AI Search Is Here — But Google Still Matters. Plan for Both.

I maintain listening to a model of the identical query from enterprise leaders: Is AI search taking prospects away from Google?

A brand new Similarweb audience-overlap determine gives a helpful reply, with an vital restrict. In September, 93.7% of ChatGPT.com’s guests additionally visited Google.com. The overlap ran the opposite approach, too: 14.1% of Google.com’s guests additionally visited ChatGPT.com, based on Search Engine Watch’s report.

That’s not proof that somebody requested ChatGPT a couple of product, then looked for it on Google. The info counts individuals who visited each web sites throughout the month. It doesn’t present which one they used first, how usually they searched, what they looked for or whether or not they purchased something. It additionally doesn’t seize the total use of both service by way of apps.

These limits matter. However they don’t make the discovering irrelevant. They inform us to not mistake channel overlap for channel substitution.

One buyer, multiple discovery path

For years, entrepreneurs have organized discovery into tidy channels: search, social, referrals, paid media. AI assistants make that division more durable to defend. An individual may ask an AI software to match choices, search a model identify in Google, watch a creator evaluate, then go to an organization’s website straight. The journey might cross a number of programs earlier than the enterprise will get a measurable click on.

The Similarweb determine doesn’t show that sequence. It does present why we shouldn’t assume {that a} customer counted in an AI viewers is a brand new one who has by no means used Google. Treating these teams as separate swimming pools can result in double-counting, inflated attain estimates and unhealthy price range choices.

I noticed a model of this error when search was nonetheless new to many corporations. Some handled search engine marketing as a separate experiment as an alternative of infrastructure for demand that already existed. The chance now could be the reverse: treating AI discovery as a substitute for the search habits prospects nonetheless have.

The higher query isn’t which platform wins. It’s whether or not your model is current, clear and credible when a buyer strikes between them.

Don’t confuse visibility with a sale

The overlap stat additionally exposes a measurement downside. A buyer may encounter a model in an AI reply and later seek for it by identify. That later go to might present up as Google natural visitors, even when the sooner discovery helped create the demand. Or the AI interplay may by no means produce a go to in any respect. The audience-overlap knowledge can’t inform us which occurred.

That’s the reason last-click reporting is a poor referee for this shift. A referral supply can let you know the place a session got here from. It normally can’t let you know which earlier interactions made the model price wanting up.

Google has begun giving website homeowners a clearer view of 1 a part of the image. Its Search Console generative AI efficiency report exhibits impressions for pages showing in Google’s AI search options, together with AI Overviews and AI Mode. That’s helpful visibility knowledge, not proof of incremental visitors or income, and it doesn’t cowl ChatGPT or different AI platforms.

Companies ought to maintain the measures separate. Monitor conventional search clicks, branded searches, certified leads and gross sales. Alongside them, monitor which pages seem in Google’s AI options and whether or not your model is cited in a constant set of related AI prompts. Add a easy query to guide kinds or post-purchase surveys: “The place did you first hear about us?” None of these indicators proves, by itself, that AI brought about a sale. Collectively, they will reveal patterns that last-click studies miss.

Construct one discovery plan, not two channel silos

The response isn’t to separate the advertising and marketing price range evenly between Google and AI. It’s to cease planning as if each platform reaches a unique buyer.

Hold investing within the components of search that assist prospects discover and consider you. Make sure that the pages that designate your merchandise, pricing and worth are present. Perceive which questions carry certified prospects to your website. Then prolong the identical self-discipline to AI discovery: Test whether or not reply engines describe your enterprise precisely, cite helpful sources and embody you when prospects examine choices.

Most significantly, join the groups and dashboards. If search engine marketing studies a decline in clicks whereas AI visibility rises, don’t declare one channel the winner and the opposite the loser. Take a look at branded search, lead high quality, direct visits and conversion developments. Ask prospects what influenced them. Examine the sample over time, not simply throughout one month’s visitors report.

I don’t learn the 93.7% determine as proof that Google is secure or that AI search is innocent to present visitors. It doesn’t reply both query. I learn it as a cause to be extra exact: Prospects can add a brand new technique to uncover a enterprise with out abandoning the outdated one.

Your prospects might meet you in an AI reply and Google you earlier than they act. The businesses that perceive each moments will make higher choices than those nonetheless asking which channel will get credit score for the entire journey.

I maintain listening to a model of the identical query from enterprise leaders: Is AI search taking prospects away from Google?

A brand new Similarweb audience-overlap determine gives a helpful reply, with an vital restrict. In September, 93.7% of ChatGPT.com’s guests additionally visited Google.com. The overlap ran the opposite approach, too: 14.1% of Google.com’s guests additionally visited ChatGPT.com, based on Search Engine Watch’s report.

That’s not proof that somebody requested ChatGPT a couple of product, then looked for it on Google. The info counts individuals who visited each web sites throughout the month. It doesn’t present which one they used first, how usually they searched, what they looked for or whether or not they purchased something. It additionally doesn’t seize the total use of both service by way of apps.