Hey Reader!
A retailer came to us recently because they were seeing weird results in Google Ads and Google Merchant Center. They knew the ad setup was good, so why was the data wrong?
We started in the usual places: broken tags, consent settings, recent code changes. It was all fine. Then we turned to the product feed. The site sends one version of the SKUs to Google Ads, but it turns out Merchant Center doesn't get its products from the site at all. It's fed by a third-party feed tool that uploads a scheduled file using manufacturer part numbers stored in a Google Sheet, all set up by a former freelancer. At some point the two stopped matching (my best guess is a platform migration a while back), so we can't reliably connect a sale back to the product it advertised.
When we uncovered this, we found that nobody currently working there knows why the feed tool exists. Maybe there was a very good reason. Maybe someone needed a workaround in a hurry and promised to replace it later. (And we know how that usually turns out!)
The easiest fix is to switch the Merchant Center feed over to the platform's built-in integration and move on. But with the old feed still running, that could cause major product conflicts, and might even list every product twice. So the first fix here isn't technical. It's a conversation with the client about how their systems talk to each other, and what that feed was supposed to solve. Not a quick fix, unfortunately, but it's the one that'll get this sorted for the long term.
I see a lot of analytics setups built from a stack of reasonable decisions made by people who have since moved on, and nobody wrote down why things were set up the way they were before they left. Next time you're working with a new client or an unfamiliar configuration, find out who built the plumbing behind it and what they were trying to do at the time, before you change anything.
How I'm Removing Bot Traffic from GA4
An Edmonton Realtor client of ours suddenly had a lot of "visitors" from China and Singapore, hitting every page on the site two to ten times per session. GA4's built-in bot filtering didn't catch them, and because they view so many pages, they count as engaged sessions too. So I tested 12 different bot checks to see which ones held up, and decided on one method that routes suspected bots to a separate GA4 property. Your production data stays clean, and nothing gets thrown away.
The code is in the post, along with what the data showed and where you might want to send the traffic you're not sure about.
How I'm Removing Bot Traffic from GA4 →
Articles Worth Your Time
———•
|
California SB 690 and the end of (some) CIPA demand letters
If you've had a client get a $25,000 demand letter because marketing tags fired before consent, this one's for you. Jason Packer runs through California's SB 690, signed September 30, which removes the private right to sue over website tracking under CIPA's pen register and trap-and-trace section. As Jason says, this isn't clarity on how consent works in the US, and the demand letters may just shift to other laws. (Not legal advice, so loop in your lawyers before acting on any of this!)
Poisoned Google SERPs
Derek Perkins at Nozzle noticed that, since mid-September, a big chunk of the Google results their tracker collects have been "poisoned", with the usual sites swapped out for Facebook, Reddit, Quora and Etsy, while ads and Shopping stayed the same. Page 2, "near me" searches and local categories like legal and home services were hit hardest, and Canada had a higher rate than the US. If a client's rankings suddenly fell off a cliff in your rank tracker, this might be why.
The GA4 User Lifetime report: should you trust it?
Jude Nwachukwu Onyejekwe skips the how-to and goes straight to the question that matters: is your setup good enough for this report to mean anything? He walks through eight checks, covering purchase tracking, User ID, data retention and sampling, plus a few surprises. Run through this list before you hand anyone a lifetime value number.
Where You Can Find Me
———•
|
Inside Google Ads Live: Toronto 2026
Thursday, October 29 | Maple Leaf Cinema at the base of the CN Tower | Toronto, ON
Crossing the Valley: Getting AI Out of the Demo and Into Your Day
This Thursday! Jyll has put together a full day of tactical sessions, with Q&A in every one, plus a reception at the end and access to the CN Tower observation deck. Tickets are limited, so grab one soon if you'd like to come. I'd love to see you there!
That's it for this edition of The Huddle. As always, if you have questions or want to share what you're working on, just hit reply!