Google Is Eating Your Traffic. What an E-Commerce Owner Should Actually Do

Last month I opened Google Analytics for three e-commerce projects I work with regularly. Google organic traffic: down 23% year over year. Not one project — all three. Different niches, different catalog sizes, different site ages. Same pattern.

This isn't a technical problem. It's not a penalty for anything. It's Google's product working exactly as designed.

I'm not an SEO specialist and I won't pretend I know how to beat AI Overviews. What I do know is how to build systems that don't break when someone else's algorithm changes. That's a different skill — and it's what I started advising clients on.

This isn't a bug. It's the product.

AI Overviews answer the query on the search results page. The user gets what they need without clicking through to your catalog. For informational queries — "which laptop should I buy," "what SPF means" — this works beautifully for the user.

For e-commerce, it means top search rankings convert to visits less and less. The click still exists, but the AI block captures it first. Your blue link moved further down.

Google built something valuable for users by taking value away from your business. The system is working correctly — your interests and Google's interests just don't align the way they did ten years ago.

Waiting for this to reverse is a strategic mistake.

Three kinds of traffic. Only one of them is yours.

I think about traffic in three categories — not by channel, but by how much control you actually have.

Borrowed traffic: Google organic, Instagram, Facebook, TikTok. You're operating on a platform you don't own. When the algorithm changes, your business changes with it. This has always been true; it's just more visible now.

Rented traffic: Google Ads, paid social, marketplaces. You pay for every contact. Predictable, but increasingly expensive.

Owned traffic: email list, direct visits from repeat customers, internal site search, push notifications. This doesn't depend on Google. It depends on the relationship you've built with your buyers.

Borrowed traffic is disappearing. Rented traffic is getting more expensive. Owned traffic stays.

When I look at an e-commerce project now, one of my first questions is: "What share of your traffic is owned?" The typical answer is under 15%. That means 85% of the business rests on algorithms you don't control.

Your internal search is a conversion machine you're ignoring

The search bar on your site is an owned tool that most stores underuse.

I worked on a catalog with 28,000 SKUs. When we set up proper search analytics, we found 340 weekly queries returning zero results. Buyers were searching — the site was silent. That's not an SEO problem. It's a retention problem with traffic that already came to you.

Your search logs are a list of buyer intentions. Not potential buyers, not a focus group — real people who opened your site and typed something into the search bar. That data doesn't exist in your Google Analytics, your ad accounts, or your competitors' dashboards.

When we set up Elasticsearch and proper faceted filters, search conversion improved 18% — with no new external visitors. All of it came from people already on the site.

More on how search logs become a product signal: zero results as a product backlog.

Google can't see what happens inside your search. That's your territory.

An email list beats 10,000 Google positions

I have a client who spent years building SEO. Good rankings, steady traffic. This year, organic is down 30%. Email performance: unchanged.

Email works differently. You send a message — your subscriber sees it. No algorithm between you. No auction. A direct relationship.

10,000 email subscribers who buy regularly are worth more than 10,000 Google positions. The positions depend on an algorithm you don't control. Subscribers depend on value you create.

If your entire traffic acquisition strategy depends on Google, that's not diversification — it's a single point of failure dressed up as a marketing plan.

I wrote about this more in the case for owned audience. Since then the trend has only sharpened.

What to do in the next 30 days

Three things that work without rewriting your site.

Turn on internal search analytics. If you're using Google Analytics 4, there's a built-in report for on-site queries. Enable it if you haven't. Look at what people are searching for right now. How many of those queries return zero results?

Capture email where intent is already high. Not a banner on the homepage — a form on out-of-stock product pages ("notify me when available"). A checkout opt-in. Someone who made it that far already told you they want to buy.

Change your reporting metric. Add owned traffic share to your regular reports: not just "how much came from Google," but "how much came direct, from email, from bookmarks." If that number isn't growing, you're optimizing an asset you don't own.

What not to do

Don't write more content to get cited in AI Overviews. That strategy works in Google's favor, not yours.

When AI Overviews cite your page, that's fine. But it doesn't convert to a click. The user read the answer and moved on. You helped Google get better without getting anything back.

Optimizing for AI citations means becoming a content supplier for a platform that's redirecting your buyers.

Your goal isn't to appear in an AI answer. It's to get the buyer to your site, so they buy from you.

What I've learned this year

Google became less predictable as a traffic source. That's the fact. But the e-commerce owners I work with were already building businesses on traffic they didn't own. AI Overviews just made that dependency visible.

I don't know what Google looks like in a year. What I do know: the email list you build today works regardless. The internal search you improve this month gives you buyer data that no ad platform has. Your repeat customers' direct visits don't depend on any algorithm.

SEO isn't going away. It's just a bad idea to bet the whole business on it.