Google Ads Is Making More Decisions Than You Think

What advertisers have already lost control of

Google Ads has not become impossible to manage. But it has become much easier to mistake visibility for control.

Think of an automated campaign like a thermostat. You set the broad parameters: budget, target, conversion goal and geographic focus. Google then makes thousands of decisions around those inputs, choosing where ads appear, which searches they match to, which creative variation is shown and, increasingly, which landing page a user sees.

For many advertisers, the problem is not that automation exists. It is that the line between what they can see and what they can actively influence has moved much further than they realise.

Google has responded to long-standing concerns about transparency. Performance Max now offers channel performance reporting. Campaign-level negative keywords and brand exclusions provide more protection around where ads can show. AI Max gives advertisers controls around text customisation, Final URL Expansion, locations and content guidelines.

These are useful improvements.

But they do not return decision-making to the advertiser. They make the system easier to observe.

And observation is not the same as control.

What you can see but cannot directly change

Performance Max channel reporting can show whether spend and conversions are coming from Search, YouTube, Display, Discover, Gmail or Maps.

That is valuable information. But it does not mean you can choose how the budget is distributed between those channels.

You may see that a large proportion of spend is being absorbed outside Search. You may believe that the traffic is weaker, less commercially valuable or less incremental. But there is no direct lever to say: move more budget into Search and less into Display.

Google’s system is designed to optimise toward the campaign’s overall goal across its available inventory. You can influence the likelihood of certain channels performing through assets, product feeds, audience signals and campaign structure, but you are not setting channel-level budget allocations.

The same principle applies elsewhere.

You can review which assets are performing well, but you cannot reliably prescribe which message should appear in every auction or placement.

You can review the landing pages reached through Final URL Expansion, but you are still allowing the system to decide which page it believes is most relevant for a given search.

You can add exclusions, negative keywords and content controls, but these define the boundaries of the system rather than directing every decision inside it.

The controls are real. They are just mostly campaign-level controls, not auction-level controls.

The risk is not bad automation. It is badly defined success.

Google’s automation is sophisticated. Smart Bidding can assess signals and react at a scale no human team could match manually.

The challenge is that it can only optimise toward the outcome it has been given.

An algorithm trained on raw conversion volume will find the most available conversions. That may include existing customers searching for your brand, returning visitors who were already close to purchasing or low-friction leads that look strong in-platform but do not translate into meaningful commercial value.

The account can appear healthy.

Conversion volume is up. Cost per acquisition is stable. Return on ad spend looks acceptable.

Yet the business may see little improvement in new-customer growth, profit, lead quality or revenue that would not have arrived anyway.

That is not necessarily an automation failure. It is often a measurement and objective-setting failure.

The algorithm is doing what it was asked to do. The issue is that the conversion signal did not properly represent the outcome the business actually cares about.

AI Max makes the gap more important

AI Max brings several automated decisions closer together inside Search campaigns.

It can expand query matching beyond traditional keyword coverage, customise ad text and use Final URL Expansion to select a more relevant landing page. Individually, each feature can be useful. Together, they create a more autonomous campaign model than many advertisers are used to managing.

This can improve reach and relevance. It can also make weak inputs more consequential.

If your conversion tracking rewards low-quality leads, AI Max has more scope to find them.

If your website contains pages that are technically relevant but commercially weak, Final URL Expansion may send users there.

If your exclusions are out of date, query expansion can surface waste faster than a manual campaign structure would have done.

More automation does not remove the need for control. It increases the importance of defining the right guardrails.

The work has moved upstream

The strongest Google Ads accounts are not trying to recreate the manual control of ten years ago. That is no longer realistic, and in many cases it would be less effective.

They are more deliberate about the inputs that shape automated decisions.

That means:

• Using conversion signals that reflect revenue, profit, qualified leads or customer value rather than simple transaction volume.
• Feeding CRM and offline conversion data back into the platform where possible.
• Reviewing search terms and exclusions consistently, not only when performance drops.
• Applying brand exclusions with a clear commercial rationale.
• Checking where automated landing-page selection is sending users.
• Separating acquisition, retention and brand demand where the business needs different outcomes from each.
• Treating campaign reporting as a prompt for investigation, not proof that performance is commercially sound.

This is where paid media management is changing.

The day-to-day role is less about manually choosing every keyword, bid and placement. It is more about making sure the platform is learning from the right version of business reality.

Visibility is useful. Control is earned.

Google Ads is giving advertisers more information about what automated campaigns are doing. That is a positive development.

But knowing where spend went is not the same as deciding where it goes next.

The advertisers who perform best will not be the ones trying to fight every automated decision. They will be the ones who understand which decisions Google now owns, which levers still matter and what data the system needs to make commercially useful choices.

You still own the thermostat.

But you need to be much more deliberate about what temperature you set, what data informs it and whether the room is actually warming in the way your business needs.