Are Directory Reviews Trustworthy?

Are Checkatrade reviews real?

In most cases yes. Individual reviews on a paid directory are usually genuine customers describing genuine work, and platforms that charge a membership fee have a commercial reason to moderate obvious fakes, because a directory full of invented praise is worth nothing to anybody. The question worth asking is not whether each entry is real. It is which jobs got a review request at all.

A directory review set is not a sample of a business. It is a sample of the jobs that business chose to put forward, collected at the moment the customer was most relieved. Everything awkward sits outside the frame: the job that overran, the second visit that was argued about, the customer who went quiet and rang somebody else. None of that is hidden by anyone. It simply never enters the system, because nobody hands a feedback card to a person they have just fallen out with.

So read the rating as evidence of one thing, that the business does a reasonable volume of work and usually finishes without a row. That is worth knowing. It is not the same as knowing the engineer can tell a scaled cylinder coil from a motorised valve that is not driving over, and the review set cannot tell you that.

How a paid directory actually collects reviews

The mechanism is simple and it is published openly by the platforms themselves. A trade business pays to be listed. Feedback is then gathered either from a customer list the business supplies, or from a card, link or code the engineer hands over at the end of the job. The platform checks that the reviewer matches a real job before publishing. At no point in that chain does anybody go looking for the customers who were unhappy.

Notice who the paying customer is. Householders do not pay to search a directory, so a directory is funded by the trades listed on it. Its incentive is to keep trades subscribed and keep the listings looking good enough that householders keep arriving. Everyone behaves rationally and nothing unlawful happens. The output is still a page built by the people being judged.

The same applies to position inside a directory. On most platforms where you can be listed, you can also pay to be more visible, through a featured slot, a higher tier or a larger tile at the top of the category. Whenever a position can be bought, it is telling you about a marketing budget, and a business that pays for each enquiry recovers that cost out of the jobs it wins. That is arithmetic rather than an accusation, and it is why the order of a results page is not a ranking of skill.

Check whether the list you are reading is ordered by relevance, by distance, or by what the listed businesses paid. Most platforms say which, usually in small text at the top of the results or under a label.

The selection bias that decides what you see

Reviews get collected at handover, the most flattering moment of any plumbing job. The water is hot again, the mess is cleared, and the invoice has either just been paid or has not yet been questioned. Ask for feedback there and you get warmth. Ask the same customer six months later, after a fault has come back, and you would get something else, but nobody is asking then.

There is a second filter on top of that one. Engineers are human, so the request gets made when the job went well and quietly skipped when it did not. Nobody has to instruct anybody. It happens at every firm, and it is why almost all directory averages cluster near the top of the scale. A rating everybody scores highly on is not a discriminator.

Timing matters more here than in most trades, because plumbing faults are slow. A plate heat exchanger blocking with scale, a cylinder coil losing performance, a weeping compression joint behind boxing, a relief valve that has started to pass, none of those announce themselves on the day. London water is hard right across the supply area, so scale related problems surface a season or two after the visit that should have caught them. The review was written in week one. The verdict arrives in month nine and is never posted.

Then there is what the customer can even judge. In a purpose built block with a thermal store behind a panel, or a conversion where pipework runs under four generations of other trades, most of the work is out of sight. You can see that the engineer was polite and that hot water came out of the tap that evening. You cannot see whether the right component was replaced or the symptom was simply masked.

What a five star average is actually measuring

It measures the visit, not the plumbing. Punctuality, communication, tidiness, whether the final figure matched what was indicated, and whether the immediate symptom went away. Those things are real and they matter, but they are the customer experience layer sitting on top of the technical work, and a job can score full marks on all of them while the diagnosis underneath it was wrong.

Consider what a wrong but satisfying job looks like. No hot water, an engineer attends, an immersion element gets replaced, hot water returns, five stars. If the element had scaled and failed, that was the right call. If the element was fine and the real fault sat in the cylinder coil or the motorised valve, the customer still got hot water that evening, from the immersion rather than the boiler, and still left a glowing review. Nobody found out until the electricity bill arrived or the fault came back.

The same gap shows up in work nobody looks at. Whether isolation valves were left serviceable rather than seized, whether new pipework was supported instead of resting on a joist, whether the discharge above a pressurised cylinder actually terminates safely rather than merely existing. None of that is visible from the hallway, so none of it ever reaches a star rating.

• A review can tell you: they turned up, they were polite, they cleared up, the price matched what was said.
• A review cannot tell you: the diagnosis was right, the correct part was changed, the work is still sound next winter.
• A review almost never covers what happened when something went wrong later.

Do vetted, checked or approved badges mean anything?

They mean paperwork was checked once, usually at joining, against documents the business supplied itself: identity, an address, insurance, sometimes a registration number and references the applicant chose. Better than nothing, and far short of what people assume, which is that somebody independent went out and inspected the work.

Checkatrade vs Trustatrader is not a comparison that decides anything for a customer, because both sell the same membership product to the same trades and both check the same kind of paperwork at the door. The badge that sits on a van tells you which platform the business chose to advertise on, not which one holds it to a higher standard.

The gap that matters for hot water is that badges attach to companies while qualifications attach to people. A firm can be a member in good standing while the individual who turns up is not signed off for the system in front of him. Directory membership is not a gas registration and it is not an unvented qualification.

Check what can be checked at source rather than trusting a logo. For gas work, look the business up on the official register and read the back of the identification card on the day, because that is where the categories the person is signed off for are listed. For a pressurised unvented cylinder, ask what the attending engineer holds and expect a straight answer.

A badge on a website is an image file. It costs nothing to place and it is not evidence of anything. The register the badge refers to is the evidence, and it is free to search.

How to read a review set critically

Ignore the average and read the fifteen most recent entries, looking for specifics rather than sentiment. A review naming the fault, the part and the property type is worth more than fifty saying great service, highly recommended, will use again. Detail is the closest thing to proof that a real job sits behind the words.

Look for jobs that resemble yours. A page full of five star bathroom installations tells you nothing about whether the same firm can diagnose a thermal store in a block or trace a leak under a liquid screed floor in a new build. Hot water work splits into separate skills and the review page never separates them, so you have to.

Check the spread of dates. A steady trickle over years looks like an ordinary working business. A cluster arriving in one fortnight after a long silence means a push to collect them, which is not sinister but tells you the set is harvested rather than natural. Watch the language too: several reviews sharing the same structure and adjectives usually came from the same prompt.

• Read the most recent fifteen, not the average.
• Prefer reviews that name a fault, a part or a system type.
• Find reviews for work like yours, not just any work.
• Watch the dates for harvested clusters after quiet periods.
• Discount anything generic enough to fit any trade in any town.

The one bad review is the most useful thing on the page

Go straight to the lowest rated entries and read the replies rather than the complaints. How a business answers a public complaint tells you almost everything about how a dispute with you would be handled, and unlike the praise it was not collected at a flattering moment.

A good reply is specific and dull. It names the date, says what was attended, what was found and what was offered next. That tells you the firm keeps proper job records, which is exactly what you want behind you if a warranty claim or a second opinion becomes necessary. A firm that can reconstruct a job months later wrote things down at the time.

A bad reply attacks the customer, or announces the person was never a customer at all, or says nothing while the complaint sits there. Silence is the weakest of the three, because it means nobody is reading the page, and nobody would be reading it when the complaint had your name on it.

Be suspicious of a long, high volume history with no low ratings in it at all. Plumbing involves opened up floors, unexpected parts, access problems and price conversations under stress, and everybody who does enough of it has jobs go sideways. A spotless record over years usually means the unhappy ones were never asked.

Are Google reviews any better than directory reviews?

They are differently flawed and worth reading alongside, not instead. Nobody has to pay a membership fee to be reviewed on a map listing, and the business does not control who gets to post, so unhappy customers can and do arrive unprompted. That removes the biggest structural filter on a paid directory. In exchange you lose the job verification, so there is nothing stopping an unrelated person leaving either praise or a complaint.

In practice that makes map reviews better for spotting a negative pattern and worse for trusting any single positive entry. If several people over a couple of years describe the same behaviour, one figure quoted and another invoiced, repeat visits for the same fault, nobody answering after the money cleared, the pattern is the signal. One furious review with no detail is not.

One thing applies to both. Position on the page is not a verdict. Sponsored blocks and ad labelled map entries are bought at auction, and a business paying per click recovers that money from the jobs it converts. Scroll past the labelled results and read the reviews of whoever sits underneath.

What to do instead of trusting a rating

Use the review set to build a shortlist, then test it on the phone, because three minutes on a call beats any number of stars. Listen for whether they ask what you have. A firm with engineers behind it wants to know whether the hot water is stored or instantaneous, whether the cylinder is vented or a pressurised unit with a tundish above it, and what changed just before the fault. A firm that only wants the booking asks for your postcode and your card.

Then ask for the charging structure in writing before anybody travels. Not a total, because nobody can honestly total an unseen fault, but the structure: fixed attendance fee or hourly rate, what the first charge covers, what the rate becomes after that, how parts are charged. Any business that prices honestly has that paragraph saved already. That one request separates firms faster than any review page.

Ask who is attending and what they mainly work on. We run 20 engineers and they are not interchangeable: the one servicing pressurised cylinders all week is not the one for a thermal store in a block, and neither is the one you want tracing a leak under a screed floor. Any firm with a real spread of skills can say who is coming and why. A business that does unvented cylinder repair most weeks will name the qualification the attending engineer holds without being pushed, and a business that does it occasionally will change the subject.

The strongest move costs almost nothing and nobody does it. Give a firm a small job on a quiet weekday, a service, a dripping tap, a stopcock that will not turn, and watch how they work with no pressure and no water on the floor. You learn more from one uneventful visit than from a week of reading reviews, and you end up with a number in your phone and an engineer who has seen your system. Choices made with water coming through a ceiling get made from the first result on the screen, which is where ratings do the least good.

Reviews are a filter, not a decision. Use them to remove obvious problems, then judge the firm on the call and on what it will confirm in writing.