Retention Rebuild Kit: fix a flat short

Your view counts all look the same. That is not the algorithm.

Short-form platforms hand almost every upload a small initial-velocity test — a seeded batch of impressions to see what happens. That is why your own videos cluster in the same few-hundred band whether they were good or bad, and why comparing your videos by view count teaches you nothing.

What decides whether a video escapes that test is watch-through. This kit works from that number.

What it does

You paste one video you already published — its script, its length, and its real analytics (average percentage viewed or average view duration, views, and where the curve falls if you have it). You get back an autopsy of that specific video and a rebuild you can shoot.

What you get

1. retention-report.html — one self-contained file. No external stylesheets, scripts, fonts, or images; no network requests. Open it in any browser, print it to PDF, keep it as a record or send it to a client. Inside: the verdict, the retention curve with the drop-off marked, a second-by-second diagnosis of the opening, and every leak ranked.

2. A verdict with the threshold printed next to it — 75%+ strong / 50-75% PASS / 35-50% REWORK / under 35% REBUILD, measured on average percentage viewed. The standard is always shown alongside the judgment, so you can argue with the threshold instead of with a black box.

3. A shot-by-shot cut list — seconds, what is on screen, what is said, and why each shot exists. Plus what to cut, what to compress and to what length, what to move earlier, and a target total length with the reason for it.

4. Five first-3-second hook variants — one each using contrast, an open loop, cost, a concrete number, and negation. Each comes with its mechanism, its failure mode, and how you would know it worked. It also tells you which variant it expects to lose, and why.

5. A pre-publish checklist and measurement plan — including verifying the upload actually reached a public state, because a video sitting in a draft or restricted-visibility state produces zero distribution and looks exactly like being buried.

It checks the things a hook rewrite cannot fix

Before it discusses your opening, it checks whether the video was distributable at all: public, not a duplicate of another upload, not competing with a near-identical title, and on a topic the account has covered before. Duplicate uploads weaken distribution for the whole account. Mixing unrelated topics makes the platform's guess about your audience worse. If either is happening, that is the finding — and no rewrite would have told you.

Honest limits

• Works only from the numbers and text you paste. No account linking, no analytics access, no scraping, no API key, no subscription.
• Never invents an analytics figure. Missing numbers are marked [FILL IN] with where to find them on your platform.
• Promises nothing about views, reach, followers, or revenue. It is a diagnostic and rewriting tool, not a growth guarantee.
• The verdict thresholds are working calibration, not platform-published pass marks. No platform publishes one, and this kit says so rather than implying otherwise.
• Declines, with the reason: buying views or engagement, bot or multi-account automation, evading platform enforcement, re-uploading third-party footage, and fabricating analytics.

Not the same as a hook generator

Hook lines are one of five deliverables here, and they get written only after the structural check. A new opening on a video whose real problem is its length, its topic fit, or its visibility status is wasted work.

Try the free version first

Retention Check (Free) gives you the verdict and the single biggest leak from one number. Use it to decide whether a video is worth rebuilding; use this kit when the answer is yes.