The counter-intuitive result of building a total cost of ownership model is that it usually changes the question rather than answering it. A fleet that starts by asking which edge is cheaper often finishes by asking how many change-outs a season its routes actually require, because that input turns out to move the total more than the purchase price does. The model’s value is that it makes the question visible.
This is a verification checklist for building one. It sets out the cost categories that have to be included, the structure that makes the model reusable across machines and seasons, how the sensitivity test identifies the decisive input, and where the model stops being useful.
Total cost of ownership plow edge: what the model has to include
Include the categories that change between the options.
A total cost of ownership model for plow edges has five categories: acquisition and landed cost, installation labour, operating cost including any rework, failure cost where a part is replaced before its expected interval, and disposal.
Each is measurable, and each behaves differently depending on which specification is being considered. Categories that do not change between the options belong outside the model, and leaving them out is correct provided the omission is stated.
The model’s horizon should cover the whole replacement cycle of the option with the longest interval. Where two edges are compared over a period shorter than that, the comparison is biased towards whichever is changed most often, because the other option’s remaining service falls outside the period and is not counted.
Cost categories to include
The table below sets out each category, what it contains, and what it does to the comparison.
| Category | What it contains | Effect on the comparison |
|---|---|---|
| Acquisition and landed cost | Edge, hardware, freight, any duty or handling | Proportional to how often the edge is bought |
| Installation labour | Removal, fitting, inspection and records | Charged per change-out, so it multiplies with frequency |
| Operating cost | Rework, re-torque checks, minor intervention | Rises with frequency and with fitment difficulty |
| Failure cost | Early replacement, route disruption, recovery | Weighted by where in the season the failure lands |
| Disposal | Scrap handling and any recovery value | Small, but different between materials |
Read across the second column and the pattern becomes clear: four of the five categories scale with how often the edge is replaced, and only the first is a price. That is why the service interval dominates the model.
Acquisition and landed cost
Acquisition cost is the price of the edge and the hardware, and landed cost adds everything required to get it to the fleet’s store. Freight is the obvious addition; handling, any duty where the goods cross a border, and the administrative cost of the order are the ones usually omitted.
Two conventions keep this category comparable between suppliers. Asking each supplier to quote on the same basis, so that delivery terms and any duties are visible, and expressing the cost per edge rather than per shipment, so that different order quantities do not produce different unit figures. The wider treatment of landed cost is a separate question for imported material.
Installation and labour
Installation labour is the category that most often decides the result, and it is the one most often estimated. The hours to record run from bringing the machine in to releasing it, and they should be separated into removal, fitting and inspection so that a change in method is visible in the model.
Rework belongs here rather than in a general overhead. A change-out that has to be repeated because a seat was not clean, or because a pattern was misread, is a cost of the assembly and it scales with frequency in the same way that planned labour does. Keeping it separate makes the frequency effect visible without inflating the planned time.
Operating and failure costs
Operating cost covers what happens between change-outs: re-torque checks, minor interventions, and any rework. Failure cost covers the events that should not have happened, such as an edge failing before its interval, which carries both the replacement and the disruption of the route it was working.
Two properties of failure cost make it worth modelling separately. It is not evenly distributed across a season: a failure in a quiet week costs workshop hours, and the same failure during a storm costs untreated lanes. And it is the category where a specification mismatch shows up first, because an edge chosen for the wrong failure mode tends not to wear out but to fail.
Disposal and recycling
Disposal is the smallest category and the one most often omitted. It contains the handling of the worn part and any recovery value. Steel bodies, carbide inserts and rubber or polymer components have different routes, and separating them at the point of disposal costs less than separating a mixed pile later.
The category matters more when the options being compared have different material mixes. A carbide-heavy option produces less steel scrap and more carbide, and where a fleet has a recovery route for one of those materials the difference is real, even though it is small relative to the labour categories.
Model structure in a spreadsheet
The structure matters more than the sophistication, and six steps are enough.
- One column per option, and rows for each cost category.
- Inputs held in a separate block at the top: interval, labour hours, rates, and quantity per change.
- Each category calculated from the inputs rather than typed in, so a change of assumption flows through.
- A horizon in seasons, with the number of change-outs per season derived from the interval.
- A total for the horizon, and a value expressed per season and per lane mile or kilometre.
- A sensitivity block that repeats the total at the upper and lower ends of each key input.
The sixth step is what makes the model worth building. Without it the output is a single number that appears more precise than the assumptions behind it; with it, the fleet can see which input would have to be wrong for the conclusion to reverse.
Sensitivity analysis
The table below shows the shape of a sensitivity test: each input moved on its own across a plausible range, with the effect on the total.
| Input tested | Range applied | Effect on the comparison |
|---|---|---|
| Service interval | Shorter and longer than measured | Largest: it changes the number of change-outs in every category |
| Labour hours per change | By task, at the fleet’s own rate | Large: charged per event across the whole horizon |
| Downtime treatment | Included and excluded | Decisive where change-outs land in weather |
| Purchase price | Full range of quotations | Moderate: proportional to frequency, and bounded by the quotations |
| Disposal and recovery | With and without recovery value | Small, but relevant where material mixes differ |
The first row is the reason the model begins with the fleet’s own wear record rather than with a supplier’s estimate of life. Where the interval is uncertain, the honest conclusion is that the comparison cannot yet be made, and the fleet should measure it before deciding. The measurement method is described on the blade wear measurement page.
Making the model reusable across a fleet
A model that has to be rebuilt for every comparison is not worth building, and three conventions make it reusable.
- Keep all inputs in one block, with units named, so a new option is added by filling a column rather than editing formulas.
- Express outputs per season and per unit of work, so machines with different route lengths can be compared.
- Record the source of every input, including whether it came from a measurement or an estimate.
The third convention is what makes the model trustworthy two years later. A figure whose source is recorded can be updated when a better measurement arrives; one without a source tends to be treated as a fact.
Where the model is not the answer
Three questions sit outside a TCO model, and a fleet should recognise them rather than expecting the model to settle everything.
- Whether a specification can actually deliver the assumed interval on the route in question.
- Whether the fleet can carry the stock and the fitting capability the chosen option requires.
- Whether the service level the route has to meet changes with the option selected.
The third is the one that occasionally reverses a decision. An option that costs less per season and leaves a rougher surface is a worse answer on a route with a standard to meet, and the model measures cost rather than service. The figure that places cost alongside service is described on the cost per lane kilometre page.
What a finished model looks like
A usable model fits on one page: inputs at the top, a column per option, totals over a horizon, and a sensitivity block below. It states its horizon, its unit of work and the categories it has excluded, and it identifies which single input the conclusion is most sensitive to.
That last line is the model’s real output. A fleet that knows its conclusion depends on the service interval knows what to measure next, and a fleet that knows it depends on the treatment of downtime knows what to decide about scheduling before it decides about edges. Material test methods that support the assumptions are published by ASTM International and ISO, cost and performance research by the Transportation Research Board, and public sector practice by the National Association of Counties and the American Association of State Highway and Transportation Officials.
SENTHAI manufactures plow edges, wear plates and carbide inserts in Rayong, Thailand, and produces edges to the pattern and profile each machine requires, which is what allows two specifications to be run on the same route for a comparison. Models that need the specification and interval confirmed can be raised through the contact page.
FAQ
What should a TCO model for plow edges include?
Acquisition and landed cost, installation labour, operating cost including any rework, failure cost where a part has to be replaced early, and disposal. The categories that decide the answer for most fleets are the service interval and the labour, not the purchase price.
How many years should a TCO model cover?
Long enough to include the whole replacement cycle of the option being considered, which is usually two to four seasons for a plow edge. A horizon shorter than the longest interval on the table will favour whichever option is changed most often.
Why does sensitivity analysis matter more than the total?
Because the total is only as good as its weakest assumption, and the sensitivity test shows which assumption that is. A model that reverses its conclusion when the service interval moves by a small amount is a model that has not decided anything, and the fleet should be measuring that input before choosing.
Send the options being considered, the intervals you have measured and the change-out hours you record. The SENTHAI engineering desk will confirm the specification details the model needs so the comparison runs on your own numbers.