Advertiser Control: Google & Microsoft in 2026

Sarah, marketing director for a fast-growing e-commerce brand selling sustainable home goods, had that familiar knot in her stomach. It was late 2025, and her team was staring at their Q4 ad performance. Revenue was up, sure, but the ROAS from their Google campaigns, especially the ones leaning on Performance Max, was tanking. “It’s like we’re throwing money into a black box,” she’d complained to her agency partner. Their branded search was holding steady, but the automated stuff, particularly on Microsoft Ads, felt like it was just incinerating the budget without any real precision. She missed the days when she could hand-pick every keyword and placement. Now it felt like a constant battle, just reacting to algorithm updates and trying to claw back some control. She wasn’t alone. In 2026, a lot of us seasoned digital marketers are wrestling with this new reality. Google and Microsoft are all-in on AI-driven optimization, pushing automated campaigns that promise to make our lives easier. They promise efficiency, but what many of us get is a frustrating lack of visibility and, too often, a lot of wasted spend.

The Automation Takeover and the Fight for Control

Google Ads’ Performance Max is the perfect example of this dilemma. When it launched in 2021, PMax was sold as the one campaign to rule them all, finding your best ads across Search, Display, YouTube, Gmail, and Discover. The pitch was simple: let Google’s machine learning figure out the best mix of assets and audiences to hit your conversion goals. And look, it can be a beast when it works, especially if you have straightforward conversion goals. The all-in-one nature of it, however, means you give up the direct control over placements and keywords that you had with old-school campaigns. “Our PMax campaigns would start strong, then the performance would just go off a cliff for no reason,” Sarah explained to her agency. “We’d see conversions, but the CPA would sometimes spike, and we had no idea why. Were we showing up on some kid’s YouTube channel? Weird Display sites?” This black box problem has been the biggest complaint about PMax since day one. Google has slowly added more reporting, like placement reports and diagnostic insights, but they’re not the kind of real-time, actionable data Sarah (and the rest of us) needs to make a decision *right now*. So how do you get a handle on it? You have to understand the levers you still have. With Performance Max, that means feeding the algorithm the right signals and setting up some guardrails. You need strong audience signals that tell Google exactly who you’re after, and (this is the big one) you need to use negative keywords at the account level. “You can’t add negative keywords directly to a Performance Max campaign,” her agency told her, “but you can give your Google rep a list of brand-level or junk terms to exclude.” It’s an annoying extra step, but you have to do it. For Sarah’s brand, this meant blocking searches for things like “cheap plastic decor” or “disposable home goods” to stop PMax from chasing irrelevant clicks.

Wrangling Microsoft Advertising’s AI

Microsoft Advertising is on the same path, pushing hard into AI features and automation. It’s a huge opportunity with its reach across the Bing search engine, Microsoft Start, and the Microsoft Audience Network, and the CPC can often be cheaper than what you’d pay on Google. But that automation needs a human watching it closely. “Microsoft’s Smart Campaigns sounded simple, but I was constantly digging around to see where the budget was actually going,” Sarah recalled. Smart Campaigns are designed to make ad creation and management easy by automating bids, budgets, and placements. That’s great for a small business with no time, but for bigger advertisers, the lack of fine-tuned control is a problem. The Audience Network, for example, can drive a ton of impressions, but if you’re not careful with your targeting and exclusions, it’s just a firehose of junk traffic. It’s no surprise that a 2025 report by the IAB on programmatic trends found that nearly 30% of advertisers were worried about brand safety and placement quality in these kinds of automated environments. The most important lever you have on Microsoft Advertising is the aggressive use of exclusion lists for websites and apps on the Audience Network. The interface lets you build shared exclusion lists you can apply to all your campaigns. This simple, proactive step stops your ads from showing up on sites known for garbage content or bad engagement, or that just aren’t right for your audience. “We update our exclusion lists for both Google Display and the Microsoft Audience Network like it’s a religion,” Sarah said. “It’s tedious work, but it saves us thousands in wasted spend every single quarter.”

Your Levers of Control in an Automated World

The platforms want you to think it’s all automated, but you still have a few old-school levers you can pull:

• Granular Geographic Targeting: The algorithms can automate bids by location, but you’re still the one who draws the map. For a local business, getting this wrong is a disaster. That Atlanta bakery needs to make sure its ads are only hitting people within driving distance of the store, not everyone in the entire state of Georgia. Both Google and Microsoft have solid geo-targeting options that let you use radiuses, target specific zip codes, and block out areas you don’t want.
• Budget Allocation and Pacing: Automated bidding might adjust the bids, but the overall budget and how fast it gets spent is still your call. You have to watch the daily spend like a hawk. “I set hard daily budgets and check in every single day,” Sarah stated. “If a campaign blows through its budget by noon with no conversions, I pause it and investigate. You can’t ‘set it and forget it,’ no matter what the platforms imply.”
• Conversion Tracking Accuracy: Your entire automated campaign strategy is built on one thing: accurate conversion tracking. If you’re feeding the algorithm bad or incomplete conversion data, it’s going to make bad decisions. Garbage in, garbage out. Making sure your conversion actions are set up perfectly in Google Tag Manager or with the Microsoft UET is probably the most fundamental control you have. Without clean data, the smartest AI in the world can’t get you good results.
• Creative Asset Management: Performance Max will mash up your images, videos, and headlines to create ads, but the quality of those assets is all on you. Giving the algorithm a good mix of high-quality, relevant creative gives it better material to test and optimize. What happens when you use bad assets? You get bad performance, every time. Automation can’t fix that. “We spend a lot of time and money on good lifestyle imagery and short video clips,” Sarah noted. “It gives the platforms something decent to work with.”

This whole story about losing control to the algorithms is often sold as an inevitable march toward total automation. But those of us in the trenches know it’s really about learning to work a new control panel. The levers look different now, but you can still steer the ship. It just takes a real understanding of how these complex systems work, a commitment to testing everything, and the discipline to feed the algorithms clear, consistent signals. For Sarah, that Q4 slump was a wake-up call. Her team went back to the drawing board. They tightened the audience signals on their PMax campaigns and worked with their Google rep to get a huge list of negative keywords implemented. On Microsoft Ads, they built out massive exclusion lists for the Audience Network and broke out their campaigns for more precise budget control. The payoff? Their ROAS was already climbing back up by Q1 2026. “It’s not about fighting the automation,” she concluded. “It’s about learning to steer it.”