With Google, or for Google?
Many agencies proudly display the Google Partner badge or the number of campaigns they run on Meta. It's a good calling card, though it leaves a question open: when an agency says it works with Google, what does that mean for its client?
The question matters more than it seems. To keep that badge, Google requires an optimization score of at least 70% and at least USD 10,000 in ad spend every 90 days across all the accounts the agency manages (Google). That score goes up as the platform's own recommendations get applied.
A partner that also has to make money
Google and Meta live off advertising. In the third quarter of 2025, Alphabet made USD 74.2 billion from ads, 72% of its total revenue. At Meta, advertising accounts for almost 98% of revenue (Storyboard18).
That doesn't make them villains. It means they have their own interests, and it's worth keeping that in mind every time the platform suggests something.
In August 2024, a US federal judge ruled that Google had illegally monopolized the search market. In the ruling, Judge Amit Mehta noted that the company raised text ad prices through auction tweaks that were "barely perceptible and rarely announced," and that it did so "without fear of losing advertisers" (AdExchanger).
That's where the paradox shows up. The same company that recommends how much you should bid is the one setting the price of the auction.
What the optimization score measures
Google defines it as "an estimate of how well your Google Ads account is set to perform." It's calculated partly from pending recommendations, and an account can reach 100% by applying or dismissing all of them (Google). In practice, it measures how much attention you've paid to the platform's suggestions.
Many of those suggestions point the same way: raise the budget, broaden keyword match types, or let automated bidding decide. Some improve a campaign. Others just make it spend more.
Dismissing a suggestion also raises the score.
In 2024, Optmyzr analyzed 17,380 Google Ads accounts. Those scoring 90 to 100 showed better returns than those below 70. At first glance that looks like a point in favor of following recommendations, until you remember how the score is reached: dismissing a suggestion raises it too. An account at 100% may well be one where someone reviewed every recommendation and rejected the ones that didn't help. Between 70 and 90, the difference was almost zero, and accounts below 70 had a better conversion rate. The analysts themselves warn that a high score, on its own, doesn't explain better results (Optmyzr).
Every recommendation deserves a question before you apply it: does this serve the client's business, or the volume of the auction?
Lots of clicks, few customers
A bad agency spreads ads everywhere. It turns on the Display Network, search partners and every new format that shows up in the account. The report fills up with impressions and clicks, and the client is still waiting for the phone to ring.
What makes a campaign profitable is where every dollar goes.
Search partners are a good example. They're sites outside Google that show your ads, and in many accounts they're switched on by default. In one account analyzed by Adalysis, cost per conversion was USD 185 on those sites and USD 11 on Google Search: almost 17 times higher. In another account from the same analysis, partners performed better (Adalysis). No general rule replaces looking at each account's numbers.
If spending more were enough, any company with a budget would be guaranteed customers. But what defines a campaign isn't just its budget, or the number of clicks.
What we've seen in our own accounts
We tested it with our own advertising. Many times, going against a Google Ads recommendation worked better than following it.
Google suggests a cost per click, and a lower bid ends up being more profitable. Other times, it proposes a target cost per acquisition that's beyond the client's budget, and the campaign performs better with a different approach. In our own account, we turned off search partners and skipped most recommendations. That brought in qualified leads and saved a good share of the budget that was going to low-quality clicks.
Going against a Google Ads recommendation often worked better than following it.
The scenarios are endless, which is why no automated recommendation fits every account the same way.
The agencies that will last
The future of paid advertising will be decided in this tension. Platforms need to show results so advertisers keep investing, and at the same time they're companies that have to hit their numbers every quarter.
Automation does a big part of the work very well, and many recommendations are worth taking. What makes a campaign succeed is combining those tools with the judgment of whoever runs the account: knowing the business inside out and being willing to say no when a suggestion inflates spend or brings in low-quality clicks.
The agencies still standing will be the ones that work for their client, with Google and Meta as tools.