How do you measure marketing performance, and which metrics are misleading?

Measure against the objective, and distinguish three tiers.

Business metrics — what you are accountable for:

• Revenue and pipeline influenced by marketing, customer acquisition cost, LTV:CAC ratio, payback period, and market share.

Channel metrics — for allocating budget:

• Cost per acquisition and conversion rate by channel, return on ad spend, and lead-to-customer rate by source.

Diagnostic metrics — for finding causes, not for reporting success:

• Impressions, click-through rate, bounce rate, time on page, follower counts.

The misleading ones, and why:

• Impressions and reach — measure spend, not effect.
• Social followers — the classic vanity metric. Easily inflated and weakly related to revenue.
• Cost per lead — dangerous because it ignores lead quality. The cheapest leads are frequently the worst, and optimising for it can reduce revenue while the dashboard improves.
• Click-through rate alone — a compelling ad that oversells produces clicks and no sales.
• Last-click attribution — systematically over-credits the final touchpoint and under-credits everything that created the demand.
• Blended averages — hide a profitable channel subsidising a losing one.

Note: The best marketers report a small number of business metrics and keep diagnostics for their own investigation. A dashboard with forty metrics communicates nothing.