Most SLG teams know roughly what their cost-per-lead is by channel. Far fewer know cost-per-MQL, cost-per-demo-attended, or cost-per-customer with confidence — and those are the numbers that actually determine whether a channel is profitable.
Reading the curve
Healthy paid channels show a smooth, predictable cost ramp from stage to stage. Cost-per-Lead might be $30, Cost-per-MQL $90, Cost-per-Demo-Attended $400, Cost-per-Customer $2,500. The ratios between adjacent stages tell you the conversion rate at each step.
When one ratio is dramatically worse than the rest, that’s where the channel is leaking. A 10x jump from CPL to Cost-per-MQL usually means the ads are attracting low-intent traffic. A 10x jump from Cost-per-Demo-Booked to Cost-per-Demo-Attended usually means a no-show problem (see lesson 2.7).
Acting on the report
Use this report to pause individual ads, not entire campaigns. Two ads in the same ad set can have wildly different cost-per-stage curves — one delivering cheap, low-quality form fills and the other delivering expensive but qualified demos. Pause the first, scale the second.
Pair the report with a target CAC derived from your unit economics. Most B2B SaaS teams aim for CAC ≤ 1/6 of LTV, with CAC payback of 6–12 months. Channels that hit those numbers are scaling candidates; channels that miss by 2x or more are usually best paused while you fix the funnel.
- Sort by Cost-per-Customer ascending to find scaling candidates
- Sort by Cost-per-Customer descending to find channels to pause
- Compare CPL → Cost-per-MQL ratios to spot intent-quality issues
- Compare Cost-per-Demo-Booked → Cost-per-Demo-Attended to spot no-show issues
What to watch for.
- Optimizing only on cost-per-lead
CPL is the easiest number to move and the least predictive of revenue. Always pair it with at least one mid-funnel cost.
- Comparing channels without the same window
Long-cycle channels look worse on a 30-day window than they actually are. Use a 60–90 day window for SLG.
- Not differentiating Demo Booked vs Demo Attended
These are different events with different costs. Tracking only one of them hides half the story.
Recap.
- Surface ads with great CPL but poor MQL conversion — usually low-intent traffic
- Identify campaigns where the cost-per-attended-demo is 3–5× the cost-per-booked-demo
- Pause ads on cost-per-customer, not cost-per-lead
- Compare cost-per-stage to your target CAC — most teams aim for CAC ≤ 1/6 of LTV
- Add account-level revenue to see true LTV ROAS by source