A flat monthly price looks good until you add a second location, cross a spend threshold, or need one more landing page. The sticker price is the opening bid, not the bill.
I wrote this for owners and small agencies comparing AI Google Ads tools that claim to include everything. These nine costs follow patterns you can catch before signing. For each one, I’ll use a $20k-a-month advertiser where the math helps and give you a question to send the vendor in writing.
1. Ask whether your fee rises with ad spend
A percentage-of-spend fee is easy to miss when the pitch calls it a management or platform subscription. The mechanism is less elegant: as your media budget grows, the fee grows with it, whether or not the work gets harder. At $20k a month in ad spend, a 15% fee is $3,000 a month. At $10k, the same rate costs $1,500. You pay twice as much because you spend twice as much, not because anyone made twice as many decisions.
I have watched that curve eat scaling budgets. Fifteen to 20% of ad spend remains a standard agency pricing model, and software can use spend-linked platform fees too. An enterprise version charges on connected media spend, including campaigns that never run through the tool. Smartly uses a percentage of total connected spend, has no public pricing page, and has reported contracts around $100k to $300k. In the draft example, a $2M account at 3% pays $60,000 a month even if only $300k uses the platform.
The first step is to ask for the fee at several budgets, not just today’s. It also helps to compare what CPA-based and spend-based pricing reward. Question to send in writing: is any part of my fee tied to spend, including connected accounts I do not run through you, and what do I pay at $20k, $40k, and $80k?
2. Price the next spend tier before you reach it
Spend tiers are a different version of the same problem. Instead of taking a percentage, the vendor puts you in a band based on connected spend or account count. The advertised entry price is real, but only while you stay inside its limits. Cross the line and the subscription changes in one jump.
Optmyzr lists Starter at $208 a month for up to $25k in spend and 25 accounts, Professional at $416 for up to $100k, and Agency at $832 for up to $500k. Start at $20k a month and a move to $40k takes you out of that Starter spend band. That is the number to put beside your growth plan, not merely the $208 you saw during the demo.
I used to tell clients tiered tools were predictable. I was wrong, or at least incomplete. They are predictable if you know the trigger and the price on the other side. Otherwise the month you raise budget is also the month your software bill jumps. Send the vendor your current spend and a plausible doubled-spend scenario. Question to send in writing: what exact plan and price am I on at 2x my current spend, and what triggers the move?
3. Multiply the price by every account and location
The one-account demo is tidy. The order form may charge per connected account, location, or domain. For a small agency, that distinction can matter more than the advertised monthly rate. Say you spend $20k a month across five locations and the tool costs $199 per location. You owe $995 a month for software before media. Add a sixth location and the fee rises even if the total ad budget does not.
That is the mechanism: the bill scales with how your business is organised, not with what the ads produce. A growing client can make the tool more expensive without becoming more profitable. Conversely, a quote for one location tells a five-location buyer almost nothing useful. Before taking a demo, write down every account and location you expect to connect, then ask for one total covering all of them.
For businesses, groas lists $1,499 a month per service regardless of ad spend. That is a different pricing unit, so compare the whole proposed setup, not two prices stripped of what each covers. Question to send in writing: what is my total if I add three accounts or locations, with nothing else changing? If the vendor cannot put that multiplied total in one line, you still do not have a usable quote.
4. Check the allowance behind “landing pages included”
“Landing pages included” can describe access to a separate builder, not unlimited pages, visitors, and tests. Visitor allowances and testing features are where the cheap-looking package becomes a bigger bill. Pages are not much use if the plan that keeps them live cannot handle the traffic you intend to send.
Unbounce Build costs $99 a month for 20k visitors; Experiment costs $149 and adds A/B testing; Optimize costs $249 for 50k visitors, plus $80 per extra 10k-visitor block. At 90k visitors, that Optimize example reaches $569 a month for pages alone. Instapage Optimize costs $199 for 30k visitors, while allowances above 50k move to a custom annual contract. Those are different ways of pricing the same practical need: keep the funnel running while traffic grows.
The overage policy matters as much as the rate. Instapage warns at 80% of its allowance, locks past 100%, and may unpublish pages after repeated overages; Unbounce keeps pages live and bills overage. For a $20k-a-month advertiser driving 40k visitors, an allowance that looks generous in a demo can become an operational problem. First, estimate the traffic those pages must handle. Question to send in writing: how many live pages, visitors, and A/B tests are included, and what exactly happens and bills at 101%?
5. Get the day-one invoice, not just the monthly quote
Implementation is setup with a better suit on. A tool can quote $499 a month, then add $2,500 in a statement of work to build campaigns, migrate history, and fix tracking. That work may be necessary. It still belongs in the price you use to decide whether to buy.
Agency onboarding commonly runs $1,000 to $5,000 on top of $2,500 to $5,000 in monthly management for mid-size B2B accounts. At $20k a month in media spend, a $3,000 setup fee plus 15% management makes $6,000 in first-month overhead before Google charges for a click. The bill lands at the start because the work does: keyword builds, feed rules, and conversion wiring need attention before steady-state management means much.
Do not treat a setup fee as automatically suspect. Treat an undisclosed setup fee as the problem. I would rather see a clear line for necessary work than discover it after planning around the subscription alone. Ask what the fee pays for and what is working when onboarding ends; otherwise you can pay for implementation and still be left sorting out tracking yourself. Question to send in writing: what is my total day-one invoice, including the subscription and every one-time fee, and what tracking is working at the end of it?
6. Put your approval hours on the invoice
Some tools do not charge for the operating labor because they leave it to you. They flag anomalies, draft rules, and recommend changes. Then someone on your team reviews, approves, pushes, and maintains those changes. The subscription buys suggestions; your payroll turns them into action.
Even a small account can need an 8-to-12-hour monthly floor, while rule-based tools can take four to six weeks to learn and still need ongoing maintenance. At a $75k salary, two hours a week works out to roughly $72 a week, or $3,744 a year. That does not count the delay when a useful recommendation sits in a queue. For a $20k-a-month advertiser, eight unbudgeted approval hours every month are not “automation.” They are another job with a nicer dashboard.
I ran accounts that way for years. An approval step can be valuable when a change needs judgment. It becomes expensive when every routine adjustment waits for a human click and nobody has time to make it. Before comparing subscriptions, ask for a live walkthrough of a change from alert to execution. Find out where your work begins, not just where the vendor’s demo ends. Question to send in writing: what changes go live without my click, who is accountable if I approve nothing for 14 days, and how many hours a week do your median $20k advertisers spend in the tool?
7. Count usable creative, not generated creative
“Unlimited generation” sounds generous until you ask what you can publish. A tool may generate hundreds of headlines or images while metering downloads, published assets, or tests. The useful unit is not the draft sitting in the interface. It is the creative you can put to work without another charge.
AdCreative.ai Starter costs $39 a month for 20 downloads, or about $1.95 each; Professional costs $249 for 300 downloads, and unused credits do not roll over. Generating more options does not change that allowance when downloading one uses a credit. If you spend $20k a month across Search, Performance Max, and Meta and need 60 fresh assets for testing, a 20-download plan does not cover the month. You either move up a tier or limit the work the tool was meant to help you do.
This is not an argument for publishing every variation a generator produces. It is an argument for pricing the amount you expect to use. Write down the ads, images, and page variants in your actual testing plan, then compare that list with the plan’s metered actions. Question to send in writing: how many published ads, images, and landing-page variants can I run per month before I pay more, and do unused credits carry forward?
8. Calculate the cost of leaving in month three
A 12-month term is a price, not paperwork. It determines how long you pay if performance falls short or the support you expected does not arrive. Legacy contracts can run for at least 12 months with 60- or 90-day auto-renewal notice; more flexible arrangements can be month to month with 30 days’ notice. Neither the monthly fee nor a cancellation button tells you the full commitment.
Say you spend $20k a month on ads and agree to $2,000 a month in fees for a year. That is $24,000 in committed fees even if CPA climbs in month two. A shorter plan can still catch you with a notice window. One creative tool requires cancellation seven days before month-end on a monthly plan and 30 days’ notice on a yearly plan. Miss the window and you buy another cycle after deciding to leave.
I would put the cancellation terms beside the price, not in a folder for later. The first step is simple: ask the vendor to calculate a month-three exit under the contract it wants you to sign. Include renewal notice, not just the initial term. Question to send in writing: what is the initial term, what is the renewal notice in days, and what is my total committed spend if I cancel in month three?
9. Confirm you keep the account when you leave
Exit is where a cheap tool can become expensive. If the vendor owns the Google Ads account, landing pages, or history, cancelling can mean rebuilding instead of simply changing who manages the work. The monthly subscription ends. The loss of accumulated account information does not.
The account-ownership guidance is to keep the ad accounts in your name from day one and grant the provider guest access. Otherwise you risk losing bid history, Quality Scores, and conversion history when you cancel. At $20k a month in spend, having to start again can cost more than a page overage or an extra software seat. You do not need a precise forecast of that loss to know it belongs in the buying decision.
Before a demo, check whose name is on the accounts and what happens to pages, audiences, and data at cancellation. Do not settle for “we make offboarding easy.” Ask for the ownership and transfer terms in writing. groas says your accounts, pages, and data stay yours, with no setup fee and month-to-month terms. That gives you a concrete line to ask any other vendor to match.
Question to send in writing: do I own the ad accounts, pages, audiences, and data from day one, and will you confirm transfer within 14 days of cancellation? If I could ask only one question before a demo, I would start here. The other eight can raise the bill. Losing the account can take away the work you already paid to build. A vendor that needs a call to explain who owns it has already told me enough.

