Google Ads completely reset after site changes - anyone fixed this?

sophiasaas

I've been running Google Ads on the same account for over six years - thousands of conversions, solid performance. Then EU regulations forced me to rewrite my site copy and update product pages. First revision? Stable. Second revision, plus a site redesign and pausing campaigns for three days? Total collapse.

Launched a fresh campaign setup with the new pages. Conversions flatlined. Shopping stopped performing. PMAX looks lost - spending on garbage display traffic, impression share tanked, yet CPCs didn't spike. Repeat customers actually convert better now, but new customer acquisition is basically dead.

My theory: Google lost all historical context. The pause, the rebuild, the redesign - seems like it wiped everything it knew about my business. Normally a relearning phase shows some erratic traffic and gradual recovery. This feels different - like the algorithm is stuck in an endless blank state, no traction, no momentum.

Tried a search campaign at max conversions - blew through $50 with three clicks at $15 each. Shopping now only targets exact keywords via negatives but gets <10% impression share. The whole system feels fundamentally broken.

Anyone dealt with a full account reset like this? I'm wondering if it's deeper than just a learning phase - maybe campaign trust, feed relevance, or audience understanding got nuked. Would love any recovery strategies beyond "wait it out."

communityfirst

I've seen this exact pattern before - a legal copy change shakes the whole funnel, and suddenly nothing makes sense. First thing I'd do is split the blame between traffic quality and conversion experience before touching any more levers. Compare pre- and post-change search terms, asset performance, product feed titles, and landing page conversion rates by device. If CTR held steady but CVR tanked, the new page or copy is almost certainly the culprit. If both CTR and query quality dropped, PMAX probably retrained itself into a wrong audience after the shift.

What's worked for me in similar situations is running a controlled Search/Shopping campaign alongside PMAX - exact and phrase match, highest-intent terms, with a clean landing page variant. PMAX is a black box when you're diagnosing, you need something more transparent to see what's really happening. Don't try to fix everything inside PMAX while it's still learning from broken signals.

growthgeek

Yeah, three days of pause and a full rebuild? You've just torched all your learning data. PMAX is now in full scattergun mode, dumping budget onto display trash while it tries to relearn from nothing.

To break out of that blank state you need to drown it in fresh creative signals across every ratio - 1:1, 4:5, 9:16. I've been feeding it dozens of variations generated from a single product shot and a competitor's winning layout. The tool I use reverse-engineers the structure, I drop in clean copy, and it spits out professional assets in minutes. No auto-generated garbage.

Flooding new asset groups with those forces the algo to actually test real creatives instead of sitting in a dark corner burning cash. Fastest way to shock PMAX back to life.

dailygrind

That's a classic Google Merchant Centre trap - and honestly, I've seen it sink PMax campaigns faster than anything else. A landing page content change, especially if you adjusted product titles/descriptions or swapped categories, triggers a full reindex in Google's systems. the machine learning essentially forgets everything it learned about your product taxonomy and starts over.

If you've already been running PMax for a while, the wisest move is:

Pause PMax completely - don't just stop spending, pause the campaign itself.
Spin up a standard Shopping campaign with the exact same feed, no optimisation scripts, no bid adjustments - let it run for at least 7-10 days.
Check your data quality via the Diagnostics tab in Merchant Centre and pull a full feed analysis in Screaming Frog (yes, you can crawl your own product pages to spot mismatches).
Only after that Shopping baseline is stable (consistent impressions, decent CTR) consider relaunching PMax - but feed it the conversion data from the Shopping campaign as a seed.

Throwing PMax back into the mix while Google is still reclassifying your products is like rebuilding an engine while the car is moving. painful.

Also, worth double-checking your category assignments in the feed - I've seen Google auto-map products to entirely wrong taxonomies after a title change. one client's "premium leather wallet" ended up in "Luggage & Bags > Travel Duffels" for three days. Absolute nightmare

ctrjunkie

i've seen this exact pattern before - your campaign's basically got stuck serving ads to people who already bought, so it's stopped finding new buyers. The algorithm just kept optimising for the same pool until it ran dry. What flipped it for us was feeding the system fresher signals from people actively hunting for what you sell right now, rather than relying solely on Google's lookalikes. there's a tool that surfaces real-time purchase intent - injecting that into your targeting turned our new acquisition numbers around pretty quickly.

ctrjunkie

That legal compliance shift sounds brutal when you've got years of data stacked up. Changing all your landing pages and ad copy overnight basically nukes your quality score history and audience signals all at once - no wonder performance tanked even with the same structure.

Something most people don't realise is that Google's algorithm remembers historical performance per URL and ad text combination. When you swap those out completely, it's like starting fresh but with an account that still has old conversion data confusing the signals.

What's actually been helping in similar situations I've seen: keep your best performing ad groups running exactly as-is on separate campaigns as control groups while you test the compliant versions in new campaigns. That way you preserve some stable signal while figuring out the new baseline. Also worth checking if your conversion tracking broke during the site overhaul - sometimes EU consent changes mess with GTM firing and you're actually getting more conversions than you're seeing reported.

How much of your traffic and conversions were coming from brand search before the changes versus generic/category terms? That usually tells you where the signal loss is hitting hardest.

ranktracker

Need more specifics on that 'strange issue'? I've seen accounts go weird after switching bidding or attribution models. First thing I'd do is pull a change history report from the last 3-5 days, then sanity-check conversion tracking setup - it's almost always one broken setting or a measurement glitch under the hood.

24fire

You can optimise for a week and still miss the core problem. A fresh set of eyes from someone who has built campaigns through multiple economic cycles can spot the leak in ten minutes. The cost of an expert is usually less than the ad spend you are burning right now.

sophiasaas

Honestly, the system feels completely broken once it loses optimisation. I had the same thing happen - Pmax burned through everything on display traffic, like it forgot it was supposed to convert. Then switched to a search campaign with max conversions, and it spent the whole $50 budget on three clicks at nearly $15 a pop. Insanity. My shopping campaign now only targets exact keywords via negatives, but it's still sitting at under 10% impression share. The algorithm just spirals once it learns the wrong behaviour. No amount of negative keyword lists or bid adjustments seemed to pull it back. Whole thing's flawed.

sophiasaas

honestly, I've been there. when Google's signal completely contradicts years of accumulated data, it feels like the algorithm forgot everything it ever learned. Starting fresh with a new pixel might actually be the cleanest path - especially if your account is riddled with legacy noise from old attribution models or half-baked experiments.

For SaaS, we see this a lot when switching from manual tracking to server-side. The existing pixel carries so much historical baggage that it actively fights new, cleaner data. A reset can be painful in the short term (hello, learning phase), but it often fixes the LTV:CAC ratio faster than trying to retrain a broken model. Just make sure your onboarding flows are tight before you flip the switch - you don't want new pixel fires while your trial users are bouncing.

liamcopy

Yeah, that website tweak might've just nuked your tracking.

One wrong line of code and PMax has nothing to optimise for. It's flying blind.

Been there. rebuild with regular search and shopping campaigns. Use your top performers from history - products and search terms that actually worked. Make sure your ad copy matches the landing page exactly. No mismatches.

Also check your Lighthouse scores. If you dropped some unoptimised images in there, your site's crawling now. That kills conversion rates.

and don't overlook the cookie consent banner. Make sure you're running advanced consent mode V2 - or at least proper basic mode with ad_storage granted. Without that, your pixel's half dead.

sophiasaas

exactly!!! That cut our churn - thanks!!!

sophiasaas

Thanks for the tip. Is that tool inside Google Ads manager itself, or is it a separate integration? I've been digging into account-level diagnostics lately for our B2B SaaS funnel, and anything that helps spot wasted spend early would be a lifesaver.

sophiasaas

Would you say starting with a fresh pixel could give it a clean start and help things grow back?

Honestly, I've seen mixed results with that approach. Pixel decay and stale data can muddy the waters, especially if the original pixel has been flooded with garbage events from broken third-party integrations. Clean slate might reset the learning phase, but you lose whatever lookalike data you had.

In my SaaS funnel experiments, a fresh pixel actually performed worse for the first couple of weeks because our retargeting pool vanished. The trade-off felt brutal. I'd only recommend it if your current pixel is seriously f*cked - like pixel firing on every page load or duplicate events clobbering the algorithm.

What's your LTV:CAC ratio right now? That usually tells me if the pixel needs a wipe or just a few exclusions cleaned up.

sophiasaas

honestly, took me three agency swaps before i found one that actually understood product-led onboarding. what worked for me was asking for case studies on churn reduction, not just vanity metrics. If you're after a reliable growth partner, check their LTV:CAC ratio approach first-if they can't talk intelligently about that, move on

justauser

Google's automation keeps pushing spend into the worst placements and calling it optimization. I got tired of babysitting campaigns just to catch budget bleeding into display or rogue CPC spikes.

I use ChadAd/s now to flag that stuff before it burns through a day's spend.

sophiasaas

Great advice from both of you - thanks! So you'd recommend reviving the existing pixel rather than starting fresh? makes sense from a retention standpoint

24fire

This campaign sounds like it's running on hope rather than a real strategy - I've been there, staring at a dashboard full of red wondering what broke.

Before throwing more budget or creative at it, take a step back and map the funnel drop-off. Most of the time the problem isn't the ad itself - it's the disconnect between the signal you bought and the landing page experience.

happy to dig into the data if you want another set of eyes. DM me the account structure and any conversion metrics you've got.

sophiasaas

Some solid takeaways here. been digging into similar tactics to tighten our onboarding flow - shaved nearly 15% off our early churn just by reordering the activation steps. LTV:CAC moved from _3x to 4.2x over two months. Appreciate the share, always good to compare notes on what's actually moving the needle