Vanity metrics vs. pipeline metrics: how B2B teams should actually score a clip

Every team publishing B2B content eventually hits the same wall: engagement is up, and pipeline isn't. It's not a content-quality problem in the way people usually mean it — the clips are often genuinely good. It's a measurement problem. You've been scoring for the wrong thing.

The problem with optimizing for reactions

Likes, views, and comments are easy to optimize for because they're fast, public, and platform-native — the algorithm hands them to you in real time. But a reaction happens in isolation, scrolling, with zero context about whether the viewer is anywhere near a buying decision. Someone can like a clip about your product's core differentiator and still have no idea what your company sells. Reactions measure attention. They don't measure conviction.

What "pipeline" actually means for a clip

A clip contributes to pipeline when it changes what happens in a real sales or evaluation conversation — a rep sends it instead of writing a paragraph, a prospect forwards it to their VP before a renewal call, a buyer references a specific line from it three weeks later during a vendor comparison. None of that shows up as a like. It shows up as a forward, a DM, a "saw your post" in a discovery call. It's slower and harder to track, but it's the thing that was actually worth making the clip for.

The forward test

Before a clip goes out, ask one question: would someone forward this to a specific colleague, and could you name that colleague? If you can picture the exact person and the exact reason ("this is the objection Sarah keeps hearing from finance"), the clip survives the forward test. If the honest answer is "I guess someone might like it," it's a reach play, not a pipeline play — and that's fine, as long as you're not scoring it as if it were the latter.

Clips that pass the forward test tend to share three traits: they make one specific, defensible claim rather than a general truism; they name the tension a buyer is actually feeling (the "why now"); and they resolve in a way that's quotable on its own, out of context, in a Slack message or an email.

Three questions to ask before you cut a clip

• Is this specific enough to be wrong? Vague claims are safe and forgettable. A claim precise enough that a competitor could disagree with it is usually the one worth cutting.
• Does it name a real moment, not a general topic? "We talk about growth" is a topic. "Most teams optimize for the wrong metric" is a moment — it's got a specific tension in it.
• Could I send this to one named person and expect them to reply? If yes, cut it. If you can only imagine a crowd reacting to it, it's a reach clip — still fine to publish, just don't expect it to move pipeline.

Scoring for this instead of reach

In practice, this means grading candidate clips on specificity and framework-strength before you ever look at predicted reach. A hook that names a real mistake, a quotable line with a defensible claim, and a framework a buyer can reuse in their own internal argument (a "does it survive X" test, a named tradeoff) all score higher than a generically upbeat soundbite — even if the soundbite would get more likes. This is the same transcript-scoring logic Unified's clip detection runs on every upload: hooks, quotable lines, and frameworks are ranked separately from raw reach potential, so the clips that surface are the ones actually worth sending to a specific person, not just the ones a feed will reward.