Reviews, yours and theirs
Google, Productreview.com.au, Facebook and industry sites. Competitors' one- to three-star reviews show the gaps you can fill.
Best for: Always. It's the fastest way to hear the market in its own words.
Buyer psychology research finds out why people choose you, or don't, in their own words. I go through what buyers have already written and typed, not just what they tell you when asked: reviews (yours and competitors'), search terms, forums, enquiry forms and sales calls. That matters because people are often poor at explaining their own choices. In one classic study, shoppers picked the right-hand pair of identical stockings about four times as often as the left, and almost all denied position had anything to do with it. What comes out is the words to use, the worries to answer, and the order to answer them in.
From my audits
These are the patterns that come up most when I look at how a business describes its customers.
The website says 'solutions' and 'integrated services'. Customers type 'blocked drain', 'tax return for sole trader', 'cheap boxes for moving'.
A stock photo, a made-up name and a list of hobbies. Nobody can say which data it came from, and nobody uses it.
The good reviews end up on the website. The three-star ones, which say exactly what nearly stopped the sale, get ignored.
Marketing writes for one person, sales talks to another, and the ads keep finding the first.
The complaints about competitors are public, specific and free. They're the clearest list of what the market wants and isn't getting.
What the research says
Asking customers why they bought is a start. It's rarely the whole answer. Three findings shape how I do this work.
Nisbett and Wilson argued people often have little direct access to why they did something, and answer from a plausible theory instead.
What I do: Trust what people did (searched, clicked, complained about) more than what they say they'd do. That's why I work from data and reviews, not surveys.
In Christensen's milkshake research, about 40% of milkshakes were bought early in the morning, by commuters wanting something to make the drive less boring. The job, not the customer profile, explained the sale.
What I do: For each service, find the job it's hired for, and the moment that triggers it: a burst pipe, a tax letter, a new hire.
Northwestern's Spiegel Research Center found a product with five reviews was 270% more likely to be bought than one with none, and purchase likelihood peaked at ratings of 4.0 to 4.7, not 5.0.
What I do: Mine your three-star reviews as hard as your five-star ones. The complaints buyers accept tell you what matters most.
Studies linked so you can read them. I'm a marketer, not a psychologist: my job is turning research like this into what your ads and pages say.
Review miner
Paste reviews, one per line. Yours, or a competitor's from Google or Productreview.com.au. It counts the themes, which ones are complaints, and the phrases people repeat. Twenty or more reviews gives a fairer picture.
The example is ten made-up reviews for a plumber. Nothing you paste leaves your browser.
Reviews read
0
Themes, by how many reviews mention them
Their words: phrases in more than one review
A first pass, using simple word lists. It can't read sarcasm or context, so read the reviews behind any theme before you act on it.
Where I look
Most of the answers already exist. They're just spread across places nobody reads together. I don't interview your customers: I work from what they've already written and searched, plus what your team knows.
Google, Productreview.com.au, Facebook and industry sites. Competitors' one- to three-star reviews show the gaps you can fill.
Best for: Always. It's the fastest way to hear the market in its own words.
Your Google Ads search terms report, Search Console queries, autocomplete and 'People also ask'. What people type when nobody is watching.
Best for: Always. It shows the questions and the order they come in.
Reddit, Whirlpool, Facebook groups and industry forums. Google now shows forum discussions in its results for questions that benefit from personal experience.
Best for: For anything people research hard, or complain about in public.
What people write in your contact form, and the questions they ask on the phone. Your team hears the real worries every day.
Best for: Always, if you have them. Often the richest source, and the least used.
What ChatGPT and Google's AI say about your category, who they name, and which worries they raise first.
Best for: For any business that wants to be recommended by AI tools.
A short set of questions for you and whoever answers the phone: who buys, what they ask first, and why people say no. Your team hears the real worries every day.
Best for: Always, at the start. It points the data work at the right questions.
How it works
I look at how you describe your customers now, and where the words don't match what they search.
Reviews, search terms, forum threads, enquiry forms and sales call notes, yours and your competitors', plus your answers to a short discovery questionnaire.
Every comment tagged by job, trigger, worry and phrase, then counted, so the big themes aren't just the memorable ones.
A short brief: the jobs, the worries in order, the phrases to use, and the ad and page angles that follow.
What you get
Signals this works on
Out of the 8 signals every page and ad sends.
Working out why people choose a business like yours, what worries them, and the words they use, from what they've written, searched and said. The output is what your ads and pages should say, and in what order.
Market research usually sizes a market or tracks opinions with surveys. This is narrower and more practical: the exact words, jobs and worries behind a purchase, gathered mostly from what buyers have already done, so it can go straight into ads and pages.
Yes, for research. Reading public reviews to understand a market is normal. Don't copy them onto your site or pass them off as yours: that's misleading, and the ACCC watches reviews closely.
No. I work from data and reviews: what your buyers and your competitors' buyers have already written, searched and asked. I also send you a short discovery questionnaire, because you and your team hear the real questions every day.
They're common, which is why I read for patterns, not single reviews. In a 2023 sweep, the ACCC found 37% of the 137 businesses it reviewed had engaged in concerning conduct with reviews, including fake positive reviews and steering unhappy customers away from posting.
Usually two to three weeks, depending on how much data you have. A first read of reviews and search terms often turns up something useful in the first few days.
Takes a minute. Audits come back within 1 to 2 business days. For anything else, I'll get back to you as soon as possible.
What happens next
Keep going