Customer acquisition cost as reported by ad platforms is almost always wrong for B2B — it's calculated from platform-side conversion events that overcount touches and inflate attribution. The agent calculates CAC from the ground up using CRM-verified data: total spend per channel in a period, divided by the number of new customers sourced by that channel based on CRM opportunity attribution. Payback period is calculated using the median ACV of channel-sourced deals and the gross margin of the product. These two metrics — real CAC and real payback period — are the foundation of the channel evaluation model and cannot be calculated accurately without connecting ad spend data to CRM deal data.

CAC measures efficiency for closed deals, but not every paid channel's value is captured at the bottom of the funnel. Some channels contribute most of their value in the middle of the sales cycle — keeping deals warm, reaching buying committee members who aren't directly engaged with the sales team, and accelerating deal velocity rather than sourcing net-new pipeline. The agent evaluates multi-touch pipeline contribution for each channel: how many opportunities had a paid touch from this channel, at what stage, and did those opportunities convert at a higher rate or faster velocity than opportunities without a touch from this channel. Channels that show strong mid-funnel influence — even when their sourced-pipeline CAC is high — are distinguished from channels that look good on paper but have no detectable influence on pipeline outcomes.

CAC-only channel evaluation misses the customer lifetime dimension — a channel that delivers higher CAC deals may be acquiring customers with higher LTV, lower churn, and stronger expansion behavior that makes the economics more favorable over a 3-year horizon than a lower-CAC channel acquiring churner-profile customers. The agent calculates LTV by channel cohort using CRM retention data and expansion revenue history, and adjusts the channel ROI calculation to reflect LTV rather than just initial ACV. The LTV-adjusted ranking can differ materially from the simple CAC ranking — a channel that looks inefficient on a CAC basis may rank first on a 3-year LTV basis, and vice versa. Both rankings are shown with the methodology made explicit.

The output of channel evaluation is a decision, not a report. The agent synthesizes CAC, payback period, pipeline contribution, sales cycle influence, and LTV-adjusted ROI into a channel tier recommendation: Tier 1 channels (invest and scale), Tier 2 channels (maintain and optimize), Tier 3 channels (test or reduce), and channels to exit. For each tier, specific budget guidance is generated: the current allocation as a percentage of total paid spend, the recommended allocation with rationale, and the projected pipeline impact of implementing the recommended shift. The recommendation accounts for diversification risk — a framework that concentrates 90% of paid budget in one channel creates fragility even if that channel has the best current ROI.