Right to Repair Research: Measuring Whether Owners Can Actually Fix What They Own

Answer first: right-to-repair statutes require you to make parts, tools, and documentation available on fair and reasonable terms. Availability is a legal test that your compliance team can pass with a web page. Whether an owner or an independent shop can actually complete the repair is an empirical question that nobody in the chain is required to answer, and the two things come apart constantly. The measurement that matters is a funnel: decide to repair, find the instructions, obtain the part, have the tool, complete the physical work, and have the device accept the part. A programme that publishes documentation and never tests it is optimising the stage least likely to be the bottleneck. Tools like Koji let you interview owners and independent technicians at the exact stage where they abandoned, and the AI asks what stopped them without a moderator scheduling a call.

What the FTC found when it looked

In May 2021 the Federal Trade Commission delivered Nixing the Fix, a report to Congress on repair restrictions. Its conclusion is unusually direct for an agency document: repair restrictions "have diluted the effectiveness of Section 102(c) and steered consumers into manufacturers repair networks or to replace products before the end of their useful lives," and "there is scant evidence to support manufacturers justifications for repair restrictions."

The report groups the practices raised in the record into eight categories:

The legal hook underneath all of this is the anti-tying provision, Section 102(c) of the Magnuson-Moss Warranty Act, which prohibits conditioning warranty coverage on the use of an article or service identified by brand name unless it is provided free of charge or the FTC grants a waiver. The report notes that since 1975, only three waiver requests have ever been made to the Commission, and all three were denied. The Commission also notes it cannot seek civil penalties when enforcing its three Magnuson-Moss rules, only an injunction, which explains why the enforcement history looks thinner than the rule.

The Commission followed the report with action. On 3 July 2024 it sent warning letters to eight companies over warranty terms that tied coverage to branded parts or authorised service, and over "warranty void if removed" stickers. Those letters and the disclosure duties behind them are covered in warranty comprehension research.

The state patchwork, and why it is a research problem

Since the report, states have legislated where the federal rule stopped. The obligations differ enough that a single national product can face several different duties at once.

Two features of this patchwork drive the research need. First, the California durations are tied to the wholesale price band and run from the last date of manufacture, not from the sale, which means the obligation outlives the warranty by years and long outlives the product marketing. Second, Oregon and Colorado reach parts pairing, the practice of serialising components so a device only functions with a part the manufacturer has validated. That is the stage of the funnel that sits after a physically successful repair, and it is invisible to every metric that stops at parts shipped.

The repair abandonment funnel

Compliance reporting counts availability. Owners experience a sequence, and each step can end the attempt.

1. Decide repair is worth attempting. Beliefs about cost, time, and whether repair voids the warranty all sit here. The disparagement category in the FTC list operates entirely on this stage.
2. Find the instructions. Published is not the same as findable. Owners search your consumer site, not your service portal.
3. Understand the instructions. A document written for a trained technician can be legally available and practically unusable.
4. Obtain the part. In stock, priced sensibly, shipped in a period shorter than the owner patience.
5. Have the tool. Proprietary fasteners and specialised tools convert a ten-minute job into a purchase decision.
6. Complete the physical work. Adhesive, soldered components, and sealed enclosures live here.
7. Have the device accept the part. Software validation, pairing, and calibration. The repair is done and the device still refuses.

This is the same shape as the funnel in product recall notice research, and it produces the same reasoning error. A recall programme that responds to a low correction rate by buying more notification is treating awareness as the bottleneck when awareness was never the problem. A repair programme that responds to complaints by publishing more documentation is doing precisely the same thing at stage two, when the failures are concentrated at stages four, five, and seven.

The corollary is the one to take to a leadership meeting. You cannot infer the funnel from your parts sales data. Every abandonment before stage four is invisible to it by construction: the owner who could not find the manual never ordered a part, so they never appear. Parts revenue measures the survivors of the first three stages and calls them the population.

Documentation availability is not documentation usability

The statutes use language like documentation and tools made available on fair and reasonable terms. Fairness of terms is a legal test. Comprehensibility is not tested anywhere, by anyone, and it is entirely measurable.

The study is straightforward and almost nobody runs it. Give a real independent technician, or a competent owner, your actual published repair document and the actual part, and observe whether the repair completes. Then interview them about where they stalled. This is a usability test with a repair procedure as the task, and the methods in usability testing transfer directly, including the discipline of measuring task completion rather than asking whether the document seemed clear.

The specific things worth scoring:

• Time to locate the correct procedure for the exact model variant, starting from your public site.
• Whether every tool the procedure requires is named before the procedure begins, rather than discovered at step nine.
• Whether torque values, adhesive cure times, and calibration steps are stated or assumed.
• Whether the part number in the document resolves to a purchasable part.
• Whether the procedure ends at reassembly or continues through the software step that makes the part work.

That last one is the most commonly missing section in published documentation, and it maps exactly onto the stage that parts-pairing legislation now reaches.

Interviewing the people who gave up

The hard population in this research is the abandoners, and they are hard for a structural reason: they are defined by not contacting you. An owner who searched, failed, and paid a third party has no relationship with your support system. An independent shop that turned the job away has no reason to file anything.

That is what makes an AI-moderated interview the practical instrument here rather than a nice-to-have. You can invite from a broad list, run in the owner own time, and let the conversation route to the right stage instead of forcing every respondent through a questionnaire built for stage seven. When a respondent says they gave up, Koji generates the follow-up from that answer, asking what specifically stopped them, what they tried next, and what it cost, without a researcher on the call. Voice responses are worth enabling for the same reason they matter in diary studies: people describe a physical task more accurately by talking through it than by typing a summary.

A workable design, using the six structured types documented in structured questions in AI interviews:

The warranty-belief question deserves emphasis. The FTC has been enforcing against tie-in statements and void-if-removed stickers precisely because they communicate a coverage rule to owners who will never read the warranty. If a meaningful share of your owners believe self-repair voids their coverage and your terms say otherwise, you have a comprehension defect that is suppressing repair, and it lives in your packaging rather than your policy.

Reading the results

Report the funnel as a funnel, with the denominator at every stage. A 90 percent completion rate among owners who obtained a part is a statistic about a survivor population, and quoting it without the earlier stages is the single most common way this research gets misread.

Segment by who is doing the repair. Owners, independent shops, and authorised servicers hit different walls, and averaging them hides all three. An independent shop is rarely stopped by physical difficulty and frequently stopped by diagnostic software access; an owner is the reverse.

Watch the cost-of-effort effect. Small increases in friction produce disproportionate drops in compliance behaviour, a pattern documented in the safety-compliance literature and revisited in the recall work. A part that ships in two days and one that ships in three weeks are not a modest difference in convenience; they are different products from the owner point of view, and the second one is a replacement sale you have handed to a competitor.

Finally, treat the repair experience as product research rather than compliance evidence. The parts most often replaced are a ranked list of your design weaknesses, and the stage where owners abandon is a ranked list of your service-design weaknesses. Those two lists are worth more to a hardware roadmap than most feature research, and they are produced as a by-product of a study you may already be obliged to care about.

Frequently asked questions

What did the FTC actually conclude about repair restrictions?

In Nixing the Fix, its May 2021 report to Congress, the Commission concluded that repair restrictions have diluted the effectiveness of the Magnuson-Moss anti-tying provision and steered consumers into manufacturer repair networks or toward replacing products before the end of their useful lives, and that there is scant evidence supporting manufacturers justifications for those restrictions. The report catalogues eight categories of restriction, from physical design choices through software locks and end user licence agreements.

Does allowing self-repair mean we have to honour warranty claims on badly repaired products?

No. Section 102(c) of the Magnuson-Moss Warranty Act prevents you from conditioning coverage on the use of branded parts or authorised service, but it does not require you to cover damage the owner caused. 15 U.S.C. 2304(c) preserves a warrantor position where the failure resulted from consumer damage or unreasonable use. The prohibited move is the blanket condition, not the case-by-case denial supported by evidence.

Which states have right-to-repair laws for consumer electronics?

New York signed the Digital Fair Repair Act on 28 December 2022, applying to devices first sold or used in the state on or after 1 July 2023. Minnesota version took effect 1 July 2024 for products manufactured after 1 July 2021. California SB 244 took effect 1 July 2024, requiring parts and documentation for three years for products wholesaling at 50 to 99.99 dollars and seven years at 100 dollars or more. Oregon SB 1596 added the first statutory ban on parts pairing, applying to devices manufactured on or after 1 January 2025, and Colorado has restricted parts pairing in specific equipment categories.

What is parts pairing and why does it matter for research?

Parts pairing is the use of software to serialise components so a device recognises and fully functions with a replacement part only if the manufacturer has validated it. It matters for research because it is a failure stage that occurs after a physically successful repair. Every metric that stops at parts shipped or repairs attempted is blind to it, which is exactly why owners report the frustrating experience of a completed repair that the device refuses to accept.

How do we reach people who abandoned a repair?

Not through your support system, because abandoning is defined by not contacting you. Invite broadly from your owner base and let the study route respondents by the furthest stage they reached, rather than screening for people who filed a ticket. An AI-moderated interview makes this practical because it can handle a wide invitation list at low cost and still ask each respondent the specific follow-up their stage warrants.

Is this compliance work or product work?

Both, and it is more valuable as product work. The compliance question is whether parts, tools, and documentation are available on fair and reasonable terms. The product question is which parts fail most, where owners abandon, and what a completed repair costs them in time and money. The second set of answers feeds durability and serviceability decisions in the next hardware revision, and the study that produces one produces the other.

---

Ready to measure your repair funnel? Sign up for Koji and get 10 free credits to interview owners and independent technicians about their last repair attempt. Route by the stage they reached, let the AI probe what stopped them, and read a stage-by-stage funnel in hours rather than months.

Related Resources

• Structured Questions in AI Interviews - the six question types that make a stage-routed study scoreable
• Warranty Comprehension Research - the tie-in and void-if-removed disclosures that shape repair beliefs
• Lemon Law Research - what happens when repeated repairs never resolve the fault
• Extended Warranty and Service Contract Research - the paid coverage that sits alongside repair access
• Implied Warranty Research - the durability expectations owners bring to a repairable product
• Product Recall Notice Research - the funnel model this study shares its shape with
• Usability Testing - measuring task completion rather than perceived clarity
• Diary Studies - capturing a multi-day repair attempt as it happens