🧩

When your metrics tell different stories

A framework for reading discovery, retention, and conversion together — instead of panicking over one number.

This page is AlgoLens's own analysis, not an official statement from YouTube. It's meant to help you read your own channel's numbers.
TL;DR — Views, retention, and subscriber growth can look like they're telling different stories — one up while another stalls. That's not a contradiction to panic over; each metric maps to a different stage of your funnel (discovery → retention → conversion). Group your numbers by stage, compare each stage to your own channel's usual range, and work on whichever stage is weakest first.

Definition

It's common to see one metric look strong while another looks weak on the same video — views climbing steadily while retention dips, or a great average view duration paired with almost no new subscribers. These aren't contradictions. Each metric measures a different step a viewer takes: whether they discover the video, whether they keep watching, and whether they decide to follow the channel.

AlgoLens has two related concepts for reading this: combined analysis — looking at two different metrics together, such as upload timing and upload consistency, rather than one at a time, used to spot patterns that a single metric alone would miss — and bottleneck interpretation — pointing to which of the three steps (discovery, retention, or subscriber conversion) is currently weakest, used to decide which step to work on first.

Reading metrics by funnel stage

AlgoLens's per-video diagnosis breaks a video into three stages instead of one number: discovery (views compared to your own channel's typical view count), retention (your video's average retention compared to your own channel's retention baseline), and conversion (net subscribers gained compared to your own channel's typical range). Comparisons always use your own channel's past videos — never other channels — and the diagnosis flags whichever stage looks weakest as the one worth working on first.

What to do

1
Instead of reading each metric on its own, group them by funnel stage — discovery, retention, conversion — before deciding what's wrong.
2
Compare each stage to your own channel's typical range, not to a fixed number or someone else's channel.
3
Work on the weakest stage first. Polishing a stage that's already fine rarely moves the needle as much as shoring up the weak one.

Where AlgoLens helps

AlgoLens's Video Analysis tab runs this three-stage comparison automatically for every video and flags whichever stage needs attention most, so you don't have to eyeball several tabs and metrics at once to figure out where to focus.

Related terms

FAQ

More on this topic

What a low subscriber-conversion rate means → Finding where viewers stop watching →