Payback Period and ROI for a French Fries Line

1. What payback measures — and what it does not

Simple payback answers one narrow question: how long does it take before the net cash the project generates equals the cash that was committed to it? It is a screening tool. It deliberately ignores what happens after the payback date, the residual value of the machines, how the project is financed and any strategic reason for owning the capacity.

Return on investment answers a related but different question: how large is the annual net benefit relative to the money committed? Expressed as a ratio, it is usually quoted over a defined period, for example the average annual net benefit across the first several years of operation.

Both are only as good as their inputs, so the model needs one more column before it means anything: a status for every assumption. Mark each line confirmed — meaning it comes from an invoice, a contract, a utility bill or a sample test — or to be confirmed. A payback figure built on unlabelled assumptions is not a number, it is a hope with a decimal point.

One rule protects the whole exercise: choose a convention once and keep it. Decide whether the net benefit is before or after tax, whether the investment includes financing cost, and whether depreciation is in or out. Changing the convention halfway through a comparison is the most common reason two models of the same project disagree.

2. The investment side: what belongs in the denominator

The denominator of a payback calculation is not the machine invoice. It is every cash item that has to be committed before the line earns anything. Buyers who use only the invoice price are, in effect, modelling someone else's project.

Working capital and buyer-side works are the two omissions that flatter a model most, and both are real cash. Material has to be bought before anything is sold, and the electrical connection, water supply and drainage have to exist before the line can be commissioned. The utility planning guide covers the supply side of that list in detail, and the cost factor guide explains why two quotations for what looks like the same line are not comparable until the scope behind them is stated in writing.

3. The output side: build from working hours, not from the rating

The revenue side of the model starts with kilograms actually sold, and those come from hours actually run. A line rating is an indicative figure for a defined product and a defined condition. As an illustration of how one module scales across a range, the published figures for the continuous belt fryer go from model LY4000 at 120 kW electric heating power with an indicative capacity of 100–200 kg/h, up to model LY8000 at 200 kW with an indicative 300–800 kg/h. Those are indicative values across the model range, not a commitment for your product, your slice thickness or your frying curve.

The working formula for annual finished output is short:

• Annual finished output = finished output per hour × productive hours per day × operating days per year × utilisation factor
• Productive hours = shift hours minus start-up and shutdown, minus cleaning between products, minus changeovers, minus planned maintenance
• Utilisation factor = the share of those hours the line is really running, after raw material availability, demand, breakdowns and operator availability are taken into account

Two of those inputs are habitually overestimated. The first is productive hours: a factory that plans two shifts does not get two shifts of output on the day it changes product twice. The second is raw material availability — where the line is fed by a seasonal crop, annual output is capped by the harvest window, not by the machine, and the same capital equipment then works a fraction of the year.

Keep raw input and finished output strictly apart in the model. Peel and sorting losses, trimming and moisture loss during frying are all real, and the yield between the two ends is a result of a sample test on your own material rather than a figure that can be borrowed from a brochure. Valuing raw intake as if it were finished product is one of the most effective ways to make a project look viable when it is not. The capacity tiers and what each one changes behind the machines are covered in the capacity planning guide, and the difference between the chilled and the frozen route — which changes both the investment and the revenue side — is explained on the buyer education site in the comparison of fresh and frozen fries production.

4. The cost side: fixed, variable and step costs

On the cost side the useful split is not accounting categories but behaviour: which costs change with every kilogram produced, which ones are paid whether the line runs or not, and which ones jump in steps.

Variable — moves with output

Raw material, oil make-up and oil replacement, packaging, the energy consumed in heating and cooling, consumables and outbound handling. In most fries line models raw material is the single largest variable line, which is why a small change in procurement price moves the payback more than a large change in machine price.

Fixed — paid per period

Operators and supervision for the shift pattern, planned maintenance and spares, insurance, quality testing, occupancy and the cost of the money invested. These have to be covered in every period, including the periods when the line is idle.

Step costs — jump and stay

Some costs do not move per kilogram and do not stay flat either: a second shift, an additional boiler or heat source, a larger transformer or a generator. Put each one in the period in which it starts, not as a smooth per-kilogram figure.

Energy deserves its own discipline, because it is recorded at two different levels. Connected load is what the site must be able to supply, and it decides the panel, the cable, the transformer and sometimes the generator. Running consumption is what the meter records, and it is the figure that belongs in the per-kilogram cost. Mixing the two is the classic utility error: it either makes the electrical scope look trivial or makes the operating cost look impossible.

5. The two formulas, written out

Once the inputs exist, the arithmetic is simple. Writing it out is still worth doing, because it shows exactly where the model can be manipulated.

• Simple payback (years) = net investment ÷ annual net cash benefit
• Annual net cash benefit = annual sales revenue − annual cash operating costs. Exclude depreciation if you are modelling cash; include interest and fees if your convention treats financing as part of the project. State the convention and stay with it.
• ROI = annual net cash benefit ÷ net investment, quoted over a defined period such as the average annual benefit across the first several years
• Incremental payback for automation = extra investment ÷ annual cost reduction actually removed, plus any quality, uptime or rework benefit the buyer can genuinely quantify
• Breakeven utilisation = the utilisation level at which the annual net cash benefit is zero

Breakeven utilisation is often the most useful output of the whole exercise, because it converts an argument into a question the buyer can answer: can this line realistically be kept above that level of output for the year? If the answer is yes, the project is worth pricing in detail. If the honest answer is no, a better price will not fix it.

6. The assumptions that move the answer most

Not all inputs deserve equal attention. A sensitivity pass over the model shows which ones change the answer enough to argue about, and which ones can safely stay as estimates.

The practical use of the table is a downside case. Take the two assumptions with the greatest effect and move them against the project at the same time — a lower selling price and a lower utilisation, for example — then check whether the payback period is still acceptable. A project that only works in its base case is not a project with a payback period, it is a project with a favourable spreadsheet.

7. How capacity tier and automation change the shape of the model

A larger capacity tier does two things at once. It spreads fixed costs across more kilograms, which lowers unit cost, and it raises both the investment and the fixed cost that has to be covered every month. The second effect is the one buyers underestimate: because fixed costs are paid in full whether the line runs or not, a higher tier is more sensitive to utilisation, not less. Choosing a tier for demand the market does not have yet is the most expensive form of optimism available to a food factory.

Automation changes the shape in a different way. Raising the automation level moves cost from labour hours to investment: the machine share of the project grows, and the dependence on crew size and operator discipline falls. Two consequences follow for the model. First, the payback period depends on how many shifts the line really runs, because a saving that is only realised on the second shift does not exist in a one-shift operation. Second, it depends on product stability, because changeover time is where the economics of a connected route are most easily lost.

The current line configurations are shown on the automatic French fries and potato chips line page and on the semi-automatic configuration page. The useful question is not which one sounds better, but how many hours per year the line will actually run, and how much of the labour it replaces will genuinely leave the payroll.

8. Five ways a payback model flatters itself

These are the recurring failures seen in project models, in rough order of how much difference they make.

• Rated capacity used as output. A model that multiplies the indicative rating by hours and days, at full utilisation and with no cleaning or changeover time, produces a revenue line the line cannot deliver.
• Working capital and factory works left out. Both are real cash and both are required before the first sale; omitting them shortens payback by a margin that has nothing to do with the machines.
• Quotations compared across different scope. When the line boundary, the automation level and the service terms differ, the cheaper quotation is not cheaper — it is smaller.
• Commissioning-week performance treated as steady state. The first operating period is a ramp: recipe tuning, operator learning and spare parts logistics all take time, and output during that period is lower than the design figure.
• Savings counted that never leave the payroll. An operator kept on for other duties is not a saving, and a rented space that is still rented is not a saving either.

Each of these can be corrected with one line of honesty in the model. None of them can be corrected by a supplier's discount.

9. What to prepare if you want the technical side confirmed

A supplier cannot build your business case, but the technical half of it can be made much more precise. The following information lets the investment basis and the performance basis be confirmed together, instead of one being revised after the other.

• Finished product, process route and the stages included, from the first machine to the last
• Raw input per hour or per day, and the working hours and shift pattern you intend to run
• Building drawing or measured floor area, with clear height at several points along the route
• Power supply basis — voltage, phase, frequency and available load — plus the position of the main panel
• Water source, pressure, drainage positions and the wastewater route
• Heating source available for frying, and any refrigeration or compressed-air provision
• Whether the site is ready or still to be prepared, so buyer-side works can be listed rather than assumed
• Delivery term intended, and the port or destination to be used
• Your own commercial inputs: expected selling price, raw material cost and finance terms

Items one to eight are what the technical offer is built on, and they are also the items that decide the size of the denominator in your payback calculation. The layout planning guide covers the building-side checks in more detail, and the RFQ checklist lists the project information that belongs with them.

Frequently asked questions