Your Clarity Score is the one number that tells you if you meant to watch what you watched.

What a number cannot tell you

Screen Time says I spent 3 hours and 12 minutes on YouTube last Tuesday. That is accurate. It is also nearly useless as a behavioural signal, because it cannot tell me whether those 3 hours were the most intentional viewing I did all month or a cascade of algorithmic drift that I barely participated in.

Duration measures presence. It does not measure agency. That distinction is what the Clarity Score is built around.

How the Clarity Score works

The Clarity Score is a per-session score, aggregated to a daily and weekly view, that estimates how intentional your watching was. It draws on several signals: how you arrived at each video (search, subscription feed, or homepage recommendation), whether you watched to completion or abandoned early, how diverse the session was, and whether the videos you watched align with channels you've explicitly engaged with versus algorithmic recommendations.

A session where you searched for something specific, watched it through, and then stopped scores high. A session where you clicked a homepage recommendation, abandoned it at 40%, clicked another recommendation, watched it halfway, and then spent 45 minutes in autoplay scores low.

The score runs from 0 to 100. I've found sessions below 50 are worth examining. Sessions below 30 are usually the ones I'm not proud of.

My own distribution over a month

Over a recent 30-day period, my Clarity Score distribution looked like this:

• Sessions scoring 80-100: 23% of sessions
• Sessions scoring 60-79: 31% of sessions
• Sessions scoring 40-59: 28% of sessions
• Sessions scoring below 40: 18% of sessions

That bottom 18% was the revealing number. Nearly one in five sessions was low-clarity, meaning I was on YouTube in a way that was mostly reactive, driven by the algorithm rather than by anything I'd decided to watch.

When I looked at what those low-clarity sessions had in common, the pattern was not subtle. 71% of my sub-40 sessions happened between 2pm and 4pm. A second cluster appeared after 10pm. Almost none of my high-clarity sessions happened in those windows.

The time-of-day signal

I had suspected I had a mid-afternoon YouTube problem. The Clarity data confirmed it precisely. What it added was the intentionality dimension: it wasn't just that I was watching YouTube at 2pm, it was that the sessions at 2pm were qualitatively different from my morning sessions. They were less purposeful, more driven by recommendations, more likely to end in a long autoplay chain.

Duration apps would have shown me the time-of-day pattern eventually, in aggregate. But they couldn't have told me that my 2pm watching was structurally different from my 9am watching. They would have just added the minutes together.

The entry-point breakdown

Another pattern the Clarity Score surfaces is the entry point into sessions. My high-clarity sessions almost always started from a search or my subscription feed. My low-clarity sessions almost always started from the YouTube homepage.

This is not a surprising finding in the abstract, the homepage is specifically designed to draw you into content you didn't plan to watch. But seeing it quantified in my own data made it concrete in a way that general knowledge about algorithm design doesn't. My homepage is, statistically, the entry point for most of my low-intention watching.

That knowledge changed a practical decision: I now default to opening YouTube to my subscription feed rather than the homepage. It's a small structural change. My Clarity Score improved noticeably in the following two weeks.

What the score cannot tell you

I want to be honest about the limits here. The Clarity Score is an estimate. It uses signals that correlate with intentional watching, not a direct measurement of your mental state. A low Clarity Score doesn't mean you did something wrong, sometimes you genuinely want to let the algorithm take you somewhere. A high score doesn't mean the time was well spent in any broader sense.

What it does is make a normally invisible quality visible. Intentionality is hard to track retrospectively. "Did I mean to watch that?" is a question most of us can't answer accurately from memory. The score gives you a structured approximation, built from behavioural signals, that you can look at honestly.

Duration apps cannot compute anything like this because they do not know how you arrived at a video. They see the watch. They don't see the decision that preceded it, or whether there was one at all.

The number I check first

I open Gazenest most days, and the number I look at first is not total watch time. It is my Clarity Score for the previous session. A score above 70 and I close the dashboard without concern. A score below 50 prompts me to look at the session breakdown and understand what happened.

It has become the most honest daily feedback loop I have on my YouTube use, more useful than a total, more legible than a raw log of what I watched.

The Clarity Score is available on Gazenest Plus.