Magoom

ChatGPT referral traffic to brand homepages just jumped 150%.

ChatGPT changed how it surfaces links in May. Brands suddenly have new visitors. The catch: 60% of that traffic is landing on the homepage. Time on site is up. Pages per visit are up. The first impression now matters more than it has in years.

Most B2B homepages were not built for this audience.

-> A homepage built for paid search lands with a feature list and a CTA. AI-referred visitors come with the question, not the intent.
-> Good design used to stand out. Now it’s expected.
-> AI-referred visitors arrived because they trusted the LLM. They're benchmarking your homepage against that trust, not against your last campaign.

Audit your homepage like a new visitor arriving from a sentence in ChatGPT. Most will not survive the comparison.

Why does your retargeting still treat every site visitor like they forgot you?

Because the entire B2B retargeting playbook was built on a 2010 assumption: visitors leave because they got distracted, so remind them and they'll come back.

B2B visitors almost never leave because they forgot. They leave because they weren't ready, didn't trust the price, weren't the decision-maker, or didn't believe the value yet.

A pricing page bounce and a blog-post bounce are not the same person. Treating them with the same ad is the most expensive shortcut in the funnel.

Segment retargeting by reason for the exit. The pricing-page visitor gets the case study. The blog reader gets a different entry point.

Domain authority does not predict whether AI cites you.

Aleyda Solis & AirOps did a study of 350,000 pages across nearly 17,000 ChatGPT queries and found something uncomfortable for SEOs: pages with high DA were not the ones getting cited. Pages with lower DA often were.

What did predict citation: being in retrieval position 1 to 3, and a headline that directly answers the question being asked.

Stop optimising for a metric that no longer maps to the visibility you actually need.

"AI on top of our CRM" only works if the CRM is clean.

Most aren't.

Our client Janis Balodis from eXpanby, who work with CRM systems every day, point out that 44% of companies lose more than 10% of annual revenue due to inaccurate CRM data and messy workflows.

That's before any AI starts summarising the pipeline.

-> Audit field-completion rates per pipeline stage.
-> Kill custom fields nobody filled in for two quarters.
-> Clean duplicate contacts this week, not next quarter.

Garbage in, AI-summarised garbage out.