What actually goes wrong on residential construction projects

Segment

Rank by

Click any failure mode to see the numbers, what customers actually said, and what removes it.

How to read this

How often counts every record naming that failure mode. How damaging is an index built from two observed signals, the star rating the customer gave, and whether they took the complaint to a formal dispute body. Nothing is modelled or estimated; there are deliberately no invented dollar figures anywhere in this research.

The two rankings deliberately disagree. Communication is the most frequent complaint in the market and one of the least severe, customers rarely escalate over it, they simply stop recommending you. Warranty response is comparatively rare and by far the most dangerous: nearly half of those complaints reach a formal dispute body, and they happen after the money has been collected.

What this research cannot see

Google is no longer absent. The ten largest custom and design-build firms were read directly, sorted lowest-rating first, 1,172 reviews screened. Google remains unread for the other 57 firms in the study, and Yelp and Houzz still block collection entirely.
Review platforms only capture people who became customers. A homeowner who asked for a bid and never heard back doesn't leave a review, they hire someone else. Estimating failures are therefore structurally under-counted, not unimportant.
Survey programmes are commissioned by the builder and formal complaint records cover a limited window. Both channels understate dissatisfaction. This research indicates a pattern; it is not a census.

Sources: verified post-project customer surveys and public complaint records, July 2026; Google Maps reviews read directly through an authenticated browser session, 26 July 2026; Oregon Construction Contractors Board complaint history, read from the state system and name-verified per firm; CCB Governor's Recommended Budget 2025-27; CCB legislative presentation, February 2025. Individual firms are not named. Record-level data with source URLs retained.

Research and analysis by Marty Kribs, acre-mgt.com. Version 3, July 2026, external release. Individual firms are anonymised; the named research file is retained internally. May be cited with attribution.

What the Google channel added

The model above was built from post-project surveys and formal complaint records. Both only capture people who became clients. Google captures the ones who did not. In July 2026 the ten largest custom and design-build firms in the study were read directly on Google Maps, sorted lowest rating first: 1,172 reviews screened, 329 loaded, 51 negatives captured. Ratings and counts are as Google displayed them.

Firm	Segment	Market	Rating	Reviews	Negatives	Negative rate
Firm A	Custom home builder	Portland Metro	4.1	49	13	26.5%
Firm B	Custom home builder	Central Oregon	4.6	38	5	13.2%
Firm C	Design-build	Portland Metro	4.6	240	21	8.8%
Firm D	Design-build	Corvallis	4.7	47	3	6.4%
Firm E	Design-build	Portland Metro	4.8	111	6	5.4%
Firm F	Design-build	Corvallis	4.9	62	1	1.6%
Firm G	Design-build	Portland Metro	4.9	204	2	1.0%
Firm H	Design-build	Corvallis	5.0	213	0	0.0%
Firm I	Design-build	Corvallis	5.0	174	0	0.0%
Firm J	Design-build	Salem	4.9	34	0	0.0%
Ten firms				1,172	51	4.35%

Firms are identified by letter. Segment, market, rating and review counts are reported as observed, because the pattern is the point and the identity is not. Letters run in order of negative rate.

1 in 6of the negatives describe a failure that happened before any contract existed. At two firms, every negative was pre-contract.

2.3% vs 6.3%One-star rate for the same firm over the same period: post-project survey versus Google. Nearly triple. This is channel bias, measured.

3 of 10firms carry a genuine zero: no review below four stars across 421 reviews between them. Excellence is visible in this data, not just failure.

Firm E reported zero negatives across 374 post-project surveys and six across 111 Google reviews. Firm D reported zero across nine surveys and three on Google, all of them pre-contract. The survey channel is not dishonest; it is structurally blind to everyone who never signed.

Caution on sample size: at 51 records the ranking of categories within the Google channel is not yet stable, and several categories carry only one or two instances. What is solid at this n: the per firm negative rates, the tier rate, the channel comparison, and the pre-contract finding.

What to do with this

This is a market read, not a verdict on any one company. Every failure mode above is a management process that can be installed, measured, and kept. Pick the path that matches where you are.

The answer

What we do about this

Every failure mode above is a missing process, not a missing person. The bar, the system that runs the day against it, and the training that gets your team there.

See how it works → The map

The System Map

Every finished job leaves 53 pieces of proof. Six rooms by seven areas, on one grid. Free, takes an email.

Open the map → Four minutes

The Operating Gap

Ten questions about how the company runs. A ranked install list, not a grade.

Run the check → Go further

Measure your own operation

A documented baseline of where information stalls in your company, with ranked findings and agreed success metrics.

Request the Diagnostic →

Working with someone who should see this?

Most of the people who recognise these patterns are not the ones who can act on them alone. If that is you, here is the short version to send along.

Found some market research on what actually goes wrong on residential construction jobs. It is 290 real customer records across 67 Oregon builders. The headline: management failures outnumber workmanship failures about four to one, and 41 percent of negative feedback never shows up in a star rating. Worth five minutes.