Noticed Your Psychology Google Ads Aren't Performing? Here's Why and How to Fix It

If your psychology practice's Google Ads have gone quiet, or each booking costs more than it used to, you're not imagining it, and it's rarely just one thing. In the account I'll walk through, the price of a click on the same keyword roughly tripled in six months, the reports were flattering the results, and a weak landing page was quietly adding to the cost of every click. The practice's calendar is still full. Here's what was behind it and how to fix it.

The short version: you can't outbid the national brands, so win by owning your own name organically, counting real bookings, and going specific instead of generic.

Why it's happening: the same keyword now costs three times as much

Take the broad keyword "psychologist". Across five reports its Quality Score sat at 3 out of 10, its click-through rate barely moved, and its price per click did this:

The last two windows overlap, so read them as one picture.

Searchers kept clicking at about the same rate, and Google's ratings for the ad and the page didn't change. If search behaviour had shifted, click-through rate would have moved with it. It didn't. The thing that moved was the price.

It's the same story on other terms. "Psychologist sydney" went from $2.84 to between $6.80 and $6.94, though the earlier figure comes from only 155 impressions. And "psychologist near me" climbed from $2.69 to $3.57 to $3.86 across three reports in about a month, while its click-through rate rose too, from 4.1% to 6.0% to 6.4%. People clicking more while the price climbs is the pattern you'd expect when more advertisers bid on the same searches.

What's pushing the price up: more advertisers want your clients

In July I noticed the competition in this space getting heavier, with large national and telehealth providers showing up in the same auctions, and I raised lifting the budget with the client. Mental health is a category where well-funded brands spend a lot on Google, and a single practice bidding on "psychologist" is in the same auction as all of them.

The account shows the signs. Rank-lost impression share, the share of auctions where the ads were beaten on ad rank, climbed from 13% to 29% to 32% across the last three reports. Budget-lost impression share ran between 63% and 83%, which means that at these prices the daily budget was running out before the day did.

Two things changed in the same stretch: more bidders, and the bidding strategy I moved to in order to compete with them. The standard reports can't cleanly separate the two, but Auction Insights, which lists who you're up against, shows the competition side. I walk through it below.

When a market gets more expensive you have three options: pay more per client, get cheaper clicks, or need fewer paid clicks. The sections below show who you're up against, what else can distort your results, and the fixes for the second and third options.

Who you're bidding against

Auction Insights is the report most practices never open. It lists every advertiser that appeared in the same auctions as you. I pulled it for two windows: 21 June to 20 July, the earliest data available, and 22 September to 21 October.

Each cell shows the June-July figure, then the September-October figure. Overlap rate is how often the competitor's ad showed alongside the practice's. The practice itself sat under 10% impression share in both windows.

The growth is concentrated in two online platforms. BetterHelp's impression share nearly doubled. In the later window it appeared in about a third of the auctions the practice entered, and when both ads showed, it ranked above the practice 66% of the time. UpDoc's overlap with the practice tripled. The directory barely moved. The rest of the list was an insurer, a couple of smaller platforms and several clinics, each under 10% impression share.

The practice's own ad reached the top of the page more often over the same period, 66.4% of impressions rising to 73.9%, while the very first slot held roughly steady at 25.1% then 22.0%. That is consistent with automated bidding paying for better positions against stronger bidders.

What does this explain? The cost per click jumped from about $1.57 to $4.15 in a single week in early June, before Auction Insights has any data, and that coincides with my move to automated bidding. The climb since, from about $4 to above $5, is consistent with BetterHelp and UpDoc taking a bigger share. So the big step looks like the bidding change, and the further climb looks like competition. Weekly cost per click on non-brand keywords would confirm it.

Your report may be flattering you

The best account upgrade I made was fixing what counts as a conversion.

In a 60-day window, Google reported 116 conversions on about $3,000 of spend, which works out at about $26 each. But 87 of those 116 were clicks on the "Book a session" link, and only 29 were confirmed bookings. The booking system (Halaxy) sits on a separate site, so the easiest thing for the ad account to record was the click that left the website, not the booking that followed.

This matters most in a crowded auction, for two reasons. Automated bidding optimises toward whatever you call a conversion, so if a button click counts, Google spends your money finding more button clickers rather than more clients. And when each click costs $5 instead of $2, every wasted click hurts more.

The conversion now fires only on a confirmed booking in the booking system. There's a smaller trap too: the conversion value in the account is a flat $175 placeholder, so the return on ad spend figure is the booking count times $175, not real revenue.

Your landing page makes every click dearer

Quality Score is Google's estimate of how well your keyword, ad and page match what the searcher wants, and it feeds into what you pay per click. In this account the impression-weighted score was 2.0 in three separate reports spread across four months, so nothing improved.

The breakdown points at one culprit. By mid-October all 7 scored keywords had a below-average landing page experience, up from 4 of 8 in the earlier export. Expected click-through rate was below average on 5 of 7. "Best psychologist sydney" and "treatment for anxiety" scored 1 out of 10 on all three components. The ad copy, meanwhile, earns average or better relevance almost everywhere, so the limit is the page, not the ads.

In an auction where better-funded advertisers push the price up, a weak page means you pay a premium on every click. It's also the one lever no competitor can take from you, and the one I'd work on next. If you want to see how we approach this, our psychologist Google Ads service page covers the setup.

The fix: rank for your own name, then stop paying for it

Here is where SEO and ads meet, and where this account got its biggest budget win.

The cheapest clicks in the account were the ones on the practice's own name. The exact-match keyword for the practice name had a 41.7% click-through rate at $1.36 a click. The practitioner's own name ran at 27% to 30% with a Quality Score of 9. The exact keyword for "psychologist" plus the suburb cost $2.96 a click. Compare that with $5.38 or more on the generic keyword. In the 60-day export, brand searches were about 5% of clicks and roughly half of the conversions Google counted.

They were also searches the practice already won for free. SEO had the site ranking first on Google for its own name and for its suburb, so every paid click on those terms was a click the practice would have received anyway. When I looked at who was clicking, about five clicks a week came from existing clients, and some of those cost as much as $7 because the same searches triggered the suburb keyword too.

So I paused brand and suburb bidding and pointed the budget at people who didn't know the practice yet. As competition pushes the price of a new-client click up, the value of every search you own organically goes up with it.

The fix: go specific, not generic

I went through the keyword data looking for any sign that people were searching differently. There wasn't one. There were about four to five unique search terms per 100 clicks in every period, and no new type of query was taking over. What changed was which keywords earned their keep:

• Broad "psychologist" stopped paying. It ran at $15 to $32 per conversion in earlier exports, though those figures include link clicks. In each of the last two reports it spent about $1,000 for 2 conversions. I paused that test keyword on 8 October.
• Niche terms held up. "Grief therapist" (exact match) gets only 15 to 38 impressions a month, but it converted in two of the periods at $3.87 a click, well under the account's $5.30 average. The counts are tiny, so treat it as a pattern to test, not a result.
• Broad match reads intent loosely. Among the search terms were queries that looked like someone hunting for a book or venting, not someone ready to book a psychologist. That is the cost of letting Google guess.

In a crowded auction the head term "psychologist" is where the big spenders hold the high ground. The searches that earned their keep were specific: a particular need, a particular person, a particular place.

Proof it works: what the honest numbers look like

With conversions tracked on confirmed bookings and the budget pointed at new clients, this is where the account stands. The practice's calendar is full, and the practitioner has thanked me for keeping it that way.

In the last three 30-day reports, the cost per confirmed booking was $169, $343 and $224. Those windows overlap by most of the month, so they are one picture, not three. In the latest 30 days the ads produced 6 booked appointments, all new clients as Google counted them. In the couple of months before that, the account produced 10 to 15 a month.

A full calendar changes what a good result is. A practice with a fixed number of appointment slots doesn't need the most bookings. It needs the right ones, at a price that makes sense for what a new client is worth. And with five to ten bookings a month, one booking more or less moves the cost per booking by tens of dollars, so judge the trend over months, not a single report.

What psychologists can't use to compensate

Most Google Ads advice leans on two tools: social proof and retargeting. A psychology practice can't use either, and that raises the stakes on everything else.

Testimonials. Under the Health Practitioner Regulation National Law, advertising a regulated health service can't use testimonials or purported testimonials, and that includes a practice's own website and social media pages. Not every review counts as a testimonial, and advertisers aren't responsible for reviews on platforms they don't control, but quoting client praise on your landing page is out. Ahpra has a testimonial tool that walks through the definition.

Remarketing. This restriction comes from Google, not from Ahpra. Google's personalised advertising policy treats mental health conditions and counselling for issues like anxiety and depression as a sensitive category, which rules out following website visitors around with remarketing audiences.

What's left is a single visit. Someone searches, sees your ad, lands on your page and decides then, with no reminder later. The search term, the ad, the landing page and the booking flow have to carry the whole decision.

Your fix checklist

1. Count confirmed bookings, not clicks. Make sure the conversion fires on the booking itself.
2. Rank organically for your own name and suburb, then stop paying for them. Check first that you rank first and nobody bids on your name.
3. Don't fight on "psychologist". Go specific: a need, a name, a place.
4. Fix the landing page. A below-average page raises the cost of every click, and no competitor can take that lever from you.
5. Check Auction Insights each quarter, and talk to your practitioner about budget before the market moves, not after.
6. Judge results against what a client is worth and how many slots you have free. A full calendar is the finish line.

If you'd like a second pair of eyes on what your own account is counting, get in touch with firstSites or book a call.