The short version: check who paid for the review, whether the reviewer used the product for more than a weekend, and whether a single negative sentence appears anywhere in it. If the answer to any of those is "unclear" or "no", treat the review as an advert and move on. I've been burned enough times buying software off glowing reviews to have built an actual checklist for this, and I'm giving you mine.
Why I stopped trusting star ratings years ago
In 2019 I bought an email marketing tool because it had a 4.8 out of 5 on a well known review site and three separate "best of" blog posts had it at number one. I paid for the annual plan, roughly £900, to save 20 percent. Within six weeks I found out the tool couldn't handle a list segmented by more than two conditions without support tickets and a two week wait. I went back and read the reviews this time, not skimmed them, and every single glowing one mentioned "getting started" or "onboarding" and nothing about using it after month three. Nobody had used the thing long enough to hit its limits before writing about it.
That's the pattern I now look for first. Reviews written in the first two weeks of using something are marketing, even when the reviewer means every word. You can't know a CRM, an AI tool, or a piece of accounting software has a fatal flaw until you've hit the edge case it wasn't built for, and that takes months, not a launch week trial.
The uncomfortable bit nobody wants to say out loud
Here's the part most people writing "how to spot a fake review" guides skip over. A huge share of tech reviews, including on sites people treat as neutral, are written by affiliates who get paid every time someone clicks through and buys. That's not automatically dishonest. I do sponsored and affiliate work myself, it's one of the income streams that kept my business alive during a run of very hard years, and plenty of affiliate reviewers are straight with their readers. But the commission structure changes what gets written, always, whether the writer notices it or not. A tool paying £150 per signup gets a glowing 2000 word review with screenshots. A tool paying nothing gets ignored, no matter how good it is. That's the actual economics of most "top 10 tools for X" posts you're reading, and almost nobody who writes those lists will tell you that plainly.
So the question isn't "is this review paid" because most are, one way or another, through affiliate links, gifted licenses, or a sponsorship deal. The question is whether the writer discloses it and whether they still found something to criticise despite the money on the table.
My actual checklist, step by step
This is the exact process I use before I trust a review enough to act on it, whether it's for a piece of AI software, a CRM, or a marketing agency.
- Step 1: Find the disclosure. Scroll to the bottom or check the top for "affiliate links" or "we were given free access". No disclosure at all on a site that clearly monetises reviews is itself a red flag, not a reassurance.
- Step 2: Look for one specific criticism. Not a vague "it has a learning curve" throwaway line, an actual named limitation. "The API rate limit is 100 calls a minute on the starter plan, which broke our integration" is a real review. "It's not perfect but overall we love it" is filler.
- Step 3: Check the date and check for updates. Software changes fast. A review from 2023 praising an AI tool's accuracy is worthless in 2026 if the model behind it has been swapped three times since. Good reviewers update posts and say when they last checked. Most don't.
- Step 4: Search for the reviewer's name plus the word "complaint" or "refund". Takes 30 seconds. If someone reviews twelve different project management tools a month and gives all twelve five stars, they're not reviewing, they're cataloguing.
- Step 5: Check where the traffic goes. Hover over the buy link. If it routes through a tracking domain rather than straight to the company's site, that's an affiliate link. Fine to still trust the content, just weight it accordingly.
- Step 6: Read the comments or replies, not just the post. On YouTube especially, the top comment under a glowing review is often someone saying the exact opposite happened to them. That's more useful than the video.
Real examples of getting this right and wrong
When I was researching AI writing tools for clients last year, one review site had Jasper, Copy.ai, and Writer all scored within 0.2 of each other on a 10 point scale, with near identical wording across the three write-ups. That's a template being filled in, not three separate people testing three separate products. Compare that to a review I read on a niche B2B software forum where the writer said plainly that a well known project management tool "crashes on Chrome if you have more than 40 tabs open, which sounds like a niche problem until you realise every marketer alive has 40 tabs open." That's a person who used the thing and hit a wall. I trusted that review and ignored the polished ones.
The same logic applies well beyond software. If you're comparing agencies rather than tools, the same red flags show up: testimonials with no name attached, case studies with no numbers, "we increased traffic" with no baseline given. I wrote a longer piece on choosing the right SEO or digital marketing agency that covers exactly this, because the vetting process for hiring a company is nearly identical to vetting a review of a piece of software. You're looking for specifics, dates, and named limitations in both cases.
What to do when reviews conflict
Sometimes you'll find one reviewer saying a tool's customer support is brilliant and another saying it's dreadful. Both can be true. Support quality varies by plan tier, by time zone, and by how busy the company is that quarter. When I see this, I check which pricing tier each reviewer was on, since companies quietly prioritise the £500 a month accounts over the £15 a month ones, every time, and no review mentions this unless you ask.
If you're weighing up whether to bring in outside help to pick and implement the right tools rather than gambling on reviews yourself, that's a fair reason to work with someone who tests this stuff for a living. I cover what that costs and what it involves on my AI consultant costs page, because paying for a few hours of proper testing is often cheaper than the £900 mistake I made in 2019.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
A quick number to keep in mind
Roughly 4 percent of online reviews across all categories are estimated to be fake according to research cited by outlets like the BBC, but that figure understates the problem for tech specifically, because the more damaging issue isn't outright fake reviews, it's real reviews written too early, under a sponsorship, or by someone who never used the product past the free trial. Fake reviews get flagged and removed eventually. Sponsored ones that read as neutral almost never do.
The five second gut check I use now
Before I read a single word of a review's content, I ask myself three quick things: would this person still recommend the tool if they'd paid full retail price out of their own pocket, would they still recommend it after using it for six months rather than two weeks, and is there anywhere in the piece where they admit it isn't right for a specific type of user. If a review passes all three, I read on. If it doesn't, I treat it as an advert with extra steps, which, honestly, is usually exactly what it is.
Frequently asked questions
How can you tell if a tech review is sponsored?
Check the top or bottom of the page for a disclosure statement, hover over any "buy" or "try it" links to see if they route through a tracking domain, and notice if every product the site reviews scores suspiciously similarly. Genuine sponsorship isn't dishonest on its own, but it should be disclosed clearly, not buried in tiny grey text.
Are star ratings on review sites reliable?
Not on their own. A 4.8 average tells you nothing about how long reviewers used the product, whether negative reviews were removed, or whether the rating reflects the free plan or the expensive one. Read the actual text of at least five reviews, including the lowest rated ones, before trusting the number.
Why do so many tech review sites recommend the same tools?
Often because those tools pay the highest affiliate commissions, not because they're objectively the best. Higher paying software companies get featured more often and reviewed more favourably across dozens of sites, which creates a false impression of consensus that's really just a shared payout structure.
Should I ever trust a review with no negatives in it?
Be cautious of it. Every piece of software has a limitation, a plan tier that's stingy, or a feature that's clunky. A review with zero criticism has either not tested the product thoroughly or has an incentive to only show the positives.
Related reading: ActiveCampaign Marketing Strategy: How They Built a Brand That Wins and Business Lessons from Alan Sugar.
