G2 recently reported a surprising gap in its referral traffic. AI discovery accounted for more than 5% of its referral sessions during the second quarter of 2026, but more than 20% of referral leads and over 10% of referral-attributed revenue.

The figures came from G2 CMO Alex London in a November 3 interview with Chief Marketer. G2 is a software review marketplace that also sells advertising and buyer-intent products to vendors, giving it a commercial interest in demonstrating the value of its audience.

For demand gen teams, those numbers tell three different stories. AI referrals account for a relatively small reported share of traffic but a much larger reported share of leads, while the revenue figures add another layer to the comparison. Traffic, leads, and revenue each tell part of the story, but none can establish a channel's value on its own. Depending on which metric gets the most attention, the same channel can look either modest or unusually effective.

Why Traffic Volume Can Be Misleading

Demand gen reporting often puts a lot of attention on traffic volume because it gives teams a consistent number to compare over time. But two thousand visits from one source and two thousand from another can represent completely different levels of buyer interest.

G2's numbers make that difference easy to see. Business of GTM estimated that, if G2's reported percentages are close to their stated floors, an AI referral session could generate roughly four times as many leads as the average referral session.

A channel generating relatively little traffic but a much larger share of leads may be performing better than its session numbers suggest.

What Revenue Reveals About AI Referrals

The third number complicates the story. G2 reported that AI discovery contributed over 10% of referral-attributed revenue, compared with more than 20% of referral leads.

If the reported percentages were exactly 5%, 20%, and 10%, the implied results would also be roughly half the average revenue per lead and twice the average revenue per session.

Those are illustrative calculations based on the reported thresholds, not G2's actual performance rates. The company hasn't published exact shares or enough detail about deal sizes and sales cycles to explain the differences.

Still, the comparison shows how much the interpretation changes depending on the metric. AI referrals could look especially productive when measured by leads per session, less impressive when measured by revenue per lead, and strong again when measured by revenue per session.

Each measurement captures a different part of channel performance, and relying on just one could lead a team to the wrong conclusion about where its demand is coming from.

Follow the Channel Through to Revenue

The answer isn’t to stop tracking traffic. Sessions still show where people arrive, which content they choose to open, and how discovery patterns change over time. They just need to sit beside metrics that show what happened afterward. G2 is already separating AI referral sessions, leads, and revenue, which makes it possible to compare how much business each source of traffic actually produces.

Demand gen teams can start looking at lead conversion rates by referral source, whether those leads become qualified opportunities, and how much pipeline and revenue each channel contributes.

Those comparisons also need context. A channel generating early-stage interest may not produce revenue as quickly as one reaching buyers who are already close to making a decision. Judging both on immediate revenue alone could create another misleading comparison.

Better measurement also changes how teams make investment decisions. This connects to RAD Intel's broader decision intelligence approach, which brings together different sources of information to help marketers evaluate their options. If one channel produces fewer visits but more qualified leads, evaluating it primarily on traffic could lead a team to reduce spending on a source that's contributing to growth.

G2's numbers show why channel performance needs to be evaluated beyond whichever metric looks strongest. Before deciding where to invest, demand gen teams need to understand how traffic turns into leads, how those leads become opportunities, and what ultimately contributes to revenue.