never rewrite a month that already happened.
- Change the driver, not the number. If volume came in 15% low, move volume - editing the total teaches you nothing.
A plan that missed and was corrected is more valuable than one that was never tested. The point of driving every line is that a miss tells you which assumption was wrong.
# Appendix - line by line
One entry per position, in statement order - the P&L first, then the balance sheet. Each gives the question to ask, the usual answers for your business model, how to build it, and what to avoid. Read the ones you need; nobody works through all of them.
# Revenues
What are we selling, to how many, at what price - and which of those three do we actually control?
This is the line to spend your time on. Direct costs follow it, receivables follow it, cash follows it, and most of the error in a plan lives here. If you get one line right, make it this one.
A single revenue line with a growth rate is still a plan, and it beats no plan at all - for a first pass, or for a genuinely stable business, it is defensible. What it cannot do is tell you why you missed or what to change. The moment you can name what moves your revenue, split it.
Split revenue until every piece has a driver you could argue about in a meeting. Then stop.
- Start from the transaction. What does a customer actually buy, and what decides how many of them there are?
- Separate what you set from what you observe. Price you control; demand you influence; the market you only watch. Only the first two are levers.
- Split where the drivers differ, not where the products differ, and stop when the next split would not change a decision.
A quick test: if two revenue lines always move together, they share a driver and should be one line. If one can halve while the other grows, they must be separate.
How you break revenue into drivers depends on how you sell. Most businesses recognize themselves in one of these - start from the closest and adapt it.
Subscription / SaaS
SaaS - MRR (simple)Keep contraction separate from churn: downgrades and cancellations have different causes and different fixes. Start with the simple MRR waterfall, then move to the funnel version once you trust your lead and conversion data - or the sales-capacity version if growth is limited by how many reps you have rather than by demand.
Ecommerce
EcommerceTraffic is the hardest of the three to move and the easiest to overestimate - split paid from organic, because only one of them scales with spend. Most of the realistic upside sits in conversion rate and order value, not in more visitors. Model returns as a rate, never a fixed deduction: it varies by category and quietly decides your margin.
Professional services
Services (billable hours)Utilization is the lever that matters and the one most often assumed constant - it rarely is. Check your realized rate too: what you actually collect after write-offs and fixed-fee overruns usually sits well below the rate card. And capacity here is people, so growth needs hiring months before the revenue shows up.
Physical product
Price x VolumeSplit by product family only where price or volume genuinely behave differently - two lines that always move together should be one. Keep price and volume apart so a miss tells you which of them moved. And check that your assumed margin survives the volume you are planning: discounts and capacity limits both bite as you scale.
Usage / transaction based
Metric-based (usage/transaction)The customer roll-forward lets you separate "more customers" from "more usage per customer" - usually two different teams' problems. Watch the ramp: new customers rarely arrive at full usage, so revenue lags signups by months. And check whether pricing tiers or volume discounts flatten your top end as usage grows.
Rental / leasing
Rent / LeasingOccupancy is your utilization equivalent - plan it explicitly rather than netting it into the price. Capacity steps rather than scales: you add units in blocks, so model the month each one comes online. Existing leases give unusual visibility on the base, which leaves renewals and void periods as the real uncertainty.
Pick the model above that matches how you sell. With no better information yet, year-over-year growth is an acceptable placeholder - a marker of work still to do, not an answer.
- Leaving a single growth-rate line in place once you know better. It is a fine start and a poor finish.
- Splitting revenue by customer name. Group by segment or channel instead - a customer list changes, a segment does not.
- Planning price and volume as one blended number. Separating them is the whole point: it tells you whether you sold more or charged more.
- Ignoring seasonality because "it averages out". It does not average out within a year, and cash is monthly.
# Other operating output
Are we capitalizing production, or holding finished goods that have not been sold yet?
Used where production and sales diverge - typically manufacturing with changes in finished goods and work in progress, or own work capitalized. If you sell what you make in the same month, you do not need this line.
If in doubt, do not plan this line at all. Leaving it at zero is honest and costs you nothing; a guess distorts your margin and your inventory at the same time.
Do not drive this independently of inventory. Either plan the inventory balance with DIO and let this line be the change in that balance, or plan production volume against sales volume and let inventory follow. One of the two must be derived from the other - never estimate both. Capitalized own work is the other half of this line and it is not inventory: plan it as the cost of the internal project, and remember it does two things at once - it lifts this year's result and it adds to Fixed assets, so it returns as depreciation for years afterwards. If you capitalize, plan the matching asset and its D&A in the same pass.
- Using this line as a plug to make contribution margin look right.
- Driving both this line and inventory off separate assumptions - they are the same physical goods, and two independent estimates will not tie.
# Direct costs
What does one produced unit actually cost us - and how many are we making?
Plan direct costs the same way you planned revenue: unit cost × the same volume. If you sell 1,200 units at 40 and each one costs 17 in materials, materials are 1,200 × 17 - and a volume change then flows through to margin on its own, which is the entire point of driving them at all.
Use the same volume driver on both sides. Where you produce and sell in the same month that is one number: revenue is price × volume, direct costs are unit cost × the identical volume. If you build stock, direct costs follow production volume, not sales volume and Other operating output carries the difference.
This also forces the question most businesses answer only roughly: what does one produced unit really cost? Selling price is usually known to the cent, unit cost far less so - and the gap between them is your margin.
Unit cost × volume is the cleanest - the same Price x Volume shape applied to the cost side. Where you cannot get to a credible unit cost, percentage of revenue is the honest fallback; check the percentage against the last 12 months before committing. Use FTE × cost where direct labor dominates.
- Planning direct costs off revenue when you actually know your unit cost. A percentage hides whether materials got more expensive or you simply sold more.
- A flat monthly amount you have not questioned. Some direct costs genuinely are fixed - a minimum cloud commitment, a permanent production crew - but if you cannot say why it is fixed, it is probably in the wrong place.
- Assuming the historic margin percentage holds while you plan a 40% volume increase. Volume discounts and capacity limits both bite.
- Putting salaried staff here just because payroll posts to one account. Production staff are a direct cost, administration is not - split them by account or by department, never by guessing inside one account.
# Operating expenditure
Which of these costs would change if we changed our plans - and which are simply contractual?
Operating expenditure is where planning effort is most often wasted. Typically three or four lines move meaningfully and the rest are contractual. Model the movers properly and hold the rest at inflation.
Do not roll last year forward with a percentage on top. Ask what each line actually buys you - and once a year, what you would spend on it if you were starting today.
The opposite mistake is assuming opex stays flat while revenue doubles. It will not - more revenue means more people, more tooling, more support. Just not one for one, and that gap between revenue growth and cost growth is your operating leverage. It is the whole reason scale is worth anything, so plan it deliberately instead of letting it fall out.
If you split opex by function - sales & marketing, R&D, general & administrative - published steady-state benchmarks exist for most industries and make a fast plausibility check. For mature SaaS the commonly cited ranges are S&M 30–40% of revenue, R&D 15–25% and G&A 10–15%. Use them to sanity-check a plan, never as a target: they move enormously with growth stage, and a fast-growing company should look nothing like a steady-state one. Do not stack the top of all three: together they come to 80% of revenue, which against a typical contribution margin is a business losing money, not a mature one.
FTE × costs for anything people-related. Inflation adjustment for rent, insurance and licenses. Percentage of revenue for genuinely variable items like marketing, if that is how you budget it.
- Building 25 individually-modeled opex lines. You will not maintain them.
- Applying one inflation rate to everything. Wages and rent rarely move together.
# Personnel
How many people, when do they start, and what do they really cost?
Personnel is usually the largest cost you actually control, and the one most often planned as a single number that grows a few percent a year. Plan headcount instead: each month is last month plus joiners minus leavers, costed at an all-in rate. That way you are planning hiring - an actual decision someone makes - rather than guessing a payroll total.
Plan the month someone starts, not the year. A January hire and a November hire cost wildly different amounts in the same plan year.
The all-in cost is where plans go wrong, because the salary is the smallest part of the question. On top of gross pay come employer social contributions, and in several countries a 13th or 14th monthly payment that lands in specific months rather than spreading evenly.
| Country | Employer contributions, roughly | Extra monthly payments |
|---|---|---|
| Germany | ~20% | None statutory; many collective agreements add one |
| Netherlands | ~20% | Holiday allowance, ~8% of salary |
| Austria | ~30% | 13th and 14th, mid-year and at year-end |
| Spain | ~31% | 14 payments |
| Italy | ~30% plus severance accrual | 13th, and 14th in many sectors |
| France | 40% or more | None statutory |
Treat these as orders of magnitude, not as rates to plan with - they move with salary level, sector and collective agreement. Take your actual ratio from last year: total personnel cost divided by total gross salaries.
If you run departments, plan headcount inside them rather than as one company total. That is where hiring decisions are actually made, it puts the number in front of the person who owns it, and it means a hiring freeze in one team does not need the whole plan rebuilt.
FTE × cost, planned as a roll-forward: an opening headcount, then monthly joiners and leavers. Keep the all-in cost per head separate from the count so you can flex either one.
- Planning headcount as a single annual total. The cost starts in the month someone joins, not on 1 January.
- Using gross salary as the cost. Depending on the country you are understating by a fifth to nearly a half.
- Spreading a 13th or 14th month evenly across twelve. It is a visible cash spike in a specific month, which is the whole reason to plan monthly.
- Forgetting that people are also capacity. In services, more revenue needs the hires to land first - plan the lead time.
# Other operating income
Is this recurring enough to plan, or should it sit in non-recurring items?
Grants, rental income, recharges and similar. The test is repeatability: if it will happen again next year without a new decision, plan it here. If not, it belongs in non-recurring items where it will not distort your trend.
Percentage of a related account where it tracks something, otherwise flat with known step changes on the months you expect them.
- Planning speculative grant income you have not been awarded.
- Using this line to absorb anything that does not fit elsewhere.
# Depreciation & amortization
What assets do we already own, and what are we going to buy?
Depreciation is an output of your asset base and your capex plan, never an input. This is the line people most often get structurally wrong, because it is tempting to plan it as a percentage of revenue - which implies that buying a machine depends on how much you sell.
Plan capex explicitly: what you are buying, in which month, over what useful life. Depreciation on existing assets then rolls off on its own schedule, and new capex adds to it from its month of purchase.
Use the dedicated D&A and capex handling rather than a generic formula: what you are buying, in which month, over what useful life.
If in doubt, hold depreciation flat at its current run rate. That is far closer to right than a percentage of revenue, and it keeps the assumption visible instead of buried.
- Depreciation as a percentage of revenue. It has no causal link.
- Forgetting that existing depreciation runs out - assets fully depreciate and the charge stops.
- Starting depreciation in the month of order rather than the month the asset is in use.
- Planning depreciation without the matching cash outflow. The cash leaves when you pay the supplier - often a deposit and then a balance - while the charge spreads over years. They are not the same event.
# Net interest
What do we owe, at what rate - and what will we earn on the cash we hold?
Two separate things nested in one line: interest paid on debt, and interest earned on cash. Plan them separately even though they net into one figure.
Interest is the annual rate divided by twelve, applied to the balance actually outstanding that month - not to a balance you fixed at the start of the plan, and not a flat annual figure split evenly across the year.
Fixed and floating behave differently and deserve separate treatment. A fixed rate you can plan exactly for the life of the loan. A floating rate is a base rate plus your margin: the margin is contractual, the base rate is not, so it is worth running the plan again with the base rate meaningfully higher and seeing what breaks.
Interest income links directly to the cash balance and an annual rate. Interest expense should come off your loan schedules, including any facility you plan to draw during the period - so a new loan brings its own interest with it rather than needing a separate estimate.
- Forgetting interest on debt you have planned but not yet drawn.
- Applying one average rate across facilities with genuinely different terms.
- Holding interest flat on an amortizing loan. The balance falls every month, so the charge should too.
- Assuming today's base rate holds for three years. If your debt is floating, test a rate two points higher.
- Overlooking commitment fees on an undrawn facility - you pay for the option whether or not you use it.
- Ignoring interest because it is small today. It scales with the debt you are planning to take on.
# Other income / loss
Is there a real recurring driver here, or is this a residual?
Typically financing FX effects, associate results and similar items below the operating line. FX on trade receivables and payables normally stays inside the operating result, so do not strip it out here. For most businesses this line is small and effectively unforecastable.
Hold flat at a trailing average unless you have specific knowledge. If FX is genuinely material, drive it off the exposed revenue or cost base rather than guessing a total.
- Modeling FX in detail when it is under 1% of revenue.
- Planning gains. Plan neutral and treat gains as upside.
# Non-recurring items
Do we know about a specific one-off, in a specific month?
Restructuring, legal settlements, disposals, one-time project costs. This line exists to keep your trend clean - anything genuinely one-off should be here rather than distorting the operating lines you use to judge performance.
Be aware that once you pull one-offs out, your EBITDA is an adjusted EBITDA and will not match the statutory one. If a lender tests a covenant on EBITDA, agree which definition applies before you rely on your own.
Plan zero by default, then enter specific known amounts in specific months. This is one of the few lines where manual entry is the correct answer.
- A recurring 'non-recurring' provision - if it happens every year it is operating.
- Using it to smooth results.
- Leaving genuine one-offs in operating expenditure, which corrupts every year-on-year comparison you make afterwards.
# Tax expense
What is our effective rate, and when do we actually pay?
Two distinct questions. The charge accrues monthly on profit before tax; the cash leaves on your jurisdiction's schedule. Both matter, and confusing them is a common source of cash forecast error.
Use the tax template matching your payment frequency - monthly, quarterly, semi-annual or annual. The charge is driven by profit before tax and your rate; the schedule gates when cash moves.
- Applying the tax rate to EBITDA instead of profit before tax.
- Using the statutory rate when your effective rate differs - check the last two years.
- Accruing tax on a loss-making month without considering loss carry-forwards.
- Assuming the charge and the payment happen in the same month.
- Assuming the cash payment tracks the current year's profit. Across most of Europe prepayments are set from the prior year's assessment: a collapsing year keeps paying last year's installments until you file for a reduction, and a growing year books a large catch-up. Some jurisdictions also charge a minimum tax in a loss year.
# The balance sheet, line by line
Balance sheet lines are stocks: what you report is the closing balance. Some you drive directly - the working capital trio, via days - and some you build as a roll-forward of movements, like fixed assets, debt and equity. Either way it is the closing balance that lands on the statement. Three of these drive most of your cash swing: inventory, receivables and payables.
Most balance sheet lines are planned with Manual change: the balance carries forward and you enter the movement. Only a handful have a real driver - the working capital trio, and debt via its schedule.
That is not a shortcut, it is the honest answer. A provision or a deposit has no formula behind it; inventing one gives you a number that looks derived and is not. Enter the movements you know about, leave the rest flat, and spend the effort where a driver actually exists.
The cash flow statement needs no planning at all - it is derived from the P&L and the movement in these balance sheet lines. That is why the balance sheet matters more than it looks: your investment plan lives here as additions to fixed assets, and it is what turns into the investing cash flow. Plan the balance, and the cash statement writes itself.
# Fixed assets
What are we buying, and what is rolling off?
Fixed assets are the mirror image of depreciation and capex. Plan the capex and this line follows: opening balance, plus additions, less depreciation, less disposals.
Driven by your capex plan. Enter additions in the month the asset arrives and disposals when they happen - Manual change on top of the existing balance. If you have planned D&A and capex properly this line largely takes care of itself.
- Planning fixed assets and depreciation independently - they will drift apart and stop reconciling.
- Forgetting disposals when you plan replacement capex.
# Inventory
How many days of direct costs are we holding on the shelf?
Days inventory outstanding is the right driver because it scales automatically with your cost base. Plan DIO in days and the balance follows from direct costs - which means a volume increase raises inventory without you touching anything.
Use DIO. Take your actual DIO over the last three to six months as the starting point rather than inventing a target - then plan improvements explicitly if you intend to make them.
- Planning inventory as a percentage of revenue - it is driven by direct costs, not by selling price. Margin changes would corrupt it.
- Assuming a DIO improvement without naming what will cause it.
- Ignoring seasonal stock build. If you build inventory before a peak, that is a real cash outflow months ahead of the sales.
# Trade receivables
How long do our customers actually take to pay - not how long our terms say?
Days sales outstanding, driven off revenue. Use your actual collection performance, not your stated payment terms. The gap between the two is often 15 days or more, and it is pure working capital.
One exception matters, and it catches most first SaaS plans. If you invoice annually or quarterly in advance, it is the billing that becomes a receivable, not the recognized revenue - measure DSO against billings or your receivable days will read as three hundred plus. That same invoice also creates the advance payments balance, so plan the two together.
Use DSO. Compute it from the last three months of actuals as a baseline.
- Using contractual terms instead of measured DSO.
- Driving DSO off EBITDA or contribution margin. It is revenue - or, if you bill in advance, billings - that becomes a receivable.
- Planning a DSO improvement as a way to make the cash plan work. If nothing changes operationally, nothing changes.
- Overlooking that fast growth increases receivables - growth consumes cash even when you are profitable.
# Other assets
Is anything in here material enough to plan separately?
Prepayments, deposits, other receivables. Usually small and stable. If any single item has grown large, pull it out into its own line rather than planning the aggregate.
Manual change, held flat unless you know of a movement. Percentage of revenue only if it demonstrably scales with activity.
- Detailed modeling of immaterial balances.
- Letting this line grow without ever asking what is inside it.
# Cash & equivalents
Nothing - cash is the answer, not the question.
Cash is the one line you must never plan directly. It is the output of everything else: profit, working capital movements, capex, financing. If you type a cash number in, you have broken the link that makes the whole plan worth having.
If your cash forecast looks wrong, the error is in another line. Fix that line - never adjust cash.
Left to derive from the cash flow statement. The only cash figure you ever enter is the opening balance; every period after that is computed. The one thing you might attach here is interest income, which links to the balance directly.
- Plugging cash to make the balance sheet balance.
- Planning a minimum cash balance as if it were an input - it is a constraint to test against, not a driver.
# Equity
Are we raising, distributing, or neither?
Equity moves for three reasons that matter: profit or loss for the period, dividends and distributions, and capital increases or buybacks. Retained earnings accumulate automatically. If you grant options, the share-based payment charge also builds equity with no cash attached.
Rolls forward from profit automatically. Add dividends, distributions and capital increases as Manual change in the months they actually occur.
- Forgetting planned dividends - they are often the single largest cash outflow of the year.
- Planning a funding round as certain. Model the base case without it and test the raise as a scenario.
# Provisions
What obligations are building up that we have not yet paid?
Warranties, pensions, restructuring, legal. Mostly slow-moving and specific to your situation - though for most small and mid-sized companies the dominant ones are personnel-related: unused holiday, severance, long-service awards.
Manual change for anything you can name - a known settlement, a restructuring. Percentage of a related account where there is a genuine link: warranty off revenue, unused-holiday and severance off payroll. Whichever you use, decide which P&L line carries the charge: a provision is not free, it is an expense that has to sit somewhere.
- Modeling provisions you cannot explain.
- Driving personnel provisions off revenue - they scale with headcount and payroll.
- Forgetting that a provision release is a real profit effect with no cash movement.
# Financial liabilities
What is the repayment schedule on what we owe, and what do we plan to draw?
Interest-bearing debt. Unlike most balance sheet lines a term loan has a known, contractual future - use it. A revolver or overdraft does not: its balance is driven by the cash you need, which is why checking whether cash ever goes negative is really a check on how much of the facility you have drawn.
Each facility needs three things before it can be planned: how much is drawn and when, how it repays, and what it costs. Get those from the loan agreement rather than estimating - this is the one part of the plan where the future is written down.
| Repayment shape | What it means | Watch for |
|---|---|---|
| Amortizing | Equal installments of principal plus interest | Interest falls each month as the balance drops - do not hold it flat |
| Bullet | Interest only, principal at the end | A single large outflow that can dwarf a year of profit |
| Balloon | Small installments, large final payment | The same trap, further out - and decide explicitly whether you assume repayment or refinancing. Assume repayment and the cash plan looks impossible; assume refinancing and you have hidden the risk. Write down which one you chose. |
| Revolver / overdraft | Drawn as needed, repaid as cash allows | No fixed schedule; the balance is an output of your cash plan |
Use loan schedules rather than a generic formula. Each facility carries its own drawdown, repayment and interest profile, and the interest flows into net interest automatically - so a change to the schedule updates the P&L, the balance sheet and cash together.
- Planning a term loan as a percentage of anything. It is contractual - read the schedule.
- Missing a balloon or bullet repayment. Check the maturity profile across the whole plan horizon, not just next year.
- Planning new borrowing without the interest and repayments that follow it.
- Forgetting the covenants. A plan that technically works but breaches a leverage or interest-cover test in month eight is not a plan.
- Netting a drawn overdraft against cash. It is debt, it costs interest, and hiding it destroys your net-debt figure.
# Trade payables
How long do we take to pay our suppliers?
Days payable outstanding, driven off direct costs. Along with advance payments, this is one of only two working capital lines that help your cash position when they rise - and the one where a change has relationship consequences a spreadsheet will not show you.
Use DPO, based on your actual payment behavior over the last three months. If you are building inventory, measure against purchases rather than direct costs - your payables sit against what you bought, not what you consumed, so using direct costs overstates your days and leaves planned payables too high.
- Planning a DPO extension you have not agreed with suppliers.
- Driving DPO off revenue rather than direct costs.
- Overlooking early payment discounts you would lose by stretching payment. Two percent for ten days against net thirty annualizes to over 35%.
# Advance payments received
Do customers pay us before we deliver?
Deferred revenue and customer deposits. For annual-prepay subscription businesses this is a major source of cash and deserves real attention - you are paid long before you recognize the revenue.
Percentage of revenue where the prepayment pattern is stable. If a meaningful share of customers prepay annually, model that share explicitly - the cash timing difference is the entire point. Manual change where the balance moves with a handful of known contracts.
- Ignoring this line if you sell annual contracts - it is often the reason cash looks better than profit.
- Confusing the deferred balance with cash. Deferred revenue is what you have billed but not yet earned; the invoice creates a receivable at the same moment, and the cash only arrives when that receivable is collected.
# Other liabilities
Is anything in here large enough to matter?
Accruals, VAT payable, tax payable, payroll liabilities and similar. Usually tracks activity levels closely. For any trading company VAT payable is often the largest item here, and it moves on your filing cycle rather than on your P&L - a quarterly filer holds three months of net VAT before paying it out.
Percentage of operating expenditure or revenue, whichever it follows more closely - check which by looking at the last 12 months. Manual change for the VAT and tax payments you can date exactly.
- Letting it become the balance sheet's version of 'Other' - apply the same 5% rule.
- Forgetting the VAT payment falling due on your filing date.
- Forgetting that accrued payroll spikes around bonus payments.