Google Ads Management Services - Flat Price, You Approve Every Change

Somebody has to sit down with the account every other week. The question is who, and on what terms.

Google Ads management services come in three shapes - an agency, a freelancer, or software that does the work and asks before it changes anything - and you've already decided you want one, which is why you're reading a page like this instead of a tutorial. Two versions of you land here. One is paying somebody now, has twelve invoices in a folder, and worked out last week that the fee went up in exactly the months the business did well. The other has never bought Google PPC ads in their life, has no intention of learning what a match type is, and wants to know who does this and what it takes to have it done.

Google Ads for business gets sold like software - a login, a dashboard, a monthly charge. What you're actually buying is somebody's time, and it comes in two parts. The build is two or three days of work and it happens once. Everything after that is roughly an hour every other week, forever: reading the search terms report for the money that left, blocking what should never have matched, harvesting the search that produced a real inquiry, checking the ad is still approved, and deciding once a month whether to feed a campaign, fix it or stop it. That hour is the whole service. It isn't difficult work. It's work that has to actually happen, in a week when the van's out, the quote is half written and nobody has time.

AdPilot: the fully automated Google Ads management service

AdPilot does that hour. It reads your numbers every morning, so a campaign that went dark overnight or spend that doubled on a Tuesday reaches you the next day rather than in a monthly report. It works the account properly on a regular cycle, and it answers the minute you open the chat and ask, which is when most people use it. Everything it wants to change arrives as a card: the change, the evidence out of your own numbers, and what it costs. You tap approve or skip, and nothing reaches Google until you do. If you'd rather not open it at all, Takeoff is the version where we get on a call, build the account and run it for you - and you still approve every change, because that part doesn't get taken away from you on any plan.

You decide every step of it:

• 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.

What managing a Google Ads account actually involves, month after month

This is the honest answer to “what am I paying for”. It isn't a build guide for one trade. It's the recurring cycle - the same order AdPilot's agent works in, and the same order a good manager works in. Read it and you can do the whole thing yourself. Read it, decide you'd rather approve cards than make forty judgment calls a month, and that's a fair conclusion too.

The worked examples all use one fictional business: Meridian Facility Care, a regional commercial cleaning company in Columbus. Nobody's industry page here, deliberately, so nothing in it is borrowed from another argument. Its click prices and volumes are a plausible account rather than a measurement, and where a real figure matters the text says where it'd come from in yours.

Both ways of having this run, flat and with no cut of your spend, sit on one page:

AdPilot pricing

1. 1

   The intake, and the five things no report contains

   A manager who starts by opening Google Ads has already skipped the part that decides whether the account can work. Five facts live only in your head, no report recovers them, and an agency that never asked for them is guessing with your money.

   What one customer is worth, and what share of inquiries you win. Not revenue - the arithmetic. Meridian's average office contract runs $1,400 a month and lasts about eleven months, roughly one inquiry in four turns into a walkthrough, and about one walkthrough in three signs. So twelve inquiries buy one signature, which makes an inquiry worth around $1,280 of contract value. Then do it again on what you keep rather than what you invoice: $15,400 of contract across eleven months isn't $15,400 in your pocket, and once the crew, the supplies and the supervisor are out of it you might hold $4,500. Across the same twelve inquiries that's about $375 of margin each. The $1,280 is the ceiling. The $375 is the decision, and it's the number a cost per inquiry actually has to beat. Guess either one high and you'll overbid all year.

   Then: what you don't sell, because that list becomes the block list and it's where most small-account waste hides. Who answers the phone and when, because ads run around the clock by default and a call at 11pm that rings out is paid for and then handed to a competitor. How far you'll genuinely travel for the work, which is drive time and job value rather than a circle on a map. And which jobs you want more of, as opposed to the ones you take to fill a gap in the schedule.

   Before any of it, one mechanical check that gets skipped constantly: can the account serve at all? No approved payment method, a suspended account or a test account all look exactly like a slow start and produce silence. The silence then gets blamed on the keywords for a month.

   • Average customer value, close rate, and what one inquiry is therefore worth. Your numbers, not a benchmark.
   • Gross margin on the work, so the cost per inquiry gets judged against money you keep rather than money you invoice.
   • Everything you don't sell, written down. It becomes negative keywords in the next step.
   • The hours somebody genuinely answers, Saturdays included.
   • How far the work is worth traveling, by job value rather than by miles.
   • Which jobs you want more of, as opposed to the ones that fill a gap in the schedule.
   • And before any of it: payment method approved, account not suspended. The only item on this list Google can tell you.
2. 2

   The build, which happens once and takes two or three days

   A campaign is a budget plus a goal, and that's the only reason to open a second one. Not per keyword, not per match type, not one per town on your service map. Each of those splits your conversion data into piles too small to read. An ad group is one theme: keywords close enough that a single ad answers all of them honestly. And don't build single-keyword ad groups. That's 2018 advice, exact match has matched close variants for years so the isolation they promised never existed, and four groups on three inquiries each teach Google nothing.

   Keywords get researched, never invented. The order of evidence matters: what already converted in this account beats the Search Console queries the site already ranks for, which beat the words on your own website, which beat any keyword tool. On a brand new account the first two are empty, so you start at the tool and you scope it to the places you actually serve - a worldwide volume for a company that works one metro is a number about nobody. Then run every term through one test: read it cold, with nobody standing there to explain what you do, and ask what else it could mean. “Cleaning” on its own is a chore, a tutorial and a job listing. “Commercial cleaning services Columbus” is a buyer. No match type rescues an unqualified word.

   Then the arithmetic, out loud, before a budget goes near the screen. At about $7.00 a click, $45 a day buys six clicks a day, roughly 190 a month. At a 4 percent inquiry rate that's 7 or 8 inquiries. Below about five clicks a day a month of data answers nothing at all, so if the number doesn't clear that, the fix is a narrower and higher-intent keyword set, never a wider one. Google can spend double the daily budget on a given day and never more than 30.4 times it in a month, so $45 a day is a $1,368 ceiling rather than a promise of exactly $45.

   Everything starts on phrase match. Broad has three prerequisites - conversion-based bidding, tracking that counts a real sale or lead, and negatives somebody maintains - and missing any one of them it'll spend the month proving it. The negatives ship inside the build, not afterwards, because a campaign that serves one impression without them starts buying job hunters within the hour.

   Budget arithmetic for Meridian, done before the campaign exists

   • $45/day at ~$7.00 a click = 6 clicks a day, about 190 a month.
   • At a 4% inquiry rate that is 7 or 8 inquiries a month. Judgeable in four weeks, barely.
   • An inquiry is worth ~$1,280 of contract, about $375 of it margin - so $170 an inquiry pays and $400 doesn't.
   • Split that $45 across three campaigns instead and each gets 2 clicks a day. Thirty days later: no verdict anywhere.
   • The monthly ceiling is 30.4x the daily figure. $45/day = $1,368, and Google may spend $90 on a Tuesday.
3. 3

   The first two weeks, which is a search terms report and nothing else

   Day one and day two, two questions. Is money leaving the account, and is the ad approved? Zero spend on day two is a fault rather than a slow start - a disapproved ad, a target set so low it throttles delivery, or a location setting that excludes the county you're sitting in. Nothing on the dashboard distinguishes an account that can't serve from one that's warming up, which is why this check has to be a habit rather than an instinct.

   The block list shipped with the build, so the obvious ones - jobs, hiring, how-to, franchises, house cleaning - never cost you a cent. Week one finds the stuff nobody could have guessed. Day seven, open the search terms report and block. Nothing else: no bids, no budgets, no harvesting, no verdicts. A bid strategy still in learning will charge you for the interruption, and week one is noise anyway.

   Two things about negatives catch nearly everybody. They don't stretch the way a normal keyword does: no plurals, no synonyms. “Job” leaves “jobs” running and “cheap” never touches “inexpensive”, so both forms go in every time. And pick the narrowest form that stops the bleeding. A negative broad on “free” also kills “free quote”, and “free quote” is how half your customers ask you to come and price the work.

   One limit worth knowing before somebody quotes you a number: Google hides low-volume search terms below a privacy threshold, so the rows you can see never add up to what the campaign spent. Anyone claiming to have accounted for every dollar of your spend is claiming something the report doesn't support.

   Week one at Meridian - 61 search terms, 5 of them worth blocking

   • “cleaning company for sale” - 3 clicks, $19 -> negative phrase “for sale”.
   • “janitorial supplies wholesale” - 4 clicks, $22 -> negatives “supplies” and “wholesale”.
   • “osha cleaning checklist template” - 4 clicks, $26 -> negatives “checklist” and “template”.
   • “carpet cleaning columbus” - 6 clicks, $39, 0 inquiries -> Meridian subcontracts carpet. Negative “carpet”.
   • “commercial cleaning franchise” - 2 clicks, $14 -> negative “franchise”. The build blocked “franchises”; negatives don't do plurals.
   • Blocked: 19 clicks, $120. Left alone: “commercial cleaning quote columbus”, 5 clicks and 2 inquiries. Come back for that one at day 30.

4. You're three steps into a cycle that repeats every other week for as long as the account runs. The agent runs it and shows you every decision on a card before it touches anything.

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

   The monthly read, and what a real decision looks like

   Once a month somebody has to look at the whole account and make one decision per campaign: feed it, fix it, or stop it. Everything else in a monthly meeting is commentary. The read has an order, because the numbers lie to each other if you take them in the wrong one.

   Split brand from non-brand first, before you look at any average. Searches for your own company name convert several times better and cost a fraction, and blended into one number they drag the average down and hide non-brand losing money underneath. Google gives you no brand flag; you split the rows yourself by name match. Meridian's blended cost per inquiry reads $114, which looks fine. Split it - 6 brand inquiries at $9, 10 non-brand at $177 - and the $177 is the number that decides whether advertising works for this business.

   Then ask what's limiting each campaign, because the fix is different. Losing impression share to budget means the campaign is capped and more money buys more of something already working. Losing it to rank means more money buys nothing and the work is in the ad, the landing page or the bid. Those two look identical on a spend chart and pull in opposite directions.

   A real decision names a number and a date. “Non-brand is at $177 an inquiry against $375 of margin, so we feed it - budget from $45 to $54, which is a 20 percent step, and we look again on the 14th.” Compare that with what a monthly report usually says, which is that impressions are up 34 percent. Impressions aren't a decision. They're not even evidence.

   • Brand and non-brand reported separately, always, with the split method stated.
   • Cost per inquiry against the margin one inquiry brings you - not against last month's cost per inquiry.
   • Impression share lost to budget versus lost to rank, because they need opposite fixes.
   • One decision per campaign: feed, fix or stop. Write down the date you'll next look.
   • A campaign that spent three times your target with real tracking and produced nothing across a full month isn't still learning.
6. 5

   The ad testing loop, which is the only free lever you own

   Ad text costs nothing to change, works the day it's approved, and is pure skill. Google takes three to fifteen headlines at 30 characters and two to four descriptions at 90, and you should give it twelve and three. Count them yourself, spaces included. A headline that repeats another headline is worse than an empty slot: it burns a combination Google could have tested and teaches it nothing.

   The check almost nobody runs comes next. Google serves any three of your twelve, in an order you don't choose, and you pay full price for whichever trio shows at 4pm on a Thursday. So find the three that name the job least, read those three together, cold, as one ad, and ask whether a stranger could say what's for sale. “Trusted Local Experts | Fully Insured | Free Quotes” is a complete ad, paid for at list price, that never says what's being sold. Don't sample any three - you'll pick three that pass.

   Give each headline one job and never the same job twice. The search term in the searcher's own words, the offer, the thing only you can claim, proof with a number in it, the risk reducer, the price, the city. What ruins a set is five headlines making the same trust claim in different clothes. “Trusted”, “reliable”, “quality service” and “you're in good hands” are one claim, and one claim gets one slot. Vary the lengths on purpose too: when everything runs 29 or 30 characters Google often finds room for only two.

   Then leave it alone long enough to mean something. A new ad needs weeks and real impressions before its performance labels say anything, and rewriting the set every couple of weeks because the numbers wobbled is how accounts end up with four years of tests and no conclusions.

   Four drafts for Meridian, counted, with two failures and their fixes

   • Office Cleaning, Columbus (25) - the search term, in the words a facilities manager types.
   • 92 Offices Cleaned Nightly (26) - proof, short, with a number in it.
   • WEAK: Trusted Local Experts (21) - could be an accountant, a locksmith, a dog groomer.
   • FIXED: Same Cleaners Every Week (24)
   • WEAK: Free Quotes Available (21) - free quotes for what?
   • FIXED: Cleaning Quote In 24 Hours (26)

7. Everything past here is the half that never finishes. That's the half people hire out, and it's exactly the half the agent takes.

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

   Budgets, bids, and the weeks you're supposed to do nothing

   The most expensive habit in a small account is changing the thing that was still learning. Google's bid strategies need roughly one to two weeks or about 30 conversions, whichever is later, and a documented list of things restarts that clock: a new or reactivated strategy, pausing and restarting one, switching strategy, and any edit to a target. There's no safe percentage on a target edit. Changing which conversion action counts isn't on Google's list at all, and it's deeper than anything on it, because it swaps the goal itself.

   Budget isn't on that list, which surprises people. Step budgets about 20 percent at a time for delivery stability rather than out of fear of learning. Targets are the opposite: step them 10 to 15 percent, never more, and set the first one at what the campaign's trailing 30 days actually produced rather than at the number you want. A target with no history behind it is a wish, and it strangles delivery to almost nothing - that's the real cause of the classic “my new campaign gets no impressions”.

   Some moves are safe during learning and genuinely worth making: adding negatives off the search terms report, fixing a landing page that doesn't load on a phone, adding sitelinks and callouts. Do those. Leave bids, targets and structure where they are until the clock has run.

   And know which bid strategy your volume can feed. Seven inquiries a month is nowhere near the roughly 30 conversions in 30 days a target CPA needs before it stops guessing. Switch it on at seven and you've bought a month of erratic spend and a bad opinion about your keywords. Start on clicks with a ceiling you set, move to conversion bidding when the account genuinely produces the volume, and price the switch out loud before you make it: the learning clock restarts, and if you change your mind it restarts again on the way back.

   • One change at a time. Never stack a second reset on an unfinished first.
   • Budget steps of about 20 percent; target steps of 10 to 15 percent, never more.
   • First target set at the trailing 30-day actual, not at the goal.
   • Negatives, landing page fixes and extensions are safe mid-learning. Bids, targets and structure aren't.
   • Under roughly 30 conversions a month, conversion-target bidding is guessing with your money.
9. 7

   Reporting that tells you the truth instead of flattering the invoice

   A report exists to support one decision, and most of them are built to survive a meeting instead. Impressions up 34 percent, click-through rate up, a chart with a nice slope, and nothing in it that could ever read as bad news. Ask a simple question of any report you're handed: which line would have looked wrong if last month had gone badly? If there isn't one, you're not reading a report, you're reading a renewal. One tell costs nothing to check, too: if the thing still quotes “average position”, Google deleted that metric in 2019, and whoever built the template hasn't looked at it since.

   Four things a straight report has. Brand and non-brand shown separately, with the method stated, because that single split is the most common way an account looks fine and isn't. Cost per inquiry against what an inquiry is worth to you, not against last month's cost per inquiry. The search terms that were blocked and the ones that were harvested, so the work is visible rather than asserted. And what the manager decided, with a date attached.

   Two admissions belong in it as well, and their absence tells you something. Google hides low-volume search terms, so the visible rows will never total your spend and nobody can honestly claim they accounted for every dollar. And nobody can see a competitor's real bids or their actual budget, whatever the slide with the competitor logos implies.

   Then the thing worth protecting above all of it: whether your tracking measures a real result. Tracking that counts page views as conversions produces beautiful reports and buys browsers all year. If the number on the report isn't a phone call, a form, a booking or a sale, the rest of the report is arithmetic about nothing.

   The same month at Meridian, reported two ways

   • FLATTERING: impressions +34%, CTR +0.4pts, top-of-page rate up, 280 clicks delivered.
   • STRAIGHT: 16 inquiries - 6 brand at $9, 10 non-brand at $177. An inquiry is worth ~$1,280 of contract, ~$375 of it margin.
   • STRAIGHT: blocked 8 terms ($196 of waste), harvested 2 converters into exact match.
   • STRAIGHT: lost impression share is 41% to budget and 12% to rank, so Office Cleaning is capped, not outranked.
   • STRAIGHT: decision - feed it. $45/day to $54/day. Next read on the 14th.
   • STRAIGHT: caveat - Google hides low-volume terms, so these rows don't total the $1,826 spent.

The finished ad for one office cleaning group

One ad group at Meridian - office cleaning contracts in Columbus. The card draws the first three headlines, because three is what a searcher gets; the other nine sit underneath as the pool Google picks from. One description is drawn, though Google won't save the ad with fewer than two and will take four. Nobody ever sees the ad you wrote. They see three of your twelve, chosen for them.

• The search term goes in one or two headlines, not twelve. Google bolds whatever matched the query, in whichever slot that headline lands, which is why “Office Cleaning, Columbus” is written the way a facilities manager types it. Google's own guidance also asks for at least three headlines that don't carry the keyword at all, so four of these never use the word: the evening crew, the site walkthrough, the rooms themselves, and janitorial as the term a facilities manager might type instead.
• Read the weakest three, because that trio is a real ad. Here the three that say least about the job are “Evening Crews, 6pm Onward”, “Book A Site Walkthrough” and “Desks, Restrooms, Floors”. Read them cold, as a stranger: somebody comes in from six, you can book a look round the building, and desks, restrooms and floors are what gets dealt with. That trio survives, just. A set whose weakest trio is “Trusted Local Experts | Fully Insured | Free Quotes” doesn't, no matter how good the other nine are.
• Twelve slots, twelve different arguments. The search term with the city, the 24-hour quote, the same crew every week, the shortest name for the category, when the crew comes, 92 offices as proof, no lock-in as the risk reducer, insured and bonded as the credential, the site walkthrough as the ask, the price floor, the scope, and the word a facilities manager might type instead of yours. What's missing is the same trust claim four times over.
• The price in the ad is the cheapest negative keyword you own. “Cleaning From $0.09/Sq Ft” filters before the click rather than after it. The office manager whose budget is half that reads it, doesn't click, and doesn't cost you $7.00 and a walkthrough you were never going to win. If you can't publish a floor price, publish how you quote instead.
• Lengths run 19 to 28 characters, deliberately. Google shows up to three headlines only when they fit. A set of all-30s often serves two, and a third of what you wrote never appears. “Commercial Cleaners” is nineteen characters and still names the category, which is the point of a short asset - short, not subject-free.
• Four sitelinks, and most advertisers skip them. Two of them answer what a facilities manager wants before filling anything in: what it costs, and whether you cover their building. The third repeats the service for anyone who skimmed the headlines, and the fourth is the ask. They cost nothing, none of it touches the landing page, and an ad without them takes up visibly less of the phone screen than the two competitors sitting above it.

What a business on a real budget should actually run

Two working campaigns and a small defensive one, four ad groups between them. Each group is narrow enough that one ad answers every keyword in it and points at one page, and no narrower: on this budget, more groups means each one sees too few inquiries to tell you anything. Everything starts on phrase match. Exact is earned later, by the searches that produce real inquiries.

Why it's shaped like this. Office cleaning and post-construction cleans are split because they need different budgets and different pages, and because one is an eleven-month contract while the other is a job that ends on Friday. Office cleaning and janitorial sit in the same campaign as two ad groups: same buyer, same budget, different words on the page they land on. Brand gets a campaign of its own for one reason, and it's an accounting reason rather than a traffic one - at $9 an inquiry it makes any blended average look like a success, so it has to be countable separately or the $177 non-brand number never surfaces. Left out on purpose: residential cleaning, which is a different buyer at a different price point and would need its own budget; and anything national, because a cleaning crew that can't drive there can't sell there. The three above come to $70 a day, about $2,128 a month, and every dollar of it is spoken for. Adding a fourth now would only starve the two that work.

Put your own invoice into it

Three numbers you already know: what you spend on ads, what the base fee is, what the percentage is. It runs them forward to what the same arrangement costs when the business doubles, and when it grows fivefold. If yours is a flat fee, set the percentage to zero and watch the line stay where it is - that's the whole point of the comparison.

What a Google Ads service actually does, shown as cards you approve one at a time

Most of what follows assumes Google Ads for lead generation - a phone call, a form, a quote request - because that's what most businesses shopping for management sell. The work itself is unglamorous and it is the whole job: separating buyers from people who'll never buy, building a campaign per theme rather than per company, matching the spend to hours somebody answers, and writing ads that produce inquiries rather than clicks.

Stripped of the deck, what you're paying a Google Ads service for is four things on a schedule. Somebody opens the search terms report at least fortnightly and blocks what shouldn't have matched. Somebody harvests the searches that produced a real inquiry. Somebody checks the ads are still approved and still pointed at a page that loads on a phone. And once a month somebody makes one decision per campaign and writes down the date they'll look again. Ask any provider which of those four still happens in a week when they're short-staffed, and listen to how long the answer takes. Here it is as cards, each with its reasoning on the front.

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

Google Ads for local businesses: the radius, the phone, and the hours somebody answers

Google Ads PPC management you can read over somebody's shoulder

Handing this over usually means handing over the understanding as well. Two years later you know your cost per lead and nothing else, and switching provider feels like starting from zero because in practice it is. AdPilot works the other way round. Every card carries four things:

• The change - in one line, in plain English. Add “jobs” as a negative keyword, not “implement query-level exclusions”.
• The evidence - the rows it came from, out of your own account. 19 clicks, $128, zero inquiries.
• The argument - why this helps, so you can disagree with it. A recommendation you can't argue with is a recommendation you can't check.
• What it doesn't know - said out loud when it matters. Google hides low-volume search terms, and nobody can see a competitor's real bids.

Approve one and a snapshot is saved first. Then a before and after view shows what happened to cost per lead, clicks and spend, in words that don't need a PPC course behind them. A couple of months in you can read your own account, and that knowledge stays with you whatever you decide to do next - including hiring somebody, with the useful side effect that you'll know what to ask them.

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.