Measure AI marketing ROI, including the traffic GA4 is hiding.

Measure AI marketing ROI by attributing pipeline and revenue to four channels: traditional organic, AI Assistants (ChatGPT, Perplexity, Claude, Gemini, Copilot), direct/branded, and paid. Set up custom channel groups in GA4 so AI referral traffic stops landing in the Direct bucket. Then compare assisted and last-touch conversion rates across the four cohorts on a 90-day window.

GA4's default channel groupings do not yet classify chatgpt.com, perplexity.ai, claude.ai, or gemini.google.com as their own channel. Referrals from those hosts often arrive without a referrer due to privacy settings in the assistant, so they default to Direct. A custom channel group using source/medium regex rules recovers the visibility.

Three dashboards. One for content health (GSC queries, impressions, position by URL). One for funnel attribution (sessions to MQL to SQL to closed, segmented by channel). One for AI visibility (citation count across ChatGPT, Perplexity, Claude, and Copilot for priority queries). More dashboards than that produce noise, not decisions.

Citation traffic appears within 30 to 90 days of shipping the technical foundation and a weekly content cadence. Pipeline impact follows the sales cycle. For B2B SMBs with a 30 to 90 day cycle, expect a clear before-and-after on a 180-day window. Anyone promising measurable ROI in week two is either lying or counting impressions.

Google Search Console's queries-with-impressions-but-no-clicks report. Pages in positions 5 to 20 with thousands of impressions and CTR under 2% are sitting on rewrites that pay back inside a week. Most teams obsess over rankings 1 to 3 and ignore the rewrite tier where the leverage lives.

Rarely, and not at first. GA4's custom channel groups and GSC's built-in reports cover the four dashboards most SMBs need without a dedicated hire. A founder or marketing lead can build and read these reports in an afternoon once the channel groups exist. Bring in outside analytics help only once the reporting itself, not the data collection, becomes the bottleneck.

Assisted conversion counts a channel that appeared earlier in a buyer's journey, even if a different channel gets the last-touch credit for the sale. AI-assistant traffic disproportionately shows up as an assist rather than a last touch, because a buyer often reads an AI answer, then later searches the brand name directly to convert. Looking only at last-touch data understates how much AI citation is actually driving pipeline.

Compare GA4 total sessions against Search Console total clicks for the same date range. If GA4 shows meaningfully fewer sessions than GSC shows clicks, either the GA4 tag is missing on some pages or a consent-mode setting is blocking measurement for a chunk of real visitors. This single cross-check catches most tracking gaps before they cost a full quarter of missing data.