Something has changed in paid search this week that's easy to miss and matters more than it looks. Google has adjusted how automated bidding behaves when your budget runs out. Meanwhile, the AI side of search is still growing faster than anyone's ability to measure it. Here's what's happening and what it means for your online shop.
Google Changes How Budget-Limited Campaigns Chase Your Targets
This is the story I'd put at the top of your list this week. Google Ads has changed how campaigns optimise to Target CPA (the cost you're willing to pay for each sale) and Target ROAS (the return you want from every pound you spend) when those campaigns are held back by their budget.
Why does that matter? Plenty of e-commerce accounts run with budgets that can't keep up with the demand available. You set a target ROAS on your Shopping or Performance Max campaign, but the daily budget caps out by mid-afternoon. How Google balances "hit the target" against "spend the budget" in that situation decides which clicks you buy, what you pay for them and what kind of returns you see.
If you've set targets and budgets and then left them alone, this is a good moment to check them. A change to the logic underneath means the settings that worked in July may not behave the same way now. It's worth looking at:
- Which of your campaigns are regularly marked "limited by budget"
- Whether your cost per sale or return on ad spend has moved since the change
- Whether your targets still reflect the margins you actually need
Going forward, budget and target settings can't be treated as two separate decisions. They work together, and Google has just changed how they interact. I'll be paying close attention to how this settles in over the next few weeks.
Amazon Is Now Selling Ads Inside ChatGPT
This one would have sounded far-fetched a couple of years ago. Amazon is piloting a managed ad service for placements within ChatGPT. It's a US-only pilot for now, so nobody in the UK needs to rush anything, but the direction is clear.
Think about what it means. People are already using AI assistants to ask things like "what's the best waterproof walking jacket under £150?" Until recently, those conversations sat mostly outside the advertising world. Now one of the biggest retail advertising businesses on the planet wants a presence in them.
If you sell products online, this is worth watching for two reasons. First, product discovery keeps moving away from the traditional search results page. Second, the platforms that already hold rich product and shopping data are the ones best placed to monetise AI conversations. Your product data, meaning clear titles, accurate prices and proper descriptions, will matter more wherever your customers end up asking their questions.
Google Admits Its AI Search Reporting Isn't Good Enough
Here's a refreshingly honest admission. Google has acknowledged that Search Console's reporting on AI search doesn't properly reflect how and where sites appear in AI-driven results.
Search Console is the free tool that shows how your website performs in Google's organic search. Many shop owners use it to see which searches bring visitors in. If AI features are taking a growing share of those searches and the reporting can't show you clearly what's happening, you're partly flying blind.
What does that mean for you in practice? Don't read too much into sudden rises or falls in organic figures that seem linked to AI search. The data may simply not be telling you the full story. It's also a reminder of why paid search remains so valuable. With Google Ads, you control where you appear and you can measure what you get back far more directly.
Last Click or Data-Driven: Who Gets the Credit for a Sale?
This one isn't breaking news, but it's a useful refresher that ties neatly into the bidding story above. Attribution is how Google Ads decides which ad or click gets credit for a sale.
With last click, all the credit goes to the final ad someone clicked before buying. With data-driven attribution, Google spreads credit across the different touchpoints along the way. Someone might find you through a generic Shopping ad, come back later through a brand search and then buy. Data-driven attribution recognises that the first ad played its part.
Why should an e-commerce owner care? Because automated bidding uses this information to decide where your money goes. If only the last click counts, the campaigns that introduce new customers to your shop can look like poor performers and get starved of budget. Your brand campaigns take the credit instead. With target bidding strategies now behaving differently, the numbers you feed Google have never been more important.
One word of caution, though. Attribution is only as good as the conversion tracking behind it. If your account is counting things that aren't real sales, no attribution model will fix that.
Plugging AI Tools Directly Into Google Ads Data
More and more PPC teams are using AI assistants like ChatGPT and Claude, but there's a catch. According to the research cited, 70% of people said their biggest concern was the quality of the answers. That makes sense, because an AI tool doesn't save you any time if you have to check and correct everything it says.
The fix gaining ground is something called MCP (Model Context Protocol). Put simply, it's a way of connecting AI tools directly to live data sources, such as your Google Ads account, so they're working from real numbers rather than guesses or whatever you've pasted in.
This is something I'm genuinely interested in. The value of AI in account management doesn't come from asking a chatbot generic questions. It comes from giving it the right context so it can spot patterns, flag problems and speed up analysis. There's a big "but", though. Better data in doesn't replace judgement. An AI tool can tell you a campaign's return has dropped. It takes someone who understands your business, your margins and your stock position to decide what to do about it.
What I'm Watching
There's a clear thread running through this week. The machines are doing more of the work, from bidding and attribution to ads inside AI conversations and even the analysis itself. The measurement tools are struggling to keep up, though.
For your online shop, that makes the fundamentals more important. Make sure your conversion tracking is recording real sales. Check that your budgets and targets still work together after Google's change. Keep your product data sharp, because it's what every platform relies on, whether that's Google Shopping or an AI assistant answering a customer's question.
Over the coming weeks, I'll be watching how budget-limited campaigns respond to the new bidding behaviour. I'll also be keeping an eye on whether Amazon's ChatGPT pilot moves beyond the US. And I'll be watching whether Google follows its admission with better AI search reporting. If any of this raises questions about your own account, get in touch. That's exactly what I'm here for.