What Changed in My Traffic the Month AI Overviews Expanded
I run content for a DTC brand doing about $3M a year, and I watched our blog traffic drop 34% in six weeks when Google expanded AI Overviews into our category. Not a rumor. Not a LinkedIn post. Our GA4 dashboard. This post is what I saw and what we did about it, because if you're searching "AI Overviews killing organic traffic what to do," you don't need theory, you need a plan.
What the problem actually looks like
It doesn't hit all at once. It hits page by page, and it hits your best pages first.
Here's the pattern we saw:
- Pages ranking positions 1-3 for informational queries lost 40-60% of clicks even though rankings stayed the same or improved.
- Impressions in Search Console held steady or went up. Clicks fell. That gap is the tell.
- "What is," "how to," and "best way to" queries got hit hardest. Transactional and brand queries barely moved.
- Average position looked fine. CTR cratered. One page went from 8.2% CTR to 3.1% CTR in a month with zero ranking change.
If you're only checking rankings, you'll miss this entirely. You have to check CTR by query and by page. That's where it shows up first.
Why it happens
AI Overviews answer the question directly in the search results page. If someone searches "how long does collagen take to work," Google now writes a 3-4 sentence answer pulled from multiple sources, sometimes citing you, sometimes not. Either way, the user got their answer without clicking anything.
The mechanism is simple: Google used to be a directory. Now it's increasingly an answer engine that happens to have a directory attached. For queries with a single factual answer, a definition, a comparison, a quick how-to, there's no reason to click through anymore. The value you built by writing a clear, well-structured answer is exactly what got you scraped into the summary.
This hits informational content the hardest because that's where "one clear answer" content lives. It hits commercial and transactional content less because AI Overviews can't complete a purchase, compare fit and price in real time, or replace a product page. Google still needs to send that traffic somewhere.
What we do about it
We didn't panic and we didn't rewrite everything. Here's the actual sequence:
- Segment your content by intent first. Pull every URL with organic clicks, sort by query type. Pure informational, no-purchase-intent pages are your triage list. On the DTC brand, that was about 40% of our blog.
- Accept the loss on pure-definition content. If a page exists only to answer "what is X," it's not coming back to old numbers. We stopped investing new hours there. Existing pages stay up for brand signal and backlinks, nothing more.
- Push commercial intent harder. We shifted our content calendar from 60/40 informational-to-commercial to 30/70. More comparison pages, more "X vs Y," more "best for [specific use case]" content. These still get clicks because AI Overviews summarize but rarely settle a buying decision.
- Restructure remaining informational content to require a click. Instead of "collagen takes 8-12 weeks to show results," we now write the summary answer plus a dosage calculator, a personalized quiz result, or a downloadable tracker. AI Overviews can't replicate an interactive tool.
- Double down on channels Google doesn't own. We moved budget into email capture and a small paid social retargeting loop, since organic informational traffic used to be a huge top-of-funnel input. We replaced roughly 15% of that volume with paid social at a blended CPC of $0.78, funneling into a lead magnet instead of a blog post.
- Track CTR by query monthly, not rankings. This is the actual leading indicator now. We built a simple Looker Studio view comparing impressions to clicks by query cluster, refreshed every 30 days.
This isn't free. It cost us about 25 hours a month in content restructuring and roughly $2,400 a month in incremental paid social to backfill the funnel gap. That's the real cost of adapting, and it's cheaper than pretending nothing changed.
What it costs to ignore
On the infrastructure project I marketed, a $2.2B build with a public information mandate, we couldn't just walk away from informational content, it was regulatory. So I've seen both sides: what happens when you can adapt and what happens when you can't.
For a business that can adapt and doesn't, the cost compounds. That 34% drop in blog traffic, if left alone, translates to roughly 8,000 fewer monthly sessions for us, and at our historical 1.8% email signup rate from blog traffic, that's about 144 fewer email subscribers a month. At an average customer value of $85 and a 4% subscriber-to-purchase rate over 90 days, that's close to $12,000 a month in downstream revenue you stop seeing, and you often don't connect it back to the traffic drop because it shows up two steps removed in the funnel. Ignore it for six months and you're not just down traffic, you've lost the compounding effect of that traffic building your email list, retargeting pool, and brand search volume.
Most common mistake
The most common mistake is treating this as an SEO problem to be fixed with better SEO. Writing "more helpful" content, adding more headers, chasing featured snippet formatting harder. That doesn't work because the issue isn't quality, it's that the answer got extracted before the click happened. The fix is a funnel fix, not a content-quality fix. Change what you're optimizing content to do, not just how well it's written.
FAQ
Will AI Overviews traffic loss reverse itself?
No. Google has expanded AI Overviews steadily since May 2024 and shows no sign of pulling back. Plan for this as permanent, not a temporary dip.
Should I stop writing informational content entirely?
No, but stop expecting it to drive clicks on its own. Use it for brand authority, internal linking, and backlink attraction, and put your real growth effort into commercial and comparison content that AI Overviews can't fully answer.
The practical takeaway: check CTR by query cluster this week, not rankings. If you see impressions flat and clicks down on informational pages, you're living this right now. Shift your next content push toward commercial intent and build one thing on your site an AI summary can't replicate. That's the whole playbook.