Research study
Where the garage door ROI number comes from
The 268% figure is real and published. It just does not mean what most people quoting it think it means.
Written by HyreGarage Research Desk Primary-source research and data analysis
Audited by HyreGarage Research Desk Citation, computation and retrieval-date audit
The finding
The 268% garage door return comes from a survey, not from home sales.
Cost vs. Value’s own methodology (September 18, 2025) says the resale side is estimates from an online survey of over 6,000 Realtors, and the cost side is modeled from BLS and BEA price indices.
From 2019 to 2022 the same project recouped only 93–98% of its cost.
What we found
No house sale is observed anywhere in the calculation. The number is a modeled cost divided into a surveyed opinion, not a sale price.
We pulled the garage door row from seven editions in a row. From 2019 to 2022 it recouped 98, 95, 94 and 93 percent: four straight years of getting back less than it cost. Then it hit 103 percent in 2023, 194 in 2024 and 268 in 2025.
Over those seven years, the modeled job cost rose 29.4 percent. The surveyed resale value rose 255.3 percent.
A third measure fits neither side. The BLS producer price index for metal doors, series PCU3323213323211 ("Metal doors (except storm doors)"), rose 69.0 percent over the same years.
None of this means a new garage door is a bad idea. The number measures what agents believe curb appeal is worth, not what buyers paid. If you are pricing a real door, start with our garage door replacement cost study.
What is this page saying, and what is it not?
The garage door ROI figure is quoted everywhere and easy to misread both ways. So the limits come first.
Cost vs. Value publishes its method openly and in plain language. That is more than most widely quoted industry figures do. Every criticism on this page comes from the report’s own description of itself.
The value side is Realtor opinion, gathered by online survey. That is a fair measure of professional judgment about curb appeal. It is not a measure of what a buyer paid, and the two get mixed up all the time.
It may well be a good one. The point is that the 268 percent figure cannot tell you so, because it never looked at a sale.
jlconline.com returned HTTP 403 to every request we made on the retrieval date. Every figure here comes from Internet Archive snapshots of that site. Archived pages can be incomplete, so confirm these numbers against the live report before relying on them.
Resale divided by cost does not always match the published percentage exactly, because the published figure is rounded. We copy what the report printed rather than our own division.
Survey respondents saw the construction costs before they were asked what value the project adds. That is an anchor, and anchoring is one of the most reliable findings in survey research. We raise it as a question about the survey, not as proof of a specific bias.
What does the 268% number actually measure?
It measures what Realtors think a project adds to a sale price, divided by a modeled cost. The Cost vs. Value report comes out every year and has for decades. It is the most quoted source in American remodeling.
The 2025 edition compares 28 projects across 119 U.S. markets. Garage door replacement tops it at 268 percent cost recouped: $4,672 in, $12,507 out.
Most people who repeat that figure treat it as if someone tracked houses with new garage doors through a sale. That did not happen, and the report never says it did. Its methodology page lays out both halves plainly.
The cost side is a model. In the report’s words: “To estimate costs, the Cost vs. Value team first set a baseline cost by creating a price index for each project category.
These indices were based on specific building-product data from the Bureau of Labor Statistics (BLS), including the Consumer Price Index (CPI) and the Producers Price Index (PPI), as well as data from the Personal Consumption Expenditures (PCE) price index, produced by the U.S. Bureau of Economic Analysis (BEA).”
It also used “BLS data to account for changes in average wages for installers from 2019 to now”, with everything “given a proportional weighting to establish a reliable national baseline” from 2019 on.
Local cost differences come from a commercial partner. The team “partnered with Verisk”, who “provided us with building cost data for each of the 28 remodeling projects drawn from XactRemodel, Verisk’s remodeling estimating solution.”
The value side is a survey. “Resale value data for each project was aggregated from estimates provided by an online survey conducted by Zonda to over 6,000 Realtors working in cooperation with the Realtors Property Resource (RPR) of the National Association of Realtors.
Survey respondents were provided with project descriptions, photos and illustrations, plus construction costs and median home prices for each city, and asked ‘What value does each of the 28 remodeling projects add to the sale price of a home?’
Respondents were instructed to avoid making judgments about the motivation of the homeowner in either the decision to undertake the remodeling project or to sell the house.”
HyreGarage analysis: put the two halves together. The top of the ratio is a large national panel of real estate agents estimating, from photos and a cost figure, what a project adds to a sale price. The bottom is a modeled national cost, adjusted by a commercial estimating tool.
So the output is agents’ shared view of curb appeal, divided by a construction cost index. That is useful. Agents see thousands of houses, and their sense of what buyers respond to is worth something. It is just not the same thing as a real return.
The instruction to respondents is careful, and it also shows the limits of the survey. Asking people to ignore the owner’s motives tries to remove a known problem: houses that get a new door before a sale differ from houses that do not.
You cannot survey your way out of that problem, and the report does not pretend to. But a reader cannot treat the result as if the problem were solved.
The seven-year series: 98% down to 93%, then up to 268%
We read the Garage Door Replacement row from every edition from 2019 to 2025. Side by side, the number stops looking like a measure of anything about doors. The last column is the yearly change in the surveyed resale value.
| Edition | Job cost | Resale value | Cost recouped | Change in resale value |
|---|---|---|---|---|
| 2019 | $3,611 | $3,520 | 98% | — |
| 2020 | $3,695 | $3,491 | 95% | -0.8% |
| 2021 | $3,907 | $3,663 | 94% | +4.9% |
| 2022 | $4,041 | $3,769 | 93% | +2.9% |
| 2023 | $4,302 | $4,418 | 103% | +17.2% |
| 2024 | $4,513 | $8,751 | 194% | +98.1% |
| 2025 | $4,672 | $12,507 | 268% | +42.9% |
| 2019 → 2025 | +29.4% | +255.3% | 98% → 268% | The whole period |
National averages as published in each edition, read from Internet Archive snapshots of jlconline.com, retrieved 2026-09-06. The 2024 edition covered 23 projects; 2023 and 2025 covered 28. Percentages are as published, not recomputed. The 2024 and 2025 editions produced the headline: before 2023 this project had never once recouped its cost.
Why did the number nearly triple in three years?
Not because of doors or costs. The change is in the survey answers. For four editions in a row (2019, 2020, 2021, 2022) the garage door lost money on paper: 98, 95, 94, 93 percent. It was drifting down. Then 103, 194, 268.
HyreGarage analysis: a ratio that nearly triples in three years has to be moved by its top number, its bottom number, or the way it is measured. Take each in turn.
The cost did not do it. The report’s own modeled job cost went from $3,611 to $4,672. That is up 29.4 percent over seven years, slower than general inflation over the same span. The cost line is the steadiest thing in the table.
The door market did not do it either. We checked on our own. The BLS producer price index for metal doors (except storm doors) went from an annual average of 267.6 in 2019 to 452.3 in 2025, up 69.0 percent.
So the factory price of the product rose more than twice as fast as the report’s modeled job cost. That gap matters on its own. It suggests the modeled cost understates what happened to garage door prices, which would push the ratio up, not explain it away.
That leaves the resale side, and the resale side is a survey. The surveyed resale value went from $3,520 to $12,507. Between the 2023 and 2024 editions alone it nearly doubled, rising 98.1 percent in one year. Then it added another 42.9 percent.
We do not know why, and we will not make up a reason. Several things are possible, and nothing we can retrieve proves any of them. The survey panel or its weighting may have changed. The way the project was described or pictured may have changed.
Agents’ beliefs about curb appeal may really have shifted, in a slow market where presentation mattered more. Or the 2024 edition covering 23 projects instead of 28 may have changed what respondents compared against.
What we can say with confidence: nothing about garage doors themselves explains a 255 percent rise in perceived value against a 29 percent rise in cost.
The lesson applies beyond garage doors. When a ratio jumps and its bottom number holds steady, ask how the top number was measured before you ask about the thing itself. Reading 268 percent as a fact about doors skips that question.
Which projects top and bottom the 2025 table?
Small outside replacements top the table; big additions sit at the bottom. Everything above 200 percent is a small, visible, exterior replacement. Everything below 50 percent is a large structural addition. Whatever the survey captures, it is mostly about how a house looks on arrival.
| Project | Job cost | Resale value | Cost recouped |
|---|---|---|---|
| Garage Door Replacement | $4,672 | $12,507 | 268% |
| Entry Door Replacement | Steel | $2,435 | $5,270 | 216% |
| Manufactured Stone Veneer | $11,702 | $24,328 | 208% |
| Siding Replacement | Fiber-Cement | $21,485 | $24,420 | 114% |
| Minor Kitchen Remodel | Midrange | $28,458 | $32,141 | 113% |
| 18 projects in between, not shown | — | — | — |
| Roofing Replacement | Asphalt Shingles | $31,871 | $21,501 | 68% |
| Major Kitchen Remodel | Midrange | $82,793 | $42,130 | 51% |
| Accessory Dwelling Unit | $166,406 | $68,656 | 41% |
| Major Kitchen Remodel | Upscale | $164,104 | $58,561 | 36% |
| Primary Suite Addition | Midrange | $170,517 | $55,097 | 32% |
2025 Cost vs. Value national averages, read from an Internet Archive snapshot of jlconline.com, retrieved 2026-09-06. The five highest and five lowest of the 28 projects. Asphalt roof replacement, also an exterior replacement, sits at 68 percent. That is a useful check on any simple “outside work pays more” reading.
How to use the ROI number without being misled
The figure is not useless. It is evidence of one specific thing, and it is worth exactly that.
Treat it as evidence about looks
A large panel of people who watch buyers react to houses for a living think a new garage door adds more value than it costs. That is a real signal about curb appeal. It fits the rest of the top of the table.
Do not budget against it
“It pays for itself two and a half times over” is not something the method supports. If you spend $4,672 expecting to get $12,507 back at sale, you are spending against a survey answer. Nothing in the report promises that result for your house.
Watch the order, not the level
A survey-based series is most useful for comparing. Garage doors have topped the table two years running, above steel entry doors and stone veneer. That order is probably more reliable than any single percentage in it.
Remember it is a national average of 119 markets
The report publishes regional and metro breakdowns for a reason. A national figure for a project valued for street looks hides huge differences. Compare a market where every house has a big front double garage with one where garages sit behind the house.
Notice who is quoting it to you
HyreGarage analysis: when we searched for this figure, the loudest sources of the 268 percent number were garage door makers, on their own marketing pages. That does not make it wrong. It does mean most homeowners hear it from someone with an interest in it, without the method that limits it.
Ask what your door actually needs
The real decision is usually not about ROI. Is the door failing? Are the panels damaged past a sensible repair? Are the springs and opener near the end of their lives? Those questions have answers; our repair or replace tool walks through them. This one has a survey.
What do the report’s terms mean?
- Job cost
- The modeled cost of having the project done. It is built from a BLS/BEA price index set to 2019 and adjusted locally with Verisk’s XactRemodel estimating data. It is not a quoted price or an average of invoices.
- Resale value
- What a surveyed Realtor estimated the project adds to a home’s sale price. It is an opinion, gathered by online survey from a panel of over 6,000 respondents, with the National Association of Realtors’ Realtors Property Resource.
- Cost recouped
- Resale value divided by job cost, shown as a percentage. Both inputs are estimates, so the ratio carries the uncertainty of both. The report does not publish a margin of error for either.
- XactRemodel
- Verisk’s commercial remodeling estimating product, named in the method as the source of local cost differences. It is proprietary and its inputs are not public. That limits how far anyone can check the cost side.
- PPI (producer price index)
- A BLS measure of the prices US makers receive. We use it as an outside check on the cost side. Series PCU3323213323211 covers metal doors other than storm doors, back to a June 1983 base.
- Anchoring
- A well-known survey effect: a number shown to a respondent shapes the number they give next. It matters here because respondents saw construction costs before they were asked what value the project adds. We raise it as a feature of the survey, not a proven bias in this result.
What we could not find, and so did not write
The live report. jlconline.com returned HTTP 403 to every request we made on the retrieval date, including the 2025 edition and the methodology page. Everything here comes from Internet Archive snapshots. Archived pages can be partial or stale, so confirm these figures against the live report before acting on them.
Regional and metro figures. The report breaks results down across 119 markets and nine regions. We did not compile them. Doing it right means pulling every regional table from every edition, and a partial regional picture is worse than none for a project whose whole story is local.
The survey itself. We have the report’s description of the survey. We do not have the questionnaire, who was on the panel, the response rate, the weighting, or whether any of these changed between the 2023 and 2024 editions. That is exactly what would explain the jump, and it is not published.
Any independent measure of real sale value. A study comparing sales of houses with and without a recent garage door replacement would answer the question people think the report answers. We know of none, found none, and did not build one. That needs sale-level data we do not hold.
And one thing we chose not to do: fit a trend line to the seven data points. Seven yearly readings of a survey that may have changed do not support a trend. A fitted line would hide the most interesting feature of this data, a sudden jump, behind a smooth slope.
Questions
Do garage doors really have a 268% return on investment?
How is the Cost vs Value report calculated?
Has the garage door always had a high ROI?
Why did the garage door ROI figure jump so much?
Will a new garage door add $12,507 to my house?
Is the Cost vs Value report reliable?
What is the highest ROI home improvement project?
Does the report use real home sale data?
Why does asphalt roof replacement score so much lower than a garage door?
Where can I check these figures?
Written and audited by
HyreGarage Research Desk
Primary-source research, data analysis and fact checking
We are not a garage door company. We read the agency file, the code record, the standards document or the public register ourselves, compute the figure from it, and publish it with the source and the date we retrieved it.
Where a number cannot be traced to a primary source, we publish the shorter page and say what we could not verify. Our own company records cover ten states; nothing national is ever derived from them.
- 10
- states our own company records cover — and the limit of any claim made from them
- 3,901
- garage door companies in the store
- 457
- license records verified against a state board
- 0
- national claims made from a ten-state store
How this desk works
- Primary sources only. Injury counts come from CPSC. Housing counts come from the Census file, not from a summary of it. Code history comes from the building commission that adopted the code. We do not cite an article that cites a source; we retrieve the source and do the arithmetic ourselves.
- Every figure carries its retrieval date. Registers change, datasets are revised and codes are amended. A number without the date it was read cannot be checked, so every study states one.
- Fact, calculation and analysis are labeled apart. A quote is a quote, a HyreGarage computation says so, and an interpretation says “HyreGarage analysis”. Presenting our reading of a dataset as something the agency stated would be the easiest way to lose the only thing this desk is for.
- Limitations go above the fold. If a figure is an upper bound, a bracket, or an association rather than a cause, that is said before the figure is quoted rather than in a footnote underneath it.
- No DIY instructions for spring, cable or track work. Those components hold enough stored energy to cause serious injury, and CPSC records the consequences. We describe what has failed and what a competent repair involves; we do not tell you how to do it.
Data as of Cost vs. Value 2019-2025 editions and BLS series PCU3323213323211, retrieved 2026-09-06. Authorship on this site is organizational: the analysis belongs to the desk rather than to a named individual, and we do not publish credentials we do not hold.
Our editorial policy sets out how we source, date and correct what we publish.
Sources & retrieval dates
Replacing a door rather than an investment?
Ask every company for the same written scope: door construction, spring cycle rating, track and hardware, disposal, and both warranties. That is a comparison you can actually make.
HyreGarage is not a garage door company and does not perform, supervise or warrant garage door work. Nothing on this page is financial advice or a prediction of what any improvement will return on any individual property.