Built to last. Priced to throw away.

A repair score is only the beginning

A product can be easy to take apart and still be a bad purchase for someone who expects to repair it. The replacement part may cost too much. The service centre may be too far away. The battery may arrive after the household has already bought another appliance. These are different failures, and a single repairability score cannot explain all of them.

The useful research question is more specific than whether repair is good: under what conditions does paying more for repairability produce better value for the household? The answer affects product design, spare-parts pricing, warranties, local service networks, and the businesses that connect them.

Our argument is that repairability has an economic threshold. The benefit depends on the probability of a relevant failure, the cost of restoring useful service, and whether someone can afford the initial premium. JDGMNT should investigate those conditions together, rather than treat a repair label as an affordability guarantee.

The gap is visible in current evidence

The European Commission's Joint Research Centre describes the smartphone and tablet repairability rating introduced in June 2025 as an assessment of factors including disassembly, tools, parts availability, software updates, and repair information. The rating captures important conditions for making a repair possible. It does not establish the total bill a particular household will face. ^[1]

BEUC's June 2026 market check makes the distinction concrete. In its selected brands and markets, smartphones had better access to parts and instructions than battery vacuum cleaners. Yet repair costs remained a concern in both categories. One battery spare part was priced at €281, about 50% more than the original vacuum cleaner. That is an observed example from the check, not the average price of a replacement battery. ^[2]

The FTC's 2021 repair-restrictions review also examined barriers beyond physical design, including access to parts, tools, and diagnostic software. It found little support for many manufacturers' justifications for restrictions. That is evidence about the repair market and the arguments made within it; it does not mean every product can be repaired safely by every owner. ^[3]

Together, these sources point to a supply-chain question. A repairable design needs a usable service market around it. If the manufacturer controls a scarce part, or an independent shop cannot obtain the required diagnostic access, a good mechanical design may leave the buyer with little bargaining power.

Put a price on the option to repair

Consider two fictional appliances providing the same service over five years. A basic version costs $220. A repairable version costs $300. Assume both have the same probability of one relevant failure during that period. The basic version is replaced for $220 after that failure; the repairable version is restored for an all-in $80.

The buyer pays an $80 premium now for the possibility of avoiding $140 later. If the chance of that failure is p, the expected cash advantage of the repairable version is:

Expected advantage = p × ($220 − $80) − $80

The break-even probability is approximately 57%. If the all-in repair rises from $80 to $140, even a 100% failure probability only recovers the upfront premium in this simplified comparison. A premium that looked reasonable can disappear into labour, delivery, or diagnostic charges.

This is a conditional calculation, not a forecast or a buying recommendation. It assumes one failure at most, a successful repair or replacement, unchanged prices, equal service after intervention, and no residual value at the end. It omits discounting, financing, downtime, maintenance, and environmental effects. Different reliability between the products would require separate failure probabilities. A higher failure rate is not a desirable product characteristic; here it only changes how often the repair option is exercised.

Expected savings do not pay an upfront bill

Even where the calculation favours the repairable product, a household may not have the extra $80 today. Financing can change the comparison again. A design that reduces expected spending over five years can still be inaccessible at the point of purchase.

Downtime is another constraint. A family with no spare appliance may need a working replacement immediately. Someone with a backup device can wait for a cheaper part. Assigning both households the same repair cost hides an important difference in what they can realistically choose.

We would therefore keep three records: cash paid, days without the service, and the largest payment required at one time. Combining all three into a single score would require assumptions about whose time and liquidity matter. Keeping them separate makes the tradeoff visible.

This also changes the business opportunity. A local loan device, a fixed-price repair, or a parts stock held near customers may create more practical value than a small improvement in disassembly time. The strongest intervention will depend on which constraint actually stops people from completing the repair.

FIG. 03 / THE HUMAN CONSTRAINT

Affordable over five years.
Out of reach today.

Reliability data can mislead without a denominator

A repair shop sees failed products. Its records alone cannot tell us how likely a product is to fail across the installed base. Online complaints have a similar selection problem: a large number of complaints may reflect high sales, poor reliability, or differences in who reports a problem.

To estimate a failure probability, we would need to know which products were in use, for how long, under what conditions, and which owners stopped responding. A product still working at the end of observation is not evidence that it will never fail. A device replaced for fashion or software support has a different history from one with a broken motor.

In the absence of suitable data, publish a sensitivity range. Show what the conclusion requires at different failure probabilities and repair prices. A model should reveal where the evidence is missing rather than turn an unsupported assumption into a precise-looking ranking.

Repair success matters too. A cheap attempt that fails is not equivalent to restored service. If the first repair is unsuccessful and the owner buys a replacement, both bills belong in the record. Repeat failures and incomplete repairs can substantially change the value of the original option.

The study we would run

We would begin with one tightly defined product category and comparable products at similar levels of performance. Battery vacuum cleaners are a useful candidate because the BEUC findings give us a concrete market question. They do not supply enough data to rank the whole category. ^[2]

For a dated sample of products and locations, we would request complete quotes for the same clearly described fault. Record parts, labour, diagnosis, delivery, quoted turnaround, warranty terms, and whether a firm quote can be obtained at all. A missing quote is an observation, not a zero-cost repair. Record the route used: manufacturer, independent shop, or a documented owner repair where appropriate.

Separately, we would seek reliability and service-life evidence, identifying its population and limitations. Without a credible estimate, the public comparison would use explicit ranges. We would not infer failure rates from the quote exercise.

The main result would be a map of where repair is both economically attractive and feasible. A useful follow-up could test whether a capped repair price or a temporary replacement removes more of the barrier than a discount on the initial purchase. That comparison would need a suitable study design and actual participation data.

What would change our view

If the premium remains favourable across credible failure and repair-cost ranges, and households can access the service, the case for paying it is strong. If the result reverses after including ordinary repair charges, the design may need a different price or support model. If downtime determines the choice, faster parts access or a loan service deserves attention.

There are also reasons to repair that this cash comparison does not capture. A person may value continuity, an environmental benefit, or a product they know how to use. Those benefits need their own evidence and should not be smuggled into a claim that the cheapest option has been identified.

The implication for JDGMNT is practical: investigate the product and the service network as one offer. We are interested in businesses that can make repair a dependable option with an understandable price. The next step is a dated, reproducible market study, not another general argument that longer-lasting products are better.