Google Ads for doctors: 862,000 searches a month, and they want a superhero.

PPC for doctors starts with the most contaminated word in search.

“Doctor” belongs to Marvel, the BBC and a Halloween costume before it belongs to your practice. That is not a figure of speech. Add up the film and television searches containing the word and you pass 860,000 a month in the United States alone.

A practice running broad keywords pays to reach a share of them. The clicks look ordinary in the report. They are simply people who wanted to know when the next season airs.

AdPilot: automated Google Ads for a medical practice

AdPilot links to the account and reviews it the way a consultant would if you had hired one. The searches that were never appointments. The evenings when the ads run and reception has gone home. The ads that earn clicks and no patients. Each finding comes back as a card in ordinary English. You approve it or you skip it. Nothing reaches your account until you do.

None of it happens behind your back:

• You approve every change. The AI proposes; you decide. Every budget, bid, keyword, or ad edit arrives as a plain-language card you approve or skip with one tap. Nothing is ever applied on its own.
• A snapshot before every change. AdPilot saves exactly how things looked before anything is touched, so you always have a clean “before.”
• One-click revert. Don’t like a change once it’s live? Undo it in a single click, straight back to that snapshot.
• A full audit log. Who, what, when, the old value and the new value, every change recorded. Nothing hidden.
• You keep the keys. You connect with one secure Google sign-in, and it only ever makes changes you approve. The account stays yours start to finish - you never transfer ownership, and you can disconnect anytime.

Where medical practice ad budgets actually leak

Real US search volumes from July 2026. Every line contains the word “doctor”, and every one of them is entertainment.

You can verify any row in about a minute. That is the point. Advice about medical PPC tends to stay vague, and vague advice is hard to argue with.

US monthly search volume, Ahrefs, July 2026. Around 862,000 searches a month, and not a patient among them.

Negative keywords for medical practices. Free, no email, no form.

Use this whether or not you ever become a customer. Ten minutes, no cost, and in a category this contaminated it is the best-value ten minutes available to you. For the rest - the ones only your own search terms report knows about - the negative keyword guide has the method.

Film and television

doctor who, the good doctor, doctor strange, doctor sleep, doctor doom, doctor odyssey, plague doctor, house md, greys anatomy, cast of, full episode, season finale, trailer for, watch on netflix, marvel, imdb

Jobs and training

salary, physician jobs, medical jobs, nursing jobs, job openings, now hiring, career, medical school, residency, mcat, how to become, nurse practitioner program, cv, locum

Self diagnosis and home treatment

symptoms of, signs of, home remedy, treat at home, cure at home, natural cure, what causes, is it normal, how long does, webmd, reddit

Free and low cost care

free clinic, low income, sliding scale, charity care, uninsured free, community health center

Definitions and comparisons

meaning, definition, md vs do, difference between, wikipedia

Costumes and merchandise

costume, mask, halloween, cosplay, funko, poster, t shirt

PPC for medical practices: five things that matter more than your bids

• Know what you may not do with health data. Google restricts personalized advertising built on health information. Remarketing lists assembled from people who viewed condition or treatment pages fall foul of it. Plenty of practices run those lists without realising, and it is the fastest route to a suspended account.
• Count booked appointments where you can. Calls are the default conversion and they overstate performance, because they include prescription requests, existing patients and billing questions. If your booking system can report a completed booking, that is the number bidding should chase.
• Advertise the condition, not the practice. Patients search for the problem, not for a clinic name. Ad groups built around conditions you actually treat consistently outperform a single campaign named after the practice.
• Set the radius by how far patients really travel. For routine care that is a short drive. For specialist and elective work it can be an hour or more. Targeting a whole state for primary care spends most of the budget on people who will never attend.
• Say nothing you cannot support. Outcome claims and guarantees are both a policy problem and a regulatory one. The safest high-performing ads describe access instead: availability, waiting time, location, whether you take their insurance.

What one new patient is worth to the practice

Four numbers you already know, and the one most practice accounts are run without. The click price is measured: “doctor near me” is 30,000 searches a month at about $2.50 and “primary care doctor near me” 17,000 at the same price, from the same pull as the table above. The patient value and the booking rate are a starting point, not a measurement of your practice. Nothing is stored and nothing is sent anywhere.

The build, step by step, in the order it has to happen

This is the order AdPilot's agent works in. We're printing it because a practice manager who follows it ends up with an account that works, and because you should be able to check a method instead of taking somebody's word for the result.

Set aside an afternoon for the build. Then half an hour every other week, for as long as the ads run. Everybody does the build. The every-other-week part is what decides whether the build was worth doing.

One thing first, and nobody selling marketing will say it to you: if your next new-patient appointment is six weeks out, don't do any of this yet. Advertising into a full panel buys angry voicemails and reviews about the wait. Fix the capacity, or point the money at the one service line that still has room.

Not a doctor? The same build, written for every other industry

1. 1

   Sort the 862,000 into fans and patients

   Add up the monthly US searches containing the word doctor that belong to somebody's entertainment and you land near 862,000. Marvel is 279,000 of it. The Good Doctor is 201,000. Doctor Who 150,000, Doctor Sleep 143,000 once the cast searches are counted alongside the film itself, and another 89,000 across the remaining shows plus a Halloween costume. Not one of those people is ill. They're on the sofa at nine at night wondering when the next season drops, their clicks are cheaper than a patient's, and cheap is exactly why they'll empty the budget before a real booking gets a look in.

   The buying side is short. “Doctor near me” runs 30,000 searches a month at about $2.50 a click. “Primary care doctor near me” is 17,000 at roughly the same price. So 47,000 people a month who might actually walk in, against 862,000 who want a cast list. Eighteen to one, and that's counting only the two buyer terms anybody here has measured. The rule writes itself: never bid the bare word. “Doctor” on its own is a Marvel keyword with a stethoscope drawn on it.

   Every candidate gets the same cold read before it goes near the account. Nobody is standing next to you explaining what your practice does, so ask whether the term could match a search for something you don't sell. Bare “clinic” is a veterinary clinic, a fertility clinic and a free legal clinic. Bare “physician” is a job board. “Sleep doctor” is a specialty you may not even offer. A term earns its place when it carries a qualifier only your patient would type: a city, a specialty, “accepting new patients”, “that takes” plus an insurer, or the name of something you genuinely treat.

   Then a middle bucket this trade has and a plumber doesn't. Somebody typing “why does my shoulder hurt at night” is neither a browser nor a buyer. They're at the stage before either one, they're expensive to educate on a small budget, and you can't follow them around afterward either - that part is policy rather than preference, and step six sets it out. Condition-only terms are a month-three problem. Launch on the people who have already decided to see a doctor and are only picking which one.

   • Ask it out loud on every term: has this person already decided to see a doctor, or are they still deciding whether they are ill?
   • Patient words: near me, your city, accepting new patients, same day, appointment, takes my insurance, self pay, your specialty.
   • Not-a-patient words: cast of, season finale, full episode, showtimes, salary, residency, mcat, symptoms of, home remedy, icd 10.
   • Never the bare word. “Doctor”, “clinic”, “physician” and “medical” each carry markets that will never book you.
   • Volume is the bill. Intent is the revenue. They are not the same column.

   Eleven rows from the table above: two to bid on, nine to block - single searches, not the grouped totals

   • doctor near me - 30,000/mo, ~$2.50 - PATIENT. The whole market in one phrase.
   • primary care doctor near me - 17,000/mo, ~$2.50 - PATIENT, and the term your first ad group gets built on.
   • doctor who - 150,000/mo - a BBC series. Block it, and read step five before you pick the match type, because this one bites back.
   • the good doctor - 129,000/mo - a hospital drama, not a hospital.
   • doctor sleep - 121,000/mo - a horror film, and a trap if you treat sleep apnea. Step five again.
   • doctor strange - 113,000/mo - Marvel.
   • doctor strange in the multiverse of madness - 87,000/mo - Marvel, at length.
   • plague doctor - 61,000/mo - a beaked leather mask, and the searches climb through the third week of October.
   • doctor doom - 56,000/mo - a comic-book villain.
   • the good doctor cast - 35,000/mo - they want the cast list, not a physical.
   • doctor odyssey - 28,000/mo - another series. Block it.
2. 2

   What the landing page has to say before they scroll

   Every ad points at exactly one page, and that page is where the money is won or lost. The homepage is almost always the wrong one. A practice homepage opens with a mission statement and a photograph of the building. Somebody with a fever and a new insurance card wants two facts in the first screen: are you taking new patients, and do you take their plan. Make them hunt for either and they're back on the results page in eight seconds, and you paid $2.50 for that.

   Open the page on a phone, standing up, in bad light. Is “accepting new patients” written on it in those words, or only implied by the fact that you exist? Are the plans listed as readable text, or as a PDF somebody uploaded in 2023? That PDF is one of the most expensive files in private practice. It opens badly on a phone, it doesn't show up when a patient searches your own site, and somebody who can't find their plan assumes the answer is no.

   Then the way in. If you have online scheduling, it belongs above the fold with the next open slot showing. If the only route is the telephone, the number has to be a tap, and the page has to say what happens when they call: answered by a person until five, voicemail returned the same day, whatever is actually true. Vague hours read as nobody is calling you back, because that's what people have been trained to expect from a medical office.

   While you're in there, write two lists. Things you do that you want more of: the nurse practitioner's open slots, the self-pay physicals, the Thursday telehealth block. And things you don't do at all: no obstetrics, no controlled substances for ADHD, no DOT physicals, no patients under two. That second list becomes your negative keywords in step five and nobody can write it for you. One more pass before you leave the site: Google reads the page as well as the ad, so the “guaranteed results” line a web designer left there in 2019 can stop an ad you wrote this morning, and the notice will point at the ad rather than the page.

   • “Accepting new patients” in the first heading, in those words, with the specialty and the city.
   • The insurance plans as readable text on the page. Not a PDF, and not “most major plans”.
   • A tap-to-call number and, if you have it, online booking. Both in the first screenful on a phone.
   • The next open appointment, or an honest range. “Usually within a week” beats silence.
   • A cash price for anything you sell self-pay, or one plain line on what a visit costs without insurance.
   • Search your own site for “guaranteed”, “cure” and “reverse” before you spend a dollar. Old copy stops new ads.
   • The written list of what you do not do. It is step five's raw material.
3. 3

   Capacity first, then the budget, then the map

   Take the practice this page uses as its example: Willowbrook Family Medicine in Tucson, two physicians and a nurse practitioner who opened in March with a half-empty schedule. Invented, but every figure below is real arithmetic. Before a single keyword, answer the question no marketing page asks. How many new patients can you seat next month? Not how many you want. How many slots exist. A practice at capacity buying more demand is paying $2.50 a click to generate calls it will turn away, and turned-away people write things down.

   Open a second campaign when you need a second budget or a second goal, and for no other reason. Keywords belong in the same ad group when one honest ad covers all of them; if the ad has to hedge, that's two groups. The playbook further up this page tells you to build ad groups around conditions rather than around the practice name, and that's right. It isn't a license to open one per condition. A family practice treats hundreds of things. Build the two or three condition themes that carry real volume for you, put the rest in the general new-patient group, and let the search terms report tell you when a fourth theme has earned its own box. Forty starved groups seeing two bookings each teach Google's bidding nothing at all.

   Now the sums, out loud. At $2.50 a click a new campaign needs roughly ten clicks a day before a month of data means anything, so $25 a day is the first floor. The second floor is the one people skip: a daily budget of at least three times what one booking costs you, so the campaign sees enough conversions to learn from. Assume one in ten clicks produces a call and two in three callers book. Those are assumptions, not measurements, and yours won't match mine. On those numbers a call costs $25 and a new patient costs about $38, so the second floor is $114 a day.

   Look at what $114 a day actually buys: about 45 clicks, four or five calls, and roughly 90 new patients a month. Three providers can't seat 90 new patients a month. The textbook floor has just handed you a budget your schedule can't cash, which is how practices end up with a ringing phone and a front desk that hates the ads. The honest resolution is to run at the click floor instead, somewhere between $25 and $40 a day, start on Maximize Clicks with a maximum cost-per-click cap near $3 so one auction can't run away with the day, and move to conversion bidding only after the booking conversion has fired thirty times. Two numbers get quoted at you here and they are not the same thing. The learning status on a bidding strategy usually clears in about a week, sometimes two, and while it shows you leave the settings alone. The other one is roughly 30 conversions in 30 days, and that is about whether a target is worth typing at all rather than whether learning has finished. At 24 bookings a month you are under that second threshold, so hold off on a target and let the account run on clicks for a while. Plan for six.

   Then the setting that wastes more primary-care money than any keyword. Draw the radius by drive time, not by county line: fifteen or twenty minutes for a sick visit, wider for anything a patient can't get down the street. Specialist and elective work is the opposite case, because people will drive an hour for it, and one radius set for both is how a practice funds two campaigns and reads neither. Underneath the radius is the location option nobody opens. Google's default reaches people merely interested in Tucson, which for a family practice means paying to be read by somebody in Ohio planning a move. Switch it to presence.

   Match types: launch on phrase and nothing else. A search that produced a booking gets promoted to exact, its own ad and its own page, and that promotion is the reward - you can't award it in advance. Broad stays off. It behaves only when conversion bidding, tracking that counts a real booking and a maintained negative list are all true at once, and in a category where the main noun is mostly Marvel, broad without all three will spend the month proving it. And the name lies. Exact match has matched close variants since 2018, and no setting turns that off.

   • Count next month's open new-patient slots before you pick a number. That is the real ceiling, not the budget field.
   • One campaign is one budget and one goal. Not one per condition, not one per zip code, not one per device.
   • Ten clicks a day is the readability floor. At $2.50 that is $25/day.
   • Maximize Clicks with a max CPC near $3 at launch. Conversion bidding after the booking conversion has fired 30 times.
   • Learning status clears in a week or two. The ~30 conversions in 30 days is a different test, and at 24 bookings a month you are below it.
   • Radius by drive time, and the location option set to presence rather than presence-or-interest.
   • Phrase match everywhere at launch. Exact is earned by a search that produced a booking.
   • Broad stays off until conversion bidding, real booking tracking and a maintained negative list are all true.

   The sums, before you touch the budget field

   • “Primary care doctor near me” is 17,000 searches a month at about $2.50. Ten clicks a day, the floor where a month of data starts to mean something, costs $25/day.
   • Second floor: at least 3x the cost of one booking. Assume 1 in 10 clicks calls and 2 in 3 callers book - your numbers, not mine. A call costs $25 and a new patient costs about $38.
   • 3 x $38 = $114/day. That buys about 45 clicks a day, four or five calls, and roughly 90 new patients a month.
   • Ninety. Go count your open new-patient slots, then decide whether that is a plan or a fantasy.
   • So run at the click floor rather than the learning floor. $30/day is 12 clicks a day, about 36 calls and 24 new patients a month.
   • Google can spend up to double the daily budget on a single day, but never more than 30.4x it in a month. $30/day is a $912 ceiling, not a promise of exactly $30.
   • What kills the month is splitting that $30 across three campaigns to see which one works. Four clicks a day each, thirty days later, and no verdict on any of them.

4. None of this is difficult. It is a long run of small decisions, and each one quietly costs money when it goes the wrong way. The agent makes them for you and shows you every single one before it touches the account.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
5. 4

   Twelve headlines, and none of them may promise anything

   The format gives you room for 3 to 15 headlines at 30 characters and 2 to 4 descriptions at 90. Ten or twelve headlines is the working number. Write an eleventh only when it says something the other ten don't, because a near-twin eats a combination Google could have learned from. Count every one by hand, spaces in, and remember the comma in “Primary Care Doctor, Tucson” costs a character like everything else. Google bolds whatever in your ad matches what the patient typed, wherever that headline lands, which is the whole reason the search term goes in word for word.

   Here is where medicine parts company with every other trade. You can't claim an outcome. No cure, no guarantee, no “get your life back”. Google's healthcare and medicines policy restricts it and your state board is stricter than Google's. There's a second gate if telehealth is on your list: Google makes US telemedicine advertisers get certified before those ads can run at all, and that application is separate from the ads themselves. Until it clears, keep telehealth out of the headlines however much of your Thursday it fills. What you can say is everything about access - who you're taking, which plans you're in-network for, how soon somebody can be seen, what a visit costs without insurance, where you are, what you're certified in. Ask ten patients why they picked their doctor and you'll hear about the plan, the wait and the parking. Nobody has ever said compassionate.

   Then the part almost nobody does. Google serves any three of your headlines together, in an order you don't choose, and you pay full price for whichever trio it picks. Don't sample the trio at random, because random passes. Go find the three that say least, read those three as one ad, and ask whether a stranger could name what's for sale. “Compassionate Care” and “Here For Your Family” and “Book Today” is a finished ad for a funeral home.

   Each headline does one job. The word doctor can repeat; the job never does. “Primary Care Doctor, Tucson”, “Doctor In-Network, Most Plans” and “$99 Self-Pay Doctor Visit” all say doctor and duplicate nothing, because one is the search term, one is insurance and one is price. What ruins a set is four headlines making the same warmth claim with the words shuffled. Vary the lengths on purpose too: twelve headlines all sitting at 29 characters and Google will fit two of them, so you paid for twelve and shipped eight.

   • 10 to 12 headlines at 30 characters or fewer, 3 or 4 descriptions at 90. Count them by hand.
   • No cure, no guarantee, no promised result. Not in a headline, not in a description, not on the page behind it.
   • Telehealth in the copy? Certification first. Google will not run those ads for a US practice without it.
   • Every headline names a doctor, a clinic or a named service. “Compassionate” names nothing.
   • One job each: the search term, insurance, price, availability, credential, location, the ask.
   • Read the three weakest together as one ad. If a stranger cannot name what is for sale, rewrite one of them.
   • Mix the lengths on purpose. All 29s and 30s and Google shows two.

   Nine headline drafts and the two that did not survive

   • Primary Care Doctor, Tucson (27) - the search term, in the words they type it in.
   • Meet Your Family Doctor (23) - what a new patient is actually shopping for.
   • Doctor In-Network, Most Plans (29) - the question that decides most bookings.
   • $99 Self-Pay Doctor Visit (25) - a real cash price, doing a negative keyword's job.
   • Board-Certified Family Doctors (30) - true of the whole roster, or it doesn't go in.
   • WEAK: Feel Better, Faster (19) - an outcome claim, and it could be a gym or a cough syrup.
   • FIXED: Next Appointment In 3 Days (26)
   • WEAK: Caring For Your Family (22) - a claim no patient can check and no competitor would dispute.
   • FIXED: Doctor For All Ages (19)
6. 5

   Block the television, keep the patients

   Negatives go in before the campaign serves its first impression. In most trades that's good hygiene. Here it's structural, because the single word your patients type is also 862,000 monthly searches of television, and an unprotected campaign finds that out on day one with your money. Campaign level, and lead with the services you refer out rather than provide. You already have a starter list further up this page. What follows is the same list with the match types spelled out and four things added that it doesn't carry: the portal and billing leak, the coding searches, the school searches, and the reversed pairs.

   A negative is literal in a way a keyword never is. No plurals, no synonyms, no close variants - it blocks the letters you typed and nothing adjacent. “Remedy” leaves “remedies” running all month. Same for “symptom”, and that one costs more, because the plural is how people actually type it. Read the finished list back and mark every word you only wrote once.

   Word order and match type matter more here than in any other trade. Negative phrase blocks those words in that order, so “doctor sleep” stops the horror film and does absolutely nothing about “sleep doctor near me”. Write the reversed pairs out separately. And “doctor who” is the one that catches everybody: as a phrase negative it also blocks “doctor who is accepting new patients”, which is a patient typing a sentence. Both of the reversible titles go in as exact instead, then you block the show's own vocabulary as phrases. “Good doctor” needs the same care for the opposite reason - “looking for a good doctor near me” is somebody describing what they want.

   Then a leak only medical accounts have. Your own patients Google their way to the portal. Your practice name plus “patient portal”, “pay my bill”, “medical records request”, “how do I get a refill”. Your ad shows because you're bidding on your own market, they tap it because it sits above your own listing, and you pay for a click to send somebody you already have to a page that's open on their phone twice a month. The click is cheap. That's the problem - nobody ever notices the row. Those wordings tend to arrive in week two rather than day one, and step seven shows you what they look like when they do.

   The last thing is the reason half of these lists cost practices money instead of saving it. A great many of the obvious negatives in this category are buyers wearing a browser's coat, and one bare word quietly deletes your best queries. The careful line at the end of the list below is the important part. Read it before you paste anything.

   • Quotation marks on every line before you paste. Without them Google files it as broad, and broad blocks things you never read.
   • Campaign level, in place before the first impression, not after the first invoice.
   • Both forms of every word. Negatives match neither plurals nor synonyms.
   • Exact match for “doctor who”, “doctor sleep” and “good doctor”, because all three reverse into real patient queries.
   • Write reversed pairs separately. “Doctor sleep” and “sleep doctor” are two different blocks.
   • Block the portal and the billing wordings. Those clicks are patients you already have.
   • Delete any line from the list that you actually sell. Somebody else's negative is your best keyword.

   What goes in before the first impression

   • Every line below goes in as phrase match, which in the Google Ads box means typing the quotation marks yourself: "free clinic". Paste a bare line and Google files it as a negative broad, and a negative broad blocks any search holding all those words in any order. Negative broad "free clinic" then kills "free sports physical clinic near me", which is a whole family walking in the door. Square brackets for the exact ones: [doctor who], [doctor sleep], [good doctor].
   • doctor strange, doctor doom, doctor odyssey, the good doctor, plague doctor, house md, greys anatomy, grey's anatomy - phrase is safe for all of these
   • [doctor who], [doctor sleep], [good doctor] - exact only, because as phrases they eat "doctor who is accepting new patients", "doctor sleep apnea near me" and "looking for a good doctor near me"
   • Then the shows' own vocabulary as phrases, both forms of everything: doctor who season, doctor who seasons, doctor who episode, doctor who episodes, doctor who cast, doctor who actor, doctor who tardis, new doctor who, watch doctor who, doctor sleep movie, doctor sleep cast, doctor sleep ending, good doctor cast, good doctor season, good doctor episode, good doctor finale
   • cast of, full episode, season finale, watch online, streaming, showtimes, spoilers, marvel, netflix, imdb, funko, cosplay, halloween costume
   • physician salary, doctor salary, family medicine salary, nurse salary, physician jobs, medical assistant jobs, front desk hiring, now hiring, job openings
   • medical school, med school, pre med, residency programs, match day, mcat, usmle, pa school, np program, how to become a doctor, locum tenens
   • symptoms of, signs of, what causes, why do i, how long does, is it contagious, home remedy, home remedies, natural cure, webmd, mayo clinic, reddit
   • icd 10, cpt code, medical coding, billing code, prior authorization form - and superbill only if you don't hand them out, because "doctor that gives a superbill" is an out-of-network patient choosing you for that exact reason
   • free clinic, free clinics, low income clinic, sliding scale, charity care, county health department, community health center - and pull "sliding scale" and "charity care" out if you run either, in which case those are pages you should be advertising rather than searches you should be blocking
   • health insurance quotes, buy health insurance, insurance marketplace, insurance agent, obamacare enrollment, obamacare sign up, best medicare advantage plan, compare medicare advantage, medicare advantage enrollment, medicare advantage broker, medicare supplement quote, medigap quotes - somebody shopping for a plan, never the plan name on its own
   • mychart, pay my bill, bill pay, medical records request, records release, fmla form, disability paperwork - and "fmla form" and "disability paperwork" come out if you charge for forms, because that is a booked visit like any other
   • Everything you do not do: dentist, dental, veterinarian, chiropractor, optometrist, eye exam, podiatrist, obgyn, emergency room, dot physical, esa letter, emotional support animal, medical marijuana card, weight loss injection. Delete every line of that you actually sell, because those are your best keywords rather than your negatives.
   • Careful, and this is the expensive part. Never bare "near me": "doctor near me" on its own is 30,000 searches a month at about $2.50, "primary care doctor near me" another 17,000, and one bare negative takes the lot. Never bare "insurance": "family doctor that takes my insurance" is the best query in this market. Never bare "cost" or "price": "how much is a physical without insurance" is a self-pay patient with a wallet open. Never bare "free": the free sports physical clinic is how half of family medicine meets new families. Never bare "symptoms": block "symptoms of" as a phrase and leave "flu symptoms doctor near me" running. Never bare "cast": a walking cast and a cast removal are both appointments. Never bare "season": flu season is a booking. And watch the ones with a plan name buried in the middle - never bare "medicare", never bare "medicare advantage", never bare "obamacare", because "doctors that accept medicare advantage plans near me" contains all of it word for word and one phrase negative takes the whole panel with it.

7. Everything past this point is the work that never finishes: reading search terms, adding negatives, checking what the budget did overnight. That is exactly the part the agent takes off your hands.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
8. 6

   The phone rings all day, and almost none of it is a new patient

   A plumber's phone rings with jobs. A practice's phone rings with everything. Refill requests, a pharmacy on hold, a billing question, somebody rescheduling for the third time, and occasionally a new patient. Google counts all of it as a conversion, which makes the raw call number the least trustworthy figure in a medical Ads account. It's also the number most practices are bidding on.

   Two things make it honest. The first is a minimum call length, so eleven-second wrong numbers don't count. Sixty seconds is Google's default and it's a judgment, not a law. Set it under your shortest real booking, never at it. Time ten genuine new-patient calls at your own front desk and take the shortest one: if that runs two minutes, 90 seconds keeps every booking and drops the pharmacy; if your desk books in ninety, 45 is the number. Set it at the average and you'll delete half the patients you paid for, then spend a month wondering why the campaign died.

   The second is better, if you can have it. When a new patient lands on a confirmation page you control, that page is a real conversion, and Google hands you two snippets to make it one - one for the site header, one for the confirmation page. Check whose page it actually is first. Most practices book through a third-party scheduler sitting on somebody else's domain, and you can't tag a page you don't own. Ask the vendor what it can send back before you plan around it. Some will pass a conversion, some will pass nothing and leave you on call length, and finding out which one you have is a ten-minute email rather than a project.

   Before anyone speaks to a developer, two switches already sitting in the account. Auto-tagging: with it off, nothing your website measures ties back to the click that caused it, so campaigns that are working look dead and get paused by whoever reads the report. And half the practice accounts I open have a conversion in there already, imported from Analytics months ago and never switched on. Go look before you build a new one.

   Then a figure only you have. What is a new patient worth, not for one visit but across a first year, multiplied by the share of callers who actually become patients. Without it Google treats a $99 cash physical and a paneled family of four as the same event. Keep the number conservative, because an inflated patient value teaches the bidding to overpay for twelve months. And one hard rule while you're in there, about what leaves your site rather than what you call things: don't send a condition to Google. No diagnosis in a URL parameter, no audience built from who read the diabetes page, nothing about a named condition in data you upload. Google restricts personalized advertising built on health information, and it won't sign a business associate agreement for Ads, so nothing identifying a patient or their health belongs in the account at any point. A general list of everyone who visited the website is a different thing and is allowed - though if your whole site is one condition, even that list stops being general.

   Last, the hours. Ads run around the clock unless you say otherwise. If the only way in is a telephone that reaches voicemail at five, every evening click sends somebody to the practice down the road. If your online booking works at eleven at night, the opposite is true and you should run those hours deliberately. Either way the schedule should match what happens in your building, not Google's default.

   • Turn auto-tagging on if it is off. One switch, nothing else changes.
   • Check whether Analytics already handed Google a conversion that is sitting there switched off.
   • Set a call conversion with a minimum length just under your shortest genuine booking, not at your average one.
   • Ask your scheduling vendor what it can report back before you plan a booking conversion around a page you may not own.
   • Value a new patient over a first year, times your booking rate. Your figures, never a benchmark.
   • No health data leaves your site: not in a URL, not in an upload, not as an audience built from a condition page.
   • Set an ad schedule to the hours somebody answers, unless patients can genuinely book online at midnight.
   • Nothing is backdated. The counter starts the day you save it, so turn it on before the ads run.
9. 7

   Every other Tuesday, forever

   Day one and day two you check two things. Is it spending, and is the ad approved. An account that can't serve looks identical to an account warming up slowly, so zero spend on day two is a fault rather than patience, and there are four usual culprits. A policy hold sitting in an email address nobody at the practice reads - medical ads sit in review longer than most, and a healthcare flag can stop the whole ad rather than only the headline that caused it, so open the asset-level policy view, not the ad status alone. A location target left on the state instead of the drive time. A landing page that 404s because the website was rebuilt in June. Or a campaign still saved as a draft. Go find out which fault it is before you touch a keyword.

   Day seven, open the search terms report and do exactly one thing: add negatives. No bids, no budgets, no harvesting, no conclusions. Week one in this category is largely television and you should expect it. A bidding strategy still in its learning period will punish you for fiddling, and seven days of data is noise wearing a chart.

   Two weeks to a month in, you can finally read something, and you read it by ad group rather than blended. Here's the arithmetic that tells you what to do with it. At $30 a day you buy about 360 clicks in a month. Those arrive as maybe 120 different search terms, 24 of them end in a booking, and those 24 come from perhaps a dozen distinct wordings. So the harvest in month one is three or four promotions, not thirty: the wordings that produced two bookings or more get added as exact match with their own ad, and everything with a single booking waits another two weeks to prove it wasn't luck. That harvest is the job, and it stays the job for as long as the account runs.

   Month two you owe each campaign a verdict: more money, repairs, or the off switch. And one decision this trade has that the others don't - if the schedule filled up, the correct move may be to cut the budget. Nobody in marketing will ever suggest that, so you'll have to. Then know where the cost really sits. It isn't the build. It's the two weeks you skipped because two providers were out and the waiting room was backed up to the door, while the search terms kept arriving anyway.

   • Day 1-2: is it spending, and is the ad approved. Check the asset-level policy view, the location target, the landing page and whether the campaign ever left draft.
   • Day 7: search terms, negatives only. No bids, no budgets, no verdicts.
   • Day 14-30: cost per booked patient by ad group. Promote the wordings that produced two bookings or more.
   • Month 2: more money, repairs, or the off switch. If the schedule filled, cutting the budget is a legitimate answer.
   • Put the half hour in the same calendar the clinic runs on, or it will not happen.

   Fourteen days of search terms, row by row

   • “primary care doctor near me that takes cigna” - 9 clicks, 3 calls -> add as exact match and give it its own ad with Cigna in a headline, because Google bolds the matched words. Two conditions. It has to be true on the day the ad runs, and a carrier can file a trademark complaint about its name in your ad text. That happens, it usually gets sorted out, and it's a call to expect rather than a surprise.
   • “family doctor accepting new patients tucson” - 7 clicks, 2 calls -> add as exact match, with its own ad.
   • “how much is a physical without insurance” - 11 clicks, 1 call -> leave it running. It reads like research and it's a self-pay patient. This row is why “cost” and “insurance” are never bare negatives.
   • “doctor for sleep problems tucson” - 7 clicks, 0 calls -> $17.50 gone. You don't do sleep medicine, and the exact negative [doctor sleep] from step five never came near this wording. Add “sleep problems”, “sleep study” and “sleep doctor” as phrases.
   • “doctors that take medicaid near me” - 8 clicks, 0 calls -> $20 for a real patient who isn't yours. Block “medicaid” as a phrase, and only because you don't take it. If you do take it, that's the best $20 in the table and it belongs in a headline.
   • “willowbrook refill request form” - 6 clicks, 0 calls -> your own patients, looking for a form. Add “refill request” and “refill form”.
   • “willowbrook patient portal login” - 6 clicks, 0 calls -> also your own patients, and the brand campaign paid for them at brand prices. Brand clicks are cheap, which is exactly why nobody ever finds this row. Negatives “patient portal” and “portal login”, in the brand campaign as well as the other two.
   • “walk in clinic near me” - 12 clicks, 1 call -> $30 for one call, and it's the row you'll argue with yourself about. You're appointment-only, but a same-day sick visit is what this person wanted. Send it to the sick-visit group with an ad that says same-day appointment, no waiting room queue, before you block it. If month two says it still costs $30 a call, block the phrase “walk in clinic” then - and leave bare “walk in” alone either way, because “walk in appointment today” is a booking.
   • “doctor who is accepting new patients” - 4 clicks, 1 call -> the row that proves “doctor who” belongs in the account as an exact negative rather than a phrase. As a phrase it would have blocked this booking and you'd never have known it existed.
   • Every one of those slipped past a list already blocking “doctor strange”, “the good doctor”, “salary” and “symptoms of”. The list is never finished, and that's the whole reason for the every-other-week read.
   • And these are only the terms Google will show you. Searches too rare to be anonymous never appear at all, so the rows won't reconcile to the invoice. Don't claim you've accounted for every dollar.

The ad one Tucson ad group actually runs

This is the ad for a single ad group - people in Tucson looking for a family doctor who is taking new patients - written for a practice that doesn't exist. Twelve headlines counted by hand, and one description. The card below draws the first three, because roughly three is what a searcher gets; the other nine sit underneath as the pool Google picks from. It assembles any three of the twelve in an order you don't choose, so the ad a patient sees is almost never the one you had in your head while writing it.

• Two headlines carry the search term. The other ten earn their place another way.. “Primary Care Doctor, Tucson” and “Doctor Near You In Tucson” are the words people type, and Google bolds them wherever those headlines land, which is the only reason they exist in that exact wording. Google's own guidance asks for at least three headlines with no keyword in them, so the rest say doctor differently or name a service instead of repeating the term twelve times over.
• Every line here is an access fact.. Who you're taking, how soon, what it costs cash, where you are, what you're certified in, which plans you're in-network for. Patients were never going to pick a doctor off an adjective. They pick the one who takes their plan and has a Thursday.
• The insurance headline is the one most practice ads leave out.. Ask patients why they picked a doctor and “they take my plan” comes ahead of almost everything else on the list. “Doctor In-Network, Most Plans” is 29 characters doing the work of a phone call to reception. If an ad group is built around a named insurer, name it in a headline - Google bolds it - with the trademark caveat from step seven in mind.
• “Board-certified” is a regulated claim, not a marketing adjective.. Most state boards allow it only for a board their own rules recognize, and it has to be true of whoever the patient actually gets. Two MDs and a nurse practitioner is not “Board-Certified MDs” on the week the NP has the openings, and it's plainly wrong if one of the two is a DO. That's why the headline reads “Board-Certified Family Doctors” - write what holds for the whole roster.
• Don't sample the trio at random. Random passes.. Here the three that name a doctor least are “Next Appointment In 3 Days”, “Accepting New Patients” and “Medicare Patients Welcome”. Read them together as one ad: a stranger still knows this is a practice taking new people who can be seen quickly. That's the pass mark. On its own, “Next Appointment In 3 Days” would fit a groomer, which is exactly why it never travels alone in this set.
• The $99 turns away the clicks you were going to lose anyway.. A cash price in the ad text is the cheapest filter in the account. Somebody with no insurance and forty dollars reads $99 and doesn't click, and you don't pay the $2.50. Use a real number or none at all, because “affordable” filters nobody. It stays in a new-patient ad even though self-pay has its own campaign, because a cash patient who wants a regular doctor is still a new patient - the separate campaign exists to protect the budget, not to hide the price. And check the figure against what reception quotes, since an ad price your own staff contradicts costs you the patient and the click.
• Twelve headlines, twelve jobs, and one trust claim between them.. The search term, the near-me variant, the offer, who you get to meet, the ages you take, insurance, the cash price, the credential, the wait, Medicare, the annual physical, and one plain ask to book. Deliberately absent: “compassionate care”, “patient-centered” and “your health is our priority”. Those are one claim in three costumes, and here it gets no slot at all, because the board certification says the same thing with evidence attached.
• The lengths run 19 to 30 characters, and that is on purpose.. Google shows up to three headlines, but only when they physically fit. Twelve headlines all sitting at 29 or 30 and it fits two most of the time, so you paid for twelve and shipped eight. “Doctor For All Ages” is nineteen characters and still names what's being sold, which is the whole trick with short assets. One line sits at the full 30 because the credential wouldn't survive being trimmed.
• Four sitelinks, and the first one answers the question that decides the booking.. Extensions are free and your competitor already has them, so an ad without them is physically smaller on the same screen. “Insurance We Accept” catches the patient who won't call until she knows, “Self-Pay Prices” catches the one with no plan at all, and neither of them needs a change to the main landing page.

The shape a three-provider practice can actually staff

Two campaigns that spend and one small one that defends. Four ad groups, and only three of them chase new demand - fewer than feels right, and the right number anyway: at roughly two dozen bookings a month, three working groups see eight bookings each. Split it once more and every group is down to six, which teaches Google's bidding nothing. Every keyword starts as phrase match. Exact is something a search term earns after it has produced a booking.

Why it's shaped like this. New-patient search and same-day sick visits share a campaign because they share a budget and a goal, but they sit in separate ad groups because the ad has to say something different for each, and they should land on different pages. One says we're taking new patients. The other says we can see you today. Self-pay is its own campaign because it's a different goal with a different value per booking; blended into one budget the cheaper clicks quietly eat the ones worth more and you'd never watch it happen. Your own name always sits alone. It converts better than anything else in the account and costs almost nothing to buy, so leave it in the main campaign and it lifts the average until new-patient acquisition looks fine, which is how a practice keeps funding a campaign that lost money in week two. Three things are missing on purpose. There's no ad group per condition, because a family practice treats hundreds of things and forty starved groups are worse than three fed ones. There's no condition-page remarketing, for the policy reason set out in step six. And telehealth has no campaign here, which is a gate rather than a preference: Google makes US telemedicine advertisers get certified before those ads run at all, and even then the targeting can't be wider than the states you're licensed in. That's a question for your credentialing file before it's a question for Google.

Google Ads management for doctors, shown as cards you approve one at a time.

What you actually pay a healthcare marketing agency for is the hands-on part. Building campaigns around the conditions you treat, digging through search terms for money leaks, keeping the account inside Google’s health rules, writing ads that bring in appointments. AdPilot’s AI handles that, and shows the reasoning behind each move.

Examples of what the cards look like - not results from your account

An empty clinic list does not wait for the monthly report.

You keep control, and you stop being dependent.

Most practice owners never learn how their own Google Ads account works. Nobody explains it, and an agency has little reason to start. AdPilot is built the other way round. Every card tells you:

• What it changes - for example, pause a keyword that spent $319 and booked nobody.
• Why it helps - the thinking, not only the instruction.
• What it is based on - figures pulled from your own account.

Approve one and a snapshot is saved first. Then a before and after dashboard shows what happened to cost per new patient, clicks and spend, in words that do not require a marketing course. After a month or two you know which levers actually fill the diary.

AdPilot is new - and we’d rather earn your trust than fake it.

We could plaster this page with five-star quotes and a “trusted by 10,000 businesses” banner. We won’t - because AdPilot is new, and inventing proof would be lying to you. Instead, here’s what actually protects you:

• You risk nothing to try it. 7 days free, no credit card. If it isn’t useful, you walk away having spent nothing.
• Your account is never at risk. Nothing is ever applied without your one-tap approval, a snapshot is saved before every edit, every change has one-click revert, and a full audit log records it. The worst case is a suggestion you don’t like - so you just don’t approve it.
• Guardrails you can’t trip by accident. AdPilot never deletes campaigns and never touches shared budgets. Some doors simply don’t exist.
• Built by an operator, not a VC-funded black box. Software made to give small businesses the Google Ads work an agency charges a retainer for - without agency prices or the loss of control.

Judge us on the work, the math, and the safeguards - on your own numbers, not our adjectives.