A Google Ads audit checklist for doctors in Singapore

Friday 6:15pm in Tampines. A parent searches “paediatrician open now near me” on mobile, taps the top ad, lands on a clinic homepage that still leads with corporate awards from 2019, and the only CTA is a contact form that asks for NRIC before a reason for visit. The clinic’s Google Ads dashboard will count a click. The diary will not count a patient.

That gap is what a proper audit of PPC for doctors Singapore is built to catch. Not vanity CPL. Not “we got more impressions this month.” Booked slots, attended consults, and cost per filled appointment after GST and no-shows. Private GP, specialist, family clinic, aesthetic doctor with medical consults in the mix: paid search still pulls the highest intent in this market when someone is ready to call, WhatsApp, or walk in. It also burns budget fastest when negatives, dayparting, landing proof, and response capacity are left on autopilot.

This piece is a self-audit you can run on a live account. Score each area, find the leaks, rank the fixes, then run a 30-day sprint. Currency is SGD. Patients often prefer WhatsApp or a direct call over a long form. Your media plan has to respect that, or Smart Bidding optimises for the wrong event.

Where Singapore clinic search spend quietly fails the diary

High-intent queries in Singapore medical search are blunt: “ENT specialist Orchard”, “GP fever child Sunday”, “dermatologist eczema adult Central”, “women’s clinic Pap smear”. They are not research journeys. They are booking journeys. When campaigns treat them like brand awareness, you pay hospital-grade CPCs for tyre-kickers, wrong specialties, and people hunting polyclinic or CHAS pathways you cannot serve.

The failure mode we see most often is simple. The account is optimised for form fills or “calls from ads” as a single blended conversion. Front desk is understaffed after 1pm. Ads keep running through lunch and late evening when no one answers. Landing pages talk about “comprehensive care” instead of the exact service in the ad. Negatives were uploaded once at launch and never reviewed against search terms. Twenty to thirty percent of SEM budgets routinely vanish into irrelevant queries, thin negatives, and mismatched pages. Medical accounts are not immune; they are often worse because keywords look “clinical” even when intent is wrong.

Mobile drives most paid search clicks. Clinic pages that pass desktop eyeball tests still fail Interaction to Next Paint on a jammed 4G commute from Woodlands. Run your top two landing URLs through PageSpeed Insights on mobile - most Singapore clinic pages we see score under 60 on mobile due to late-loading chat widgets and uncompressed hero images. If your click-to-call button is below the fold and the WhatsApp widget loads late, you are buying traffic Google cannot convert even when the bid is perfect.

Another Singapore-specific trap: bidding on symptom clusters without service and location guardrails. “Cough” and “sore throat” pull volume. They also pull people who want same-day polyclinic queues, traditional Chinese medicine, or home remedies. Your unit economics need named services and clear private-fee framing, not pure symptom harvest, unless the clinic truly runs a walk-in urgent model with capacity to match.

If you already suspect the account is busy but the rooms are not, a structured SEM / Google Ads review beats another month of “raise budget and hope.”

Audit scorecard: score your doctors Google Ads account

Work through each item with the live account open. Score 0 if missing, 1 if partial, 2 if solid. Total out of 24. Under 12 means stop scaling spend. 12-18 means fix before you increase bids. 19+ means tighten and expand carefully.

Audit scorecard

• 1Intent map vs diary reality List the top 20 converting search themes from the last 90 days and map each to a real service you book this week: general consult, child consult, chronic review, named specialty procedure, health screening package. If half the spend sits on themes the diary cannot productise (vague “doctor near me” with no clinic radius, competitor hospital brands you will never win, insurance schemes you do not accept), mark this low. PPC for doctors Singapore only works when query language matches appointment types the front desk can confirm in one call.

List the top 20 converting search themes from the last 90 days and map each to a real service you book this week: general consult, child consult, chronic review, named specialty procedure, health screening package. If half the spend sits on themes the diary cannot productise (vague “doctor near me” with no clinic radius, competitor hospital brands you will never win, insurance schemes you do not accept), mark this low. PPC for doctors Singapore only works when query language matches appointment types the front desk can confirm in one call.

• 2Campaign structure and match types Separate brand, high-intent service, and (if you must) broader symptom or condition themes. Do not dump everything into one Performance Max or one broad Search campaign and call it modern. PMax can carry efficient volume later; it should not be the only place medical compliance, location, and negatives live. Check that exact and phrase coverage still exists on money terms even if you use broad with strong signals. Google still needs keywords and negatives in healthcare, despite the industry shift toward signals.

Separate brand, high-intent service, and (if you must) broader symptom or condition themes. Do not dump everything into one Performance Max or one broad Search campaign and call it modern. PMax can carry efficient volume later; it should not be the only place medical compliance, location, and negatives live. Check that exact and phrase coverage still exists on money terms even if you use broad with strong signals. Google still needs keywords and negatives in healthcare, despite the industry shift toward signals.

• 3Negative keyword hygiene Pull search terms for 30 and 90 days. Build shared negative lists for jobs, salary, PDF, login, free, polyclinic-only intent you cannot serve, student notes if you do not issue them, wrong specialties, and location terms outside your catchment if you refuse those patients. Negatives are not a launch chore. Clinics that treat them as monthly ops usually reclaim wasted spend faster than clinics that only tweak bids. If you cannot show last review date, score zero.

Pull search terms for 30 and 90 days. Build shared negative lists for jobs, salary, PDF, login, free, polyclinic-only intent you cannot serve, student notes if you do not issue them, wrong specialties, and location terms outside your catchment if you refuse those patients. Negatives are not a launch chore. Clinics that treat them as monthly ops usually reclaim wasted spend faster than clinics that only tweak bids. If you cannot show last review date, score zero.

• 4Call vs form vs WhatsApp as primary conversion Decide the primary action by service. Same-day fever and acute ENT usually need click-to-call or WhatsApp with staff coverage. Elective screening and specialist second opinions can use short forms. If Smart Bidding optimises for a long form while 70% of quality patients insist on calling, your CPA looks healthy and your rooms stay empty. Track calls with duration thresholds that match real booking conversations, not three-second misdials.

Decide the primary action by service. Same-day fever and acute ENT usually need click-to-call or WhatsApp with staff coverage. Elective screening and specialist second opinions can use short forms. If Smart Bidding optimises for a long form while 70% of quality patients insist on calling, your CPA looks healthy and your rooms stay empty. Track calls with duration thresholds that match real booking conversations, not three-second misdials.

• 5Dayparting against response capacity Overlay hourly conversion and call data with front-desk rosters. If ads run full blast through lunch, public holidays, and after last slot when no clinician can take a walk-in, you are buying abandoned intent. Singapore patients expect fast WhatsApp replies. A 40-minute lag on a high-CPC click is a lost booking to the clinic that answers in four minutes. Score high only if bid adjustments or schedules mirror real capacity in Central, East, West, and satellite clinics separately when staffing differs.

Overlay hourly conversion and call data with front-desk rosters. If ads run full blast through lunch, public holidays, and after last slot when no clinician can take a walk-in, you are buying abandoned intent. Singapore patients expect fast WhatsApp replies. A 40-minute lag on a high-CPC click is a lost booking to the clinic that answers in four minutes. Score high only if bid adjustments or schedules mirror real capacity in Central, East, West, and satellite clinics separately when staffing differs.

• 6Location and catchment control Presence targeting, radius, and location exclusion must match where you actually want patients from. National campaigns for a single clinic in the East will attract North and West traffic that never travels. Presence vs presence-or-interest settings still get mis-set. Verify Singapore geo, exclude countries that leak through, and split multi-outlet accounts so Punggol and Orchard are not fighting the same budget pool without intent.

Presence targeting, radius, and location exclusion must match where you actually want patients from. National campaigns for a single clinic in the East will attract North and West traffic that never travels. Presence vs presence-or-interest settings still get mis-set. Verify Singapore geo, exclude countries that leak through, and split multi-outlet accounts so Punggol and Orchard are not fighting the same budget pool without intent.

• 7Ad copy and asset compliance fit RSA headlines should name the service, the clinic type, and a concrete next step (book GP today, call for same-day slot, WhatsApp for appointment). Avoid unprovable superlatives and drug claims that invite policy flags. Sitelinks should deep-link to service pages, fees overview if you publish ranges, and directions. Callouts for languages spoken, parking, nearest MRT, and opening hours reduce bounce more than clever brand lines.

RSA headlines should name the service, the clinic type, and a concrete next step (book GP today, call for same-day slot, WhatsApp for appointment). Avoid unprovable superlatives and drug claims that invite policy flags. Sitelinks should deep-link to service pages, fees overview if you publish ranges, and directions. Callouts for languages spoken, parking, nearest MRT, and opening hours reduce bounce more than clever brand lines.

• 8Landing page message match and proof Every money ad group needs a dedicated page or tight service URL, not the homepage carousel. Above the fold: service name, who it is for, what happens next, primary CTA (call / WhatsApp / book), and trust proof that Singapore patients actually use (MOH-relevant credentials presentation, doctor names and languages, reviews that mention wait time and clarity, clinic photos that look like the real fit-out). Fee transparency where you can publish package bands in SGD beats vague “contact us for pricing” on elective services.

Every money ad group needs a dedicated page or tight service URL, not the homepage carousel. Above the fold: service name, who it is for, what happens next, primary CTA (call / WhatsApp / book), and trust proof that Singapore patients actually use (MOH-relevant credentials presentation, doctor names and languages, reviews that mention wait time and clarity, clinic photos that look like the real fit-out). Fee transparency where you can publish package bands in SGD beats vague “contact us for pricing” on elective services.

• 9Page speed and mobile form friction Run field data and lab checks. Heavy chat widgets, autoplay video, and multi-step forms with unnecessary fields kill mobile conversion. Keep required fields minimal for first contact. Offer PayNow or card only at payment stage, not as a barrier to enquiry. If INP feels sticky when tapping call, fix that before you raise CPC caps.

Run field data and lab checks. Heavy chat widgets, autoplay video, and multi-step forms with unnecessary fields kill mobile conversion. Keep required fields minimal for first contact. Offer PayNow or card only at payment stage, not as a barrier to enquiry. If INP feels sticky when tapping call, fix that before you raise CPC caps.

• 10Tracking integrity and enhanced conversions GTM or equivalent firing cleanly for calls, WhatsApp clicks, thank-you pages, and booked events that sales or clinic software can verify. Enhanced Conversions and first-party signals matter when ad blockers and privacy changes strip 20%+ of raw conversion data. If offline import from the practice management system is missing, Smart Bidding is guessing. Last-click vanity will double-count across channels; at minimum, reconcile Google Ads conversions to attended appointments weekly.

GTM or equivalent firing cleanly for calls, WhatsApp clicks, thank-you pages, and booked events that sales or clinic software can verify. Enhanced Conversions and first-party signals matter when ad blockers and privacy changes strip 20%+ of raw conversion data. If offline import from the practice management system is missing, Smart Bidding is guessing. Last-click vanity will double-count across channels; at minimum, reconcile Google Ads conversions to attended appointments weekly.

• 11Unit economics: cost per booked job Define booked job as confirmed appointment that shows or is prepaid per your policy. Include GST reality in fee math. Compare cost per booked job by campaign, not blended CPL. A S$18 form lead that never answers is more expensive than a S$45 qualified call that books a S$180 consult. If finance only sees media CPL, your audit is incomplete.

Define booked job as confirmed appointment that shows or is prepaid per your policy. Include GST reality in fee math. Compare cost per booked job by campaign, not blended CPL. A S$18 form lead that never answers is more expensive than a S$45 qualified call that books a S$180 consult. If finance only sees media CPL, your audit is incomplete.

• 12Creative and offer refresh cadence Even in Search, assets fatigue. Rotate proof, seasonal capacity (haze, flu peaks, school holiday clinics), and offer framing (new patient slots, health screening windows) without constant structural rebuilds that reset learning. Score low if RSAs have not been reviewed in 90 days or if every headline still says “quality healthcare” with no service noun.

Even in Search, assets fatigue. Rotate proof, seasonal capacity (haze, flu peaks, school holiday clinics), and offer framing (new patient slots, health screening windows) without constant structural rebuilds that reset learning. Score low if RSAs have not been reviewed in 90 days or if every headline still says “quality healthcare” with no service noun.

After you score, pick the three lowest items that sit on the path from click to booked slot. Those become the only priorities for the next two weeks. Spreading effort across all twelve is how audits die in slide decks.

Playbook fixes for high-intent medical search in Singapore

Once the scorecard shows the leaks, fix in this order: measurement, query control, schedule and response, then landing proof. Bidding last. Raising bids on a broken path just buys more of the same waste.

DIY playbook

• 1Export 90 days of search terms. Tag each row bookable, wrong specialty, wrong channel (jobs, research, free), wrong location, or competitor brand. Build negatives from the non-bookable set before you touch bids.

• 2Split or label conversions: qualified call (e.g. 90+ seconds), WhatsApp click with UTM, form submit, and offline attended appointment if you can import. Make the primary conversion the one that correlates with revenue.

• 3Build one landing page per top service cluster (e.g. child consult, women’s health screening, named specialty). Match H1 to the ad’s service noun. Put call and WhatsApp first; form second.

• 4Align ad schedules to rostered phones. Cut or reduce bids in hours with zero pickup. Add staff WhatsApp coverage notes for peak East and Central evening demand if that is your pattern.

• 5Rewrite RSAs with service + location + action. Pin sparingly. Test one proof angle at a time (languages, same-day, parking/MRT), not five at once.

• 6Tighten geo to realistic catchments. For multi-clinic groups, break budget by outlet performance, not vanity national reach.

• 7Only then move bidding: if conversion volume is thin, stay on a controlled strategy until 30+ primary conversions per month stabilize signals. Do not assume automation alone fixes quality.

• 8Weekly 45-minute ops: search terms, disapproved assets, landing speed, cost per attended booking, and one creative or offer change. Avoid daily structural rebuilds that reset learning.

DIY tools for this audit

• UTM link builder

• Core Web Vitals checker

• SERP snippet preview

• Meta description checker

• H1 checker

• SEO competitor analysis

On post-click behaviour, stop guessing why forms die. Session recordings and heatmaps show whether patients bounce at fee ambiguity, scroll past the doctor bio, or abandon when the form asks for too much up front. A short pass with HeyLead Insights on your top two service pages usually surfaces one obvious friction point you can fix in a day.

Clinics that want a partner-built view of the full path from query to booked consult can start from dedicated Doctors marketing rather than a generic lead-gen template.

Two Singapore clinic scenarios: busy phones versus filled rooms

East-side family clinic, single outlet. Spend sat around S$3,500-S$4,000 per month on broad “doctor near me” and symptom terms. CPL looked acceptable. Attended rate did not. The mechanism was schedule: ads ran until 10pm while phones stopped at 1pm Saturday and 5pm weekdays. Search terms were full of “polyclinic” and “CHAS” variants with no negatives. Fix path: negatives first, ad schedule clipped to rostered hours plus a two-hour buffer, primary conversion switched to calls over 90 seconds, landing H1 changed from “Family healthcare” to “Private GP consults in Tampines - call or WhatsApp to book.” Within a few weeks the team stopped celebrating form volume and started tracking cost per attended new-patient slot. The odd detail that unlocked it was not a new bid strategy. It was a printed roster taped next to the Ads schedule screen.

Central specialist group, three doctors. Strong brand search, weak service search. Performance Max was “driving the bulk” of spend with soft lead definitions. Marketing saw efficient CPA in-platform. Operations saw tyre-kickers asking for procedures the group does not offer. Mechanism: asset groups mixed unrelated services and sent traffic to a generic consultants page. Fix path: carve high-intent Search campaigns per specialty with exact service URLs, move PMax to remarketing-heavy support only after clean offline conversion import from the booking system, and add sitelinks to fees FAQ and languages. One landing test moved the primary CTA above biographies; enquiry-to-book rate lifted because patients stopped hunting for how to act. Quantify carefully: they cared about a single metric shift, cost per completed first consult, not five dashboards.

If your story matches either pattern, resist the urge to “just add Meta” until Search hygiene and response speed are honest. Cross-channel expansion on a leaky booking path multiplies noise.

Patterns we see repeatedly

One Tampines GP group cut after-hours spend by 60% over six weeks and saw attended bookings rise 18% in the same period - the phones finally matched the clicks, and booked paid patients beat the month they had maximised impressions.

At multi-doctor specialist clinics, CPL often looks fine until Ads is tied to attended appointments. Roughly half the cheap leads turn out to be wrong specialty; shared negatives and dedicated service pages routinely fix more than any bid change.

FAQs

What should a healthy cost per booked appointment look like for PPC for doctors Singapore?

There is no universal number. Anchor to average revenue per attended new consult or package, contribution after GST, and no-show rate. If a first visit yields roughly S$120-S$250 net contribution depending on specialty mix, a cost per attended booking that stays a sensible fraction of that contribution can scale; a cheap lead that never attends cannot. Build the model from your fee card, not a national average CPC slide.

Should we use Performance Max as the main campaign for a medical clinic?

Use it as a support layer once Search captures high-intent service and brand terms with clean conversions. PMax without tight assets, location control, and offline booking feedback tends to chase volume. Many accounts still need classic Search structure for medical query control and negatives.

Call extensions or WhatsApp: which should we prioritise?

Prioritise whatever your front desk answers fastest during paid hours. Many Singapore patients open WhatsApp by default. Track both, but pick one primary for bidding so algorithms are not torn between conflicting goals. Staff the channel you choose.

How often should we audit negatives and search terms?

Weekly on active medical accounts, with a deeper 90-day cleanse monthly. Flu seasons and haze periods change query mix quickly. A one-off negative list at launch goes stale.

Can we guarantee a CPL before seeing the account?

No honest operator should. Outcomes depend on landing proof, capacity, fees, specialty, and conversion setup. Expect meaningful signal over weeks to a few months of disciplined iteration, not overnight ROAS promises. Benchmarks vary too widely by specialty and fee structure to cite meaningfully here. Build your threshold from your own fee card: if a first consult nets S$150 contribution after GST and no-shows, your maximum tolerable cost per attended booking is a fraction of that - not a number from a cross-vertical average.

Putting it to work

This week

• Score all 12 audit items with the live Google Ads and booking data open.

• Export 30 days of search terms and add a first negative pass for non-bookable intent.

• Verify primary conversions: call length, WhatsApp, form, and whether offline attended imports exist.

• Clip ad schedules to real phone and WhatsApp coverage; document the roster beside the account.

• Run Core Web Vitals and H1 checks on the top two service URLs that receive paid traffic.

• Build or fix UTM conventions so clinic software and Ads tell the same story.

Next 30 days

• Launch or harden one dedicated landing page per top service cluster with call and WhatsApp first.

• Reconcile weekly cost per attended booking by campaign; pause themes that never book.

• Split brand vs service structure if everything still lives in one bucket.

• Review RSA assets and sitelinks against real patient objections (fees, languages, directions).

• Only then test controlled budget increases on campaigns that clear unit-economic thresholds.

Pull your last 60 days of paid clicks beside attended new appointments by hour and by query theme, then fix the worst three mismatches before you raise any bids. When you want that loop owned end to end - query control, schedules tied to capacity, landing proof, and cost per booked consult rather than vanity CPL - HeyLead can run the ongoing Google Ads programme for Singapore clinics so your team stays on clinical operations; reach Martin on WhatsApp (415) 420-4059.