.webp)
Last updated: February 9, 2026 | Reading time: 16 minutes
I am going to tell you a story that will sound familiar if you have ever worked with a Google Ads agency. It does not have a dramatic inciting incident. Nobody stole money or committed fraud. The agency was not incompetent. They were perfectly average, which turned out to be the problem.
We spent 14 months paying an agency $4,200 a month to manage our Google Ads. That is $58,800 we will never get back, spent on a service that, looking back, delivered a fraction of what we assumed we were getting. The painful part is not that they were bad. The painful part is that we did not know what good looked like until we saw it.
This is the story of how we figured that out, what we did about it, and what happened to our numbers when we made the switch.
We started running Google Ads ourselves. Did the YouTube tutorials, read the Google help docs, built some campaigns, started spending money. The results were okay but not great, and we knew we were leaving performance on the table. We did not have the time or expertise to optimize properly, and the platform kept getting more complex.
So we hired an agency. A mid-sized shop with a decent website, case studies that looked impressive, a few Google Partner badges, and a sales pitch that hit every note we wanted to hear. "Dedicated account manager." "Data-driven strategy." "Weekly optimization." "Monthly reporting with actionable insights." We signed a six-month contract at $4,200 per month (a flat retainer, not percentage of spend, which we thought was the smart move at the time). Our monthly ad spend was around $18,000 to $22,000, so the management fee felt proportional. They seemed professional. We felt good about it.
For the first two months, things actually went well. The onboarding was smooth. They restructured our campaigns, which needed doing. Performance improved, which reinforced our decision. The account manager was responsive. The reports were thorough. We thought we had made a great hire.
Then month three happened. And months four through fourteen were month three on repeat.
The thing about agency mediocrity is that it does not announce itself. There is no single moment where you say "this is not working." It is a gradual erosion of attention that you only recognize in hindsight.
The reports got thinner. The first two monthly reports were 12 to 15 pages with detailed analysis, strategic recommendations, and clear next steps. By month five, they were 4 pages of charts with one paragraph of commentary. Same template every month. The "insights" section started repeating the same observations. "Mobile CPA remains higher than desktop, recommend bid adjustments." We saw that exact sentence in three consecutive reports.
The calls got shorter. Our weekly strategy call was supposed to be 30 minutes. By month four, we were averaging 15 minutes. The account manager would walk through the highlights ("spend is on track, CPA is stable, we adjusted bids on a few keywords"), ask if we had any questions, and wrap up. There was never a sense that they were bringing new ideas or proactively identifying opportunities. It felt like a status update, not a strategy session.
The changes stopped. We started checking the Google Ads change history ourselves because we were curious. In a good month, we would see maybe 30 to 40 changes: some bid adjustments, a handful of negative keywords, an occasional new ad variation. In month eight, there were 12 changes total for the entire month. In month eleven, 9. Our $4,200 was buying a few hours of someone adjusting bids and adding a couple of negative keywords. For $4,200 a month.
The playbook became obvious. Once you have worked with an agency long enough, you start to see the pattern. They have a standard operating procedure that they apply to every client. Review the top spending keywords, adjust bids on the ones that are obviously over- or under-performing, check the search terms report, add some negatives, glance at the ad copy performance, write the report. It is not a bad process. It is just the bare minimum, applied uniformly regardless of what your account actually needs.
We asked our account manager once why they had not tested any new ad copy in three months. The answer was revealing: "Your current ads are performing within acceptable ranges, so we have not prioritized creative testing." Within acceptable ranges. Not excelling. Not improving. Acceptable.
The senior strategist disappeared. During the sales process, we met a senior strategist who impressed us with their knowledge and insight. They were on the first few calls. Then they "transitioned" the account to the day-to-day team. We found out later that this is standard practice: seniors close the deal, juniors do the work. Our actual account manager had about two years of experience. Nice person, hard-working, but not the expert we thought we were paying for.
The tipping point was not frustration with the agency. It was exposure to what autonomous AI could actually do.
A colleague at another company had been using groas for a few months and would not stop talking about it. We were skeptical. We had heard plenty of "AI will replace your agency" pitches that turned out to be glorified rule-based automation with a chatbot slapped on top. But the colleague was not a tech evangelist. He was a skeptic like us, and he showed us his actual numbers.
His account was similar in size to ours. Similar industry. Similar monthly spend. But his cost per acquisition was 38% lower than ours. His ROAS was nearly double. And he was not spending anything close to $4,200 a month on management.
We asked him what had changed. He said the biggest shock was not the performance improvement. It was seeing the volume of optimization activity. He pulled up the change history and showed us what the autonomous system had done in a single week: hundreds of bid adjustments, dozens of negative keywords added, ad copy variations tested and rotated, budget reallocation across campaigns multiple times per day. In a week, it had done more than our agency had done in three months.
That was the moment. Not anger. Just a quiet realization that we were paying $4,200 a month for someone to check our account twice a week and make a handful of changes, while an autonomous system could be doing hundreds of optimizations daily for a fraction of the cost.
We went back and looked at our agency's change history for the past six months. We counted everything: bid changes, keyword additions, negative keyword additions, ad copy changes, budget adjustments, targeting modifications. The total was 247 changes across six months. About 41 per month. Roughly 10 per week.
Ten changes per week. On an account with 400+ keywords across 12 campaigns. That is not management. That is checking a box.
We waited until our contract renewal was up. That was the pragmatic move but also the cowardly one. We could have pushed for an early exit, but we did not want the confrontation, and there was a termination fee that we did not feel like negotiating.
When the renewal came up, we told the agency we were not continuing. The conversation was predictably uncomfortable. They asked what was wrong, and we gave them the honest answer: we had seen what autonomous management could do, and we could not justify paying $4,200 a month for the level of activity we were receiving.
Their response was exactly what you would expect. They said AI could not replace human judgment. They said our account needed strategic oversight. They said autonomous tools were risky and could "blow up" our campaigns. They offered to reduce the retainer to $3,200 if we stayed for another six months.
We almost took it. That is the honest truth. The fear of change is real. What if the AI really did mess things up? What if we lost the institutional knowledge the agency had built? What if we needed to go crawling back? It took about a week of going back and forth before we committed.
Looking back, the hesitation was irrational. We were paying for institutional knowledge that manifested as 10 changes per week. The downside risk of switching was lower than the ongoing cost of staying.
The transition was anticlimactic in the best possible way.
Day 1. We connected groas to our Google Ads account. The process took about 15 minutes. We set our business objectives (target CPA for our lead gen campaigns, target ROAS for our e-commerce campaigns), defined our budget parameters, and added a few brand-level constraints (competitor names we did not want in our ad copy, geographic restrictions, compliance language requirements). That was it.
Days 2 through 7. The system spent the first week doing what it calls a baseline analysis. It crawled through our entire account: campaign structure, keyword performance history, search term data, ad copy performance, audience data, budget allocation patterns, and competitive positioning. We could see this happening in real time on the dashboard. It felt like watching someone do a thorough audit of our account, except it was happening at machine speed.
By day three, it had already started making changes. Small ones at first. Bid adjustments on keywords where the data was clearest. Negative keywords for search terms that were obviously irrelevant (and that, embarrassingly, our agency had been letting burn money for months). Budget shifts between campaigns based on real-time conversion data.
Days 8 through 14. The optimization volume increased. The system started testing new ad copy variations. It restructured one of our ad groups that had too many keywords competing against each other. It identified several keyword opportunities that nobody, not us and not the agency, had ever targeted. Our CPA started dropping. Not dramatically yet, but the trend was visible.
Days 15 through 30. This is when the compounding effect became obvious. The system had enough data from its own changes to start making second-order decisions. "I adjusted bids on these 30 keywords two weeks ago. Here is what happened. Based on the results, I am now doing X." The optimization was building on itself in a way that a human checking in twice a week simply cannot replicate.
We logged into the dashboard every day during that first month because we were nervous. By day 20, we stopped being nervous and started being amazed. The volume of activity was staggering. In 30 days, the system made more optimization changes than our agency had made in the entire 14 months.
Here is what happened in the first 90 days after switching from our agency to groas. These are real numbers from our account.
Cost per acquisition. Dropped from $67 under agency management to $41 after 90 days with groas. That is a 39% reduction. At our volume of roughly 300 conversions per month, that saved us approximately $7,800 per month in wasted ad spend.
ROAS on our e-commerce campaigns. Went from 3.2x under the agency to 5.1x after 90 days. Same products. Same landing pages. Same market. The difference was the speed and granularity of optimization.
Wasted spend. This one hurt the most to look at. In the last full month under agency management, we identified (retroactively) roughly $3,400 in spend on search terms that were clearly irrelevant and should have been negative keywords. In our third month with groas, that number was under $200. The system catches irrelevant searches in real time and blocks them within minutes. The agency caught them during their weekly search term review, if they caught them at all.
Management cost. Went from $4,200 per month (agency retainer) to $299 per month (groas plan). That is a savings of $3,901 per month, or $46,812 per year. And the $299 delivers dramatically more actual optimization activity than the $4,200 ever did.
Time spent. Under the agency, we spent roughly 4 to 6 hours per month on the relationship. Weekly calls, email exchanges, reviewing reports, answering questions, providing feedback. With groas, we spend about 30 minutes per week checking the dashboard and occasionally adjusting a business goal or constraint. That is roughly 2 hours per month, and it is optional. We could spend zero hours and the system would keep running fine.
Total economic impact. In the first 90 days after the switch, we saved $11,703 in management fees (three months of the $3,901 monthly savings). We saved approximately $23,400 in improved CPA across our lead gen campaigns ($7,800 per month in reduced waste). And we gained measurable ROAS improvement on our e-commerce side. The total value of the switch in just the first quarter was over $35,000. In one quarter. On an account spending $20,000 per month.
Over a full year, the math gets even more dramatic. The annualized management savings alone are $46,812. The performance improvements, if they hold (and they have, now six months in), add six figures in value.
The numbers matter. But what I did not expect was how much the emotional weight would lift.
When you have an agency managing your Google Ads, there is a constant low-grade anxiety. Are they actually doing their job? Are we getting what we are paying for? Is the account manager really spending time on our account, or are we number 18 on their priority list? You sit in the weekly call and wonder whether the "updates" are real insights or just rephrased data from the Google Ads interface.
You never quite trust it, but you also do not have the expertise to verify it. So you live in this ambiguous space where you are paying a significant amount of money on faith and hoping it is working out.
That anxiety evaporated overnight when we switched to groas. Not because we stopped paying attention, but because we could see exactly what was happening. Every change, every decision, every optimization, logged in real time on the dashboard. There was no wondering whether someone was doing their job, because the system does not have good days and bad days. It does not get distracted by other clients. It does not take long lunches or call in sick.
The other thing nobody tells you is how liberating it is to not manage the people who manage your ads. No more scheduling calls. No more reviewing reports that tell you what you already know. No more delicate conversations about why performance has been flat for two months. No more worrying about what happens when the account manager leaves the agency and someone new has to "get up to speed" on your account.
We got our time back, our money back, and our peace of mind. The time part was actually the most valuable. Those 4 to 6 hours per month we spent managing the agency relationship now go toward things that actually grow the business.
If we could go back, we would change a few things about how we handled this process.
We would have benchmarked earlier. For 14 months, we accepted "within acceptable ranges" because we did not know what good looked like. We should have compared our CPA, ROAS, and conversion rates against industry benchmarks from the start. The data was available. We just did not look.
We would have checked the change history sooner. The Google Ads change history is a reality check that every advertiser should review monthly. If your agency is making fewer than 50 changes per month on an account with more than 100 keywords, you are not getting active management. You are getting maintenance.
We would not have signed a six-month contract. Lock-in periods benefit the agency, not the client. If the service is good, you will stay without a contract. If it is not good, you should be able to leave. Any agency that requires a six-month or twelve-month commitment is telling you something about how confident they are in their ability to retain you on merit.
We would have run a parallel test before fully switching. We went cold turkey, which worked out fine but caused unnecessary anxiety. The smarter move is to connect the autonomous platform while the agency is still active, compare results for 30 to 60 days, and let the data make the decision. There is zero risk in this approach and it eliminates the fear of the unknown.
We would have done this a year earlier. The $58,800 we spent on agency fees over 14 months is gone. A meaningful chunk of that, probably half, was spent on a service that was not delivering proportional value. If we had made the switch after month six instead of month fourteen, we would have saved roughly $25,000 in fees alone, plus the performance improvements we missed during those additional eight months.
If any of the following sounds familiar, you are probably in the same position we were.
Your agency's reports look the same every month. The charts change, but the commentary could have been copy-pasted from last month. The "recommendations" section has the same three items it has had for six months, none of which have been implemented.
Your weekly calls feel like a formality. The account manager summarizes the numbers (which you could read yourself) and asks if you have any questions. There is no proactive strategy discussion. No one is bringing you ideas.
You have no idea how many hours are actually being spent on your account. You are paying a flat retainer, and the agency has never broken down how that time is allocated. When you ask, the answer is vague.
Performance has been flat. Not declining, not dramatically improving, just flat. The agency explains this with "market conditions" or "seasonal trends" or "we are testing." After several months of testing, nothing has materially changed.
You are managing the manager. You spend more time preparing for your agency call, reviewing their reports, and following up on action items than feels reasonable. The purpose of hiring an agency was to free up your time, and instead it has created a new time commitment.
You have a vague sense that you could be getting better results. You read about other companies with similar budgets getting significantly better CPAs or ROAS, and you wonder why your numbers are stuck. But you do not have a clear comparison point, so you accept "acceptable."
If three or more of those hit home, you already know the answer. The question is whether you act on it or spend another six months hoping things improve.
They will not improve. The agency's structural limitations are not going to change. The number of clients per account manager is not going to decrease. The junior staff running your account are not going to suddenly become senior strategists. The weekly check-in cadence is not going to become continuous optimization.
The alternative is not "hire a better agency." The alternative is to recognize that the agency model itself has been surpassed by technology that does the job better, faster, and cheaper. That is not a criticism of any individual agency or person. It is a recognition that continuous autonomous optimization is structurally superior to periodic human review, in the same way that GPS navigation is structurally superior to pulling over and checking a paper map.
We made the switch. Our only regret is that we did not do it sooner.
The technical transition takes less than an hour. You connect your Google Ads account, set your business objectives, and the system starts working. The performance transition takes 30 to 60 days as the system analyzes your account, establishes baselines, and begins optimizing. During this period, you may see incremental improvements that build over time rather than an immediate dramatic change. By day 30, you should have enough data to see clear trends. By day 90, the performance differential is typically well established.
Everything stays in your Google Ads account. Your campaigns, keywords, ads, conversion data, audience lists, and performance history all live in Google's system, not the agency's. When the agency removes their access, nothing changes in your account. The only thing you lose is any proprietary reporting, strategic documentation, or analysis the agency created outside of Google Ads. Ask for copies of everything before the engagement ends.
That is a personal call, but we did not, and we would make the same choice again. Agencies understandably view autonomous AI as an existential threat and may react by either trying to talk you out of it or by suddenly increasing their activity level to justify the retainer (which raises the question of why that level of activity was not happening before). Run your parallel test quietly and let the data guide the conversation.
Complexity is actually where autonomous systems shine. A complex account with dozens of campaigns, hundreds of keywords, multiple conversion goals, and varied geographic targeting has more optimization levers than any human can monitor simultaneously. The more complex your account, the larger the performance gap between periodic human review and continuous autonomous optimization. That said, if your complexity involves unique business logic, offline conversion data, or custom attribution models, spend 30 minutes during setup making sure those constraints are properly configured.
Yes, but for a narrower set of situations than most people assume. If you need strategic marketing consulting that extends beyond Google Ads, if you are running multi-channel campaigns that require human coordination across platforms, or if you are a brand-new advertiser who needs someone to teach you the fundamentals, a human advisor adds value. But for the core work of Google Ads campaign management, bidding, keywords, ad copy, budget allocation, and ongoing optimization, autonomous AI does it better. The smart play for most businesses is to pair a strategic consultant (for big-picture guidance) with an autonomous platform (for day-to-day execution) rather than paying an agency to do both adequately.
Management fee savings are straightforward: most businesses save 70% to 95% on management costs. If you are paying $2,000 to $5,000 per month for an agency, groas costs $99 to $499 depending on your plan. Performance improvements vary by account but typically range from 25% to 50% improvement in CPA or ROAS over the first 90 days. On a $20,000 per month ad spend, a 35% CPA improvement saves roughly $7,000 per month in more efficient spend. Combined with fee savings, the total first-year economic impact commonly exceeds $50,000 for mid-size accounts.
That the number of changes in your Google Ads change history is the truth that no sales pitch can override. If an agency is making fewer changes per month than an autonomous system makes per day, you are paying for presence, not performance. Check the change history. Count the changes. That number tells you more about the value of your agency relationship than any report or call ever will.
