The pilot ran. Here's what came back

61 practices, 1,966 real reviews, 34 cents. Three groups: practices whose own websites read as offering injectable testosterone to women, national menopause telehealth brands, and a control of 22 ordinary OB/GYN practices pulled straight off a category search in six cities so they couldn't be cherry-picked.

The headline is that "menopause" is the wrong thing to search for. It barely separates the groups at all. 41% of ordinary OB/GYN practices have at least one review mentioning menopause, which is actually higher than the 35% among the testosterone practices. Search reviews for the word menopause across the country and you will mostly find gynaecologists, because menopause is simply something gynecology patients talk about.

What does separate them, sharply, is hormone language: any mention of hormones, HRT, BHRT or TRT.

Percentage of practices in each group with at least one review mentioning the term. Based on up to 50 recent reviews each.

Three things fall out of that table. Hormone language is a genuine classifier, hitting roughly three quarters of both target groups and one in eleven of the control. Testosterone and pellet language appear in the target groups and never once in 939 control reviews. And injection language cleanly separates the two business models: 71% of the in-person testosterone practices, zero of the telehealth brands, because telehealth ships a prescription rather than putting a needle in your arm. You can read a practice's delivery model off its reviews.

The two reviews that justify the whole idea

Only two of the 22 control practices had any hormone language. Both are worth reading, because they are exactly what this method is for and neither could be found from a website.

The first is a plain OB/GYN group in Columbus with no menopause branding anywhere. A patient wrote: "I've slept two full nights in a row and haven't done that in at least two years. I went to this group for HRT and am delighted with their collaborative approach." That is a practice quietly doing menopause care and not saying so publicly. It is a prospect, and no website scrape would ever have surfaced it.

The second is the mirror image, from a Tucson practice: a patient describing a provider "reluctant to put me on HRT because I should expect to have less libido since I am old and in menopause." That is a practice actively losing this market, and a patient going unserved. Whether that reads to you as a prospect who needs help or a competitor worth knowing about, it's information a website will never volunteer.

That asymmetry is the real argument for reviews over websites. A website tells you what a practice wants to sell. Reviews tell you what patients actually received, including when it went badly.

Where it breaks: telehealth

The match rates came out as feared, and the gap is the finding. Ordinary OB/GYN practices matched a Google listing 100% of the time, because a physical clinic in a city is exactly what Google Maps is for. The testosterone practices matched 77%. Menopause telehealth matched 53%.

Worse, the telehealth failures are silent unless you check. "Midi Health" resolves to MIDI Product Development, an engineering consultancy in Smithtown, New York. "Winona" resolves to Winona State University. "Hone Health" resolves to Home Town Health Care. Each of those has reviews, and each would have been read as evidence about a menopause telehealth service. The pilot only avoided that by verifying every match against the practice's own website domain, and rejecting any listing whose website didn't match.

So the method is strong for practices with a front door and weak for the ones without. That's an awkward result, because telehealth is where much of this market is going. If telehealth prospects are the goal, Google reviews are the wrong instrument and Trustpilot or app-store reviews would be the place to look instead.

Why the literal version is impossible

Google's Places API does have a text search, and it does return reviews. It sounds like the answer. It isn't, for four separate reasons, and each one alone would be fatal.

The search matches business names, addresses and categories. Review text is something you can receive back, never something you can search against. So there's no query that means "show me practices whose patients wrote about menopause."

Even once you've found a practice, Place Details returns at most 5 reviews, with no way to page through the rest. That cap has been an open feature request since 2015. Five reviews is nowhere near enough: a practice with 200 reviews might mention menopause in twenty of them and none of those twenty need be in the five you're shown.

The Google Business Profile API, which does expose full reviews, only works on locations you already own or manage. It's built for a business reading its own reviews, not for researching anyone else's.

And the part that budget can't solve: the Maps Platform terms let you cache latitude and longitude, for 30 days. That's the whole caching permission. Review text isn't covered by any exception, and the license separately forbids exporting Maps content for use outside the service. Storing review text in your own database breaks Google's terms even when you paid Google for it through the official API.

What the data vendors can and can't do

There's a whole industry retrieving this data: DataForSEO, Outscraper, Apify, SerpApi, Bright Data and others. They return full review text, all of it, with real pagination, and they are dramatically cheaper than Google. You have an account with one of them already.

But none of them solves the actual problem either. Not one vendor lets you search across businesses by review content. A couple let you filter reviews by keyword once you've named the business, which is a different and much smaller thing. So the shape of the work is fixed no matter who you buy from: choose your practices, pull all their reviews, then search the text yourself. The only lever you control is how many practices you pay to check.

Worth being straight about the legal position. These vendors obtain the data by scraping, which is how they can offer what Google won't. That's a different risk profile from Google's own terms, not an absence of one. It's widely done, it's how most local-SEO tooling works, and the practical exposure for research at this scale is low. But it isn't zero, and I'd rather you knew that before spending than after.

The number that changes everything

If "every US clinician" means every clinician, the pool is about 9.2 million registered providers and the project is absurd. It doesn't have to mean that.

Every US clinician is listed for free in the federal NPI registry, downloadable in bulk, updated monthly, with self-declared specialty codes. That lets you cut the pool before spending a cent:

Then the second cut, which is the one that really matters. Clinicians are not businesses. A practice with six OB/GYNs has one Google listing and one set of reviews. Individual doctors usually have no listing of their own at all. Expect somewhere between three and eight clinicians per listing, so 112,000 clinicians collapses to roughly 25,000 to 35,000 Google listings. You pay per listing, not per clinician.

What it actually costs

Using your own DataForSEO account's live rate card, for about 30,000 practice listings averaging 50 reviews each:

For comparison, doing the same thing through Google's official API would cost about $750, would cap you at 5 reviews per practice so it wouldn't answer the question anyway, and would breach the terms the moment you saved the text.

Your DataForSEO balance is currently $41.26, which is enough to prove the whole thing works on a few hundred practices before committing to the rest. The review pricing needs one live task to confirm the exact billing unit, which is the first thing a pilot would settle.

What running it would have looked like

Kept as a record rather than a plan, since the answer was no. Steps 2 and 4 are the ones worth carrying forward if review data ever comes up again on another platform.

The honest caveats

Matching is the weak joint. Expect 50% to 70% of practices to match their Google listing automatically on a clean pass. Registry addresses are often billing addresses, the legal name frequently differs from the name on the door, and telehealth practices may have no listing at all. That last one bites here specifically, because telehealth menopause practices are exactly the kind of prospect you want.

Reviews are patient language, not clinical fact. A patient writing "she finally listened to me about menopause" tells you something a services page never will, which is the real appeal of this idea. But a practice can be excellent at menopause care and have no review that names it, and a practice can collect one passing mention without offering anything. Treat a hit as a reason to look, the same as with the testosterone sweep.

You may already have the cheaper answer. You're holding full text for 2,983 clinic websites, gathered and searchable. If the goal is finding practices that treat menopause, their own website says so more reliably than their reviews do, and you've already paid for that. Reviews earn their cost for a narrower question: finding practices that treat menopause well but never say so in their marketing. That's a real and interesting gap. It just isn't the same question, and it's worth being deliberate about which one you're asking.