What still gets repaired, and what stops the rest
Every guide on this site answers a price question, and the price has to come from somewhere: a repair company that publishes medians built from its own completed work orders, with the number of jobs behind them. For a whole class of machines that source does not exist — nobody publishes what a vacuum cleaner, a sewing machine or a stereo repair costs, with a job count attached. This page reports what does exist instead: the largest open record of repairs carried out in public, which logs what was carried in, how old it was, whether it was fixed, and what stopped the ones that were not.
What the record is, and how big it is
The Open Repair Alliance gathers the repair data that community repair groups record at their events and publishes it as open data. Its 2025 release counts 441,068 repair attempts, logged at 31,436 events, by 1,482 repair groups, in 32 countries. The unit being counted is a repair attempt: one item, at one event, with one recorded outcome.
The electrical and electronic part of the same dataset grew from 208,491 records in the 2024 release to 305,649, which the source describes as an increase of 46% in just one year. In the twelve months from August 2024 to July 2025 its groups logged 131,790 attempts at 9,123 events in 22 countries, and the source's arithmetic from that is an average of nearly 11,000 items logged per month — around 350 repair attempts every day.
Two limits come with the size, and both are the source's own. The record covers the items people chose to carry to an event, so it is not a sample of the machines sitting in homes. And it is not complete even for the events that happen: the 2024 report estimated that over 4,000 groups are active globally, performing 190,000 successful repairs a year, against the 1,158 groups that recorded data in its 2024 snapshot.
How much of it gets fixed
| Item class | Repair rate |
|---|---|
| Electrical and electronic items | 54% |
| Everything else: textiles, furniture, bicycles, jewellery, household items | 83% |
The source's own sentence is blunt about the gap: electrical and electronic items are much harder to repair, and no EEE category in the dataset has a higher success rate than a non-EEE category. The 2024 report, working from the smaller dataset of that year, put the electronic figure at just over 53%, with 25% of all attempts ending with the product deemed end of life and ready to be recycled, and 18% of products recorded as repairable but not repaired because the repair needed more time, a spare part that was not there, or a skill that was not there.
The categories the data covers
The source publishes two ranked lists from the same release, and they are not the same list: one ranks categories by how many items were logged, the other by success rate. Both are reproduced here as published, and a cell reading “not in this list” means the category is absent from that ranking — not that its figure is zero.
| Electrical and electronic category | Items logged | Repair rate |
|---|---|---|
| Vacuum cleaners | 25,106 | not in this list |
| Lamps | 21,002 | 70% |
| Power tools | 20,602 | not in this list |
| Hi-fi separates | 18,280 | not in this list |
| Coffee makers | 18,116 | not in this list |
| Sewing machines | not in this list | 71% |
| Hair dryers | not in this list | 64% |
| Non-EEE category | Items logged | Repair rate |
|---|---|---|
| Textiles | 58,657 | 88% |
| Household items | 29,683 | 83% |
| Bicycles | 13,435 | 83% |
| Jewellery | 8,321 | not in this list |
| Furniture | 7,631 | not in this list |
The sewing machine earns its place at the top of the electronic list with the oldest kind of mechanism in it: a 1980s machine is a motor, a belt and a set of adjustments, and the record says it is fixed more often than any other electrical category the groups see.
How old the machines are
Age is the part of this dataset that a price list cannot supply. The source states its own denominator: age data exists for over 69 thousand repairs, or 33% of the total, and the bands below describe those products rather than the whole file.
| Age band | 2024 snapshot | 2024 calendar year |
|---|---|---|
| Below 5 years | 31% | 34% |
| 5 to 10 years | 35% | 34% |
| 10 to 15 years | 16% | part of the 10 to 20 band below |
| 15 to 20 years | 11% | part of the 10 to 20 band below |
| 10 to 20 years | the two bands above | 26% |
| Beyond 20 years | 7% | 6% |
The two columns are the source's own two statements of the same distribution: one splits the middle into 10 to 15 and 15 to 20 years, the other reports a single 10 to 20 band, and its words for the second are that the 2024 figures are almost the same. They are shown as published rather than merged into a band that neither statement uses.
Average ages by category, again from the age-submitted records: mobile phones come in at less than 5 years old, coffee makers, laptops and printers at 7 years, hair dryers and power tools at 10 years, hi-fi integrated systems at 18 years, and projectors at a remarkable 28 years. For the two thirds of the file with no submitted age, the record carries no age to average.
What stops a repair
The same dataset records why a repair did not finish. The source's denominator is 18,921 records carrying an indicated barrier, which it describes as 36% of all end-of-life products; the shares below are shares of those barriers, not of all repair attempts.
| Barrier recorded | Instances | Share of all barriers |
|---|---|---|
| Spare part not available during the session or on the market | 4,676 | 25% |
| Item too worn out to repair | 3,617 | 19% |
| Part perceived as too expensive | 3,362 | 18% |
| No clear way to disassemble the product to reach the part | 3,126 | 16% |
| Repair information or manual not available | 2,289 | 12% |
| Required equipment not available | 1,851 | 10% |
The 2024 report draws the same four leading causes together in its own summary: in 25% of cases it is a lack of available spare parts, in 18% the parts are too expensive, in 16% it is impossible to open the product, and in 12% the repair information is missing. Read against the prices on the rest of this site, the first two entries are the same mechanism seen from the other end: a repair that a household declines because the part costs too much is a repair that a volunteer cannot finish for the same reason.
One machine broken down by fault
Vacuum cleaners are the most-logged category in the dataset, and they are also the one the alliance has analysed by fault type: more than one in five faults seen at repair events is a cable or a cord, and motor faults are the second most frequent at 14.5% while being by far the least likely to be repaired at an event, at 37%, which the source attributes to the high price of spare parts. Blockage is the third frequent fault at 11%, and it has the highest fix rate at an event of any of them, 83%.
| Fault type | Share of faults | Repaired at the event |
|---|---|---|
| Cable or cord | more than 1 in 5 | 81% |
| Motor | 14.5% | 37% |
| Blockage | 11% | 83% |
Two different samples sit behind that table, and the source is explicit about the second one. The counts — 25,106 vacuum cleaners in the 2025 release, and 16,056, or 7.7% of all products, in the 2024 supporting data — come from the whole record. The fault split comes from a microtasking exercise in which participants sorted the descriptions of over 3,000 previous repair attempts into fault types, so its denominator is a floor rather than an exact figure, and this page does not present it as if it were the same sample as the counts.
The second record: RepairMonitor 2024
A separate volunteer dataset covers the same ground from a different network. Repair Cafe International's RepairMonitor collected more than 37,000 repairs in 2024, and 62% of them were successful, in the source's own sentence. Its three most popular items that year were vacuum cleaners, coffee makers, and trousers or pants, and for vacuum cleaners it lists the same faults the alliance found: damaged cords or plugs, broken switches, and motor problems. The published factsheet behind those sentences is a PDF; the figures here are the ones the source states in its own news page text.
The gap the law leaves
The dataset also measures how little of it is covered by a right to repair. The 2024 report states that for 96% of all products brought to community repair events there is no right to repair in sight, and that only 40 products out of the more than 208 thousand logged at that point — 0.0002% — fall under legislation it could identify. The categories with the most attention, tablets, televisions and phones, account for just under 9,000 repair attempts, which the source reports as 4% of everything brought in. That is the context in which a 71% sewing machine repair rate should be read: the category with the best outcome in the data is one where the volunteer network, rather than the manufacturer, is the repair channel.
What this data cannot tell you
- It contains no prices. Every figure above is a count, a share or an age; not one of the four sources publishes what a repair cost, so nothing on this page can be compared with a quote.
- It is not a sample of households. The rates describe items carried to a repair event, which is a self-selected group: broken, repairable and worth carrying.
- Its two rankings are separate lists from the same release, and the counts and the fault split rest on different sample sizes.
- Its age figures cover the third of records that submitted an age, and none of the sources explains what the other two thirds would add.
- It is a record of outcomes, not of machines in service. A high repair rate means the item was fixable at the table, not that the model is reliable.
Why there is no guide for these machines on this site
A guide here needs two things from one publisher: a cost band, and the number of completed jobs behind it. For vacuum cleaners, sewing machines, hi-fi separates, projectors and the small kitchen machines, this site's source register finds no publisher that has both, and it records the search that established that rather than the conclusion alone. So these categories get this page instead of a guide, built only from what the published data does measure. Where the data stops, the page stops.
The decision side of the same question is measured elsewhere on this site from a repair company's own job records: what households actually did when a technician quoted a price is on do people repair or replace. For the machines that do have a published cost band, expected life and price sit side by side on how long appliances last and what appliance repairs cost, and the arithmetic for a quote of your own is in the repair or replace calculator.
The same questions, answered from the data
What share of the items brought to repair events gets fixed?
The Open Repair Alliance reports a repair rate of 54% for electrical and electronic items and 83% for non-EEE items, and states that no EEE category has a higher success rate than a non-EEE category. Its 2024 report, on the smaller dataset of that year, put the electronic figure at just over 53%, with 25% of all attempts ending with the product deemed end of life, ready to be recycled.
Which categories have the highest repair rates?
In the 2025 release the highest success rates among electrical and electronic categories are sewing machines at 71%, lamps at 70% and hair dryers at 64%. Among non-EEE items they are textiles at 88%, bicycles at 83% and household items at 83%. The most commonly seen categories are a different list: vacuum cleaners at 25,106 items, lamps at 21,002, power tools at 20,602, hi-fi separates at 18,280 and coffee makers at 18,116.
How old are the machines people bring to repair events?
Of the products for which an age was submitted - over 69 thousand repairs, or 33% of the total - 7% were beyond 20 years old in the 2024 snapshot, and the source repeats the figure as 6% for the 2024 calendar year. Average ages run from less than 5 years for mobile phones to 7 years for coffee makers, laptops and printers, 10 years for hair dryers and power tools, 18 years for hi-fi integrated systems and 28 years for projectors. For the two thirds of records with no submitted age, the source publishes no age at all.
What stops a repair when it does not succeed?
Of 18,921 records carrying an indicated barrier, which the source describes as 36% of all end-of-life products, a spare part was not available at the session or on the market in 4,676 instances, 25% of all barriers; a part was perceived as too expensive in 3,362 cases, 18%; the item was too worn out in 3,617 cases, 19%; there was no clear way to open the product in 3,126 instances, 16%; repair information was missing in 2,289 instances, 12%; and required equipment was missing in 1,851 cases, 10%.
Why is there no repair price on this page?
The dataset records outcomes, not invoices: neither the Open Repair Alliance nor the RepairMonitor publishes a price for any repair. Prices on this site come from repair companies that publish medians built from their own completed work orders together with the number of jobs behind them, and the source register records that no such publisher covers these categories. The absence is stated rather than filled in with an estimate.
Where these numbers come from
The scale figures, the repair rates, the category counts and the two rankings are from the Open Repair Alliance's 2025 data release of 16 October 2025, which counts 441,068 repair attempts. The barrier counts, the vacuum cleaner fault split and the age cohorts are from the supporting data the alliance published with its 2024 report, which is also where the age denominator of 69 thousand records and 33% of the total is stated. The 2024 repair-rate sentence, the end-of-life share and the legislation coverage figures are from that report itself, and the description of the fault-split sample is from the alliance's own vacuum cleaner insight page, which does not publish the percentages used here. The 2024 RepairMonitor figures are from Repair Cafe International's news page of 13 May 2025, whose underlying factsheet is a PDF and was not used directly. None of the four sources publishes a price for any repair. They cover different populations and periods, and no two of them are added together on this page: each table carries the sample its own source states.