Paid CTR Fell From 19.7% to 6.3% on AI Overview Queries: Rebuilding Healthcare Google Ads

If your hospital's Google Ads account shows falling click-through rates on informational searches while impressions hold steady, the cause may not be your ads. It may be that an AI Overview now sits above them. This article is for performance marketers and CMOs who run or oversee Google Ads for hospitals and clinics, and who want a practical response rather than a headline.

It covers what the published data says, what Google says about ads in AI Overviews for healthcare, how to measure the effect in your own account, and how to restructure campaigns so that spend goes where a click is still likely to turn into an appointment. It assumes you know the basics of campaign structure; for those, see our complete guide to healthcare PPC and the piece on multi-department campaign architecture.

What the data says, and what it does not

Seer Interactive tracked click-through rates across a large set of queries and reported in its AI Overview CTR study that on queries where an AI Overview appears, paid CTR fell from 19.70% to 6.34%, and organic CTR from 1.76% to 0.61%. The same study found that brands cited inside the AI Overview earned more clicks than brands that were not: 35% more organic clicks and 91% more paid clicks.

Read those numbers with care. The data is not limited to healthcare, it measures a set of accounts and queries that may not look like yours, and Seer's own methodology notes matter. It is a strong signal about direction, not a forecast for a hospital in Indore.

Healthcare is exposed because the format is common there. BrightEdge reported that AI Overviews appeared on 88% of healthcare queries in December 2025, rising to 93% for educational health queries, while "near me" searches largely stayed in the local pack. That split tells you where the exposure is: broad, informational, symptom and condition searches.

The part many accounts miss: healthcare ads are not shown inside AI Overviews

Google's help page on ads and AI Overviews lists India among the countries where ads can appear within AI Overviews, drawn from existing Search, Shopping and Performance Max campaigns. It also says ads are not shown within AI Overviews for sensitive verticals, including healthcare.

The consequence is uncomfortable. On a query like "knee pain causes", an AI Overview may answer the question while your search ad is pushed down and cannot appear inside the Overview. Competitors in other industries can show ads within the answer. Healthcare advertisers get the cost of the format without the placement.

Confirm this against the page itself at the time you read this, because Google updates these policies often.

Step 1: Measure the exposure in your own account

Google Ads does not report a clean "AI Overview present" segment for your queries, so you need your own sample. Here is a workable method.

1. Export the search terms report for the last 90 days, with impressions, clicks, cost, conversions and conversion value by campaign.
2. Classify each high-spend term by intent: symptom or condition (informational), treatment or procedure (consideration), doctor or hospital plus city (selection), and brand.
3. Run a sample of the top 100 terms by spend on a phone and desktop, logged out, from your city. Record whether an AI Overview appears and whether your ad appears above or below it.
4. Compare click-through rate and cost per conversion by intent class and by AI Overview presence in your sample.
5. Repeat the sample every month, because presence changes.

The aim is not precision. You are looking for the classes of query where CTR has fallen and cost per conversion has risen, and the classes that have not moved. If symptom terms have lost clicks while selection and brand terms are steady, you have your answer.

Step 2: Rebuild around intent, not around keywords

Most hospital accounts are organised by department. A better first split is by intent, with departments inside.

Selection campaigns. Terms where the person has chosen a type of care and is picking a provider: "bariatric surgeon in Pune", "IVF clinic near me", "hospital name plus department". These are the queries least affected by AI summaries and closest to booking. Give them the most budget and the tightest geographic targeting.

Consideration campaigns. Treatment and procedure terms: "knee replacement cost", "hernia surgery recovery time". These may show Overviews, so test carefully and pair them with landing pages that answer the cost and process questions the Overview will not.

Informational campaigns. Symptom and condition terms. For most hospitals these are the weakest value for ad spend now. Cut them back unless you have proof they produce enquiries, and let organic content and the rest of your marketing handle them. This is a decision to test, not a rule: keep a small control budget so that you can see whether it matters.

Brand campaigns. Defend your hospital and doctor names. Brand searches are also where reputation queries land, which makes the ad copy and sitelinks worth reviewing.

Local Services Ads and Maps. Where eligible, these sit outside the AI Overview question and answer high-intent local searches. See our assessment of Google Local Services Ads for doctors.

Step 3: Change what you optimise for

With fewer clicks per impression on informational queries, click-based metrics mislead. Click-through rate falling is not by itself a failure if conversions per rupee hold.

• Judge campaigns on cost per enquiry and cost per booked appointment, not CTR or cost per click. If your account counts form fills only, add call tracking and WhatsApp click events so that call-led departments are not penalised. Our guide to Google Ads for healthcare covers conversion set-up.
• Import offline outcomes. Send "appointment booked" and "patient arrived" back to Google Ads as offline conversions where your consent and policies allow. Bidding on arrived patients prevents the account from favouring cheap, low-quality leads.
• Review bid strategies by intent class. Smart bidding learns from conversion volume, so low-volume selection campaigns may need portfolio bidding or a manual floor until they gather data.
• Watch impression share on selection terms. If you are losing impression share to budget or rank on high-intent terms while spending on informational ones, move money.

Keep health data out of audiences. Google's healthcare and medicines policy and its personalised advertising restrictions limit how sensitive health information can be used for targeting, so remarketing lists based on pages about specific conditions are a policy risk. Check what your account actually collects before you assume a list is permitted.

Step 4: Rebuild landing pages for the click you still get

The visitors who click through after reading an AI summary have already seen a general answer. They arrive with a narrower question. Give them what the summary does not.

• Doctor names, qualifications and consultation locations in text.
• Clear next steps: call, WhatsApp, book. State what happens after the enquiry.
• Cost ranges or what affects the bill, reviewed by finance.
• Clinical content reviewed by the specialist and dated.

Keep claims within NMC and advertising rules: no outcome promises, no comparative superiority claims, no patient testimonials in doctors' ads. Test one change at a time with a proper experiment rather than rewriting every page together.

Step 5: Win the citation, not only the click

Seer's finding that cited brands get more clicks changes the value of content. Being named in the Overview is now part of paid performance. That is an organic content job: clear, quotable pages from named clinicians, structured so that models can extract the answer. Our AI search optimisation guide covers the content structure, and the same pages can serve as the destination for consideration campaigns.

Track citations by running your query sample monthly and noting whether your hospital is cited. It will not appear in Google Ads reports, so keep a separate sheet.

A hypothetical example of the reallocation

Take a 200-bed hospital in Pune spending 6 lakh rupees a month across orthopaedics, cardiology and women's health, with every department running broad condition keywords. The search terms export shows that roughly a third of spend goes to symptom searches, and the sample check finds an AI Overview on most of them. The same terms produced few booked appointments, while a small set of "surgeon in Pune" and hospital-name terms produced most of them but ran out of budget by mid-afternoon.

A sensible first move is to shift the symptom budget into the selection terms that were limited by budget, keeping a small holdout on symptom terms for four weeks. After the test, the team compares cost per booked appointment for both groups. If the holdout shows that symptom campaigns still assist bookings, restore part of it. If not, the saving stays with selection and brand. The numbers here are invented to show the sequence; your own export decides the split.

How to report it upward

Hospital boards tend to see click-through rate on the first slide because the platform shows it first. Change the order. Lead with cost per booked appointment and booked appointments by department, then show the intent split, then the AI Overview sample as the explanation for any CTR decline. A short note that Google does not show healthcare ads inside AI Overviews helps the board understand why the account cannot simply "bid its way back".

What not to do

• Do not raise bids to recover CTR. Higher cost per click on a query with a lower click rate compounds the loss.
• Do not switch off informational campaigns blindly. Some produce assisted conversions. Test with a holdout.
• Do not trust the platform's recommendation tab to fix this. Auto-applied recommendations tend to widen match types and budgets.
• Do not hide the decline from the board. Present cost per booked appointment alongside CTR, with the AI Overview sample as context.

Where to start

1. Export 90 days of search terms and classify by intent.
2. Sample 100 top terms for AI Overview presence and your ad position.
3. Compare CTR and cost per conversion by class.
4. Move budget towards selection and brand campaigns, with a small holdout on informational terms.
5. Add offline conversion imports and call tracking.
6. Rewrite landing pages for the narrower question.
7. Review monthly and report cost per booked appointment.

Our Google Search Ads service page describes how we run this restructure for hospital accounts.