Scores & Methodology

This page explains every score busypipe computes. The first half covers Post Intelligence — the five per-post scores. The second half covers the other computed metrics: Creator Insights, the Audio Hook Score, scoreboard velocity, portfolio pulse, and the weekly digest’s AI insights.

Post Intelligence

busypipe scores each TikTok post on five dimensions: Hook Strength, Audience Resonance, Reach Efficiency, Format Fit, and Topic Momentum.

Every score is 0–100, normalized as a percentile rank within a peer set (the creator’s last 90 days × same region by default). This means:

A score of 84 = this post is better than 84% of the creator’s recent posts on this dimension.
Scores are NOT comparable across creators absolutely — they’re each creator’s own relative performance.

When a post doesn’t have enough data for a score — too few peer posts, no first-day views yet, or no hashtags we’ve seen on enough posts — the tile shows “Not enough data” rather than a misleading low number.

Hook Strength

Question: Did the opening land?

Hook Strength is a composite of three sub-signals:

Sub-signal	Weight	What it measures
Early velocity	0.5	View growth in the first 24 hours, as a fraction of the post’s 7-day total — captures the algorithmic kickstart
Engagement density	0.3	(likes + comments + shares + saves) / views — measures whether people who watched cared enough to interact
Hook archetype quality	0.2	AI-classified opening type — curiosity_gap (0.9), direct_question (0.8), claim (0.7), story (0.6), generic_intro (0.3)

The first two are derived deterministically from engagement curve data. The third is an AI classification of the first ~5 seconds of speech, weighted by how well each archetype typically performs.

Formula

hook_strength = round( (0.5 × velocity_percentile)
                     + (0.3 × density_percentile)
                     + (0.2 × archetype_quality) ) × 100

All inputs are clamped to [0, 1]. The output is rounded to the nearest integer, 0–100.

Audience Resonance

Question: Did people care, or just scroll?

Audience Resonance blends three signals into one 0–100 score:

Signal	Weight	What it measures
Comment intensity	0.4	Comments per view, ranked against the creator’s last 90 days
Save & share signal	0.3	(Saves + shares) per view, ranked the same way
Comment sentiment health	0.3	Positive minus negative share of the post’s analyzed comments

Sentiment health

When a post has at least 10 stored comments, the latest 50 are analyzed for sentiment. Sentiment health is the positive share minus the negative share, scaled to 0–1 (all-positive = 1, balanced = 0.5, all-negative = 0).

Posts with fewer than 10 stored comments have no sentiment reading yet; the score is computed from the two engagement signals alone, re-weighted so the result stays comparable. Comments labeled positive, negative, or neutral show a small sentiment badge in the post’s comments section.

Reach Efficiency

Question: Did the algorithm reward this post?

Reach Efficiency compares a post’s first-day views to what we’d expect for this creator — then ranks the result against their recent posts. The expectation isn’t just the creator’s average; it’s adjusted for two things:

Posting time: some weekday-and-hour slots reliably over- or under-perform for a given creator. A post in a strong slot has a higher bar to clear.
Hashtag trend: if the post’s hashtags were already hot, more reach was expected; beating expectations on a cold topic counts for more.

A post that clears its adjusted expectation scores high. Posts without first-day view data yet show “Not enough data.”

Format Fit

Question: Was this on-brand for what works for this creator?

Format Fit measures how closely a post’s composition matches the creator’s best-performing posts. We summarize each post by its format — duration, caption length, hashtag count, whether the audio is reused, whether it has spoken content, orientation, and topic overlap — and compare the post to the average of the creator’s top 10% of posts by reach over the last 90 days. The closer the match, the higher the score.

This rewards sticking to a creator’s proven formula and flags experiments that stray from it. New creators without enough top-performing posts to compare against show “Not enough data.”

Topic Momentum

Question: Was the subject timely?

Topic Momentum looks at the post’s hashtags and measures whether each was being used by a growing or shrinking number of posts in the week around when the post went up. Riding hashtags while their usage was climbing scores high; posting on hashtags whose usage was falling scores low. The per-hashtag signals are averaged and ranked against the creator’s recent posts.

A post with no hashtags — or hashtags we haven’t seen on enough posts across multiple days — shows “Not enough data.”

Peer set

The default peer set for percentile calculation is the creator’s last 90 days of posts in the same region. Here’s why:

Same creator: Different creators have different audiences and algorithms — comparing @nike to @gordonramsay’s hook strength is meaningless. Comparing @nike to its own past posts is actionable.
Same region: TikTok’s algorithm varies by region. A post that scores 84 in the US may not be reproducible in JP.
Last 90 days: Long enough to span seasonality and content cadence, short enough that the comparison reflects recent performance — not posts from years ago.

Peer set configuration will become available in later milestones.

Creator Insights

Question: What is working for this creator overall?

Creator Insights appears on every profile page. The summary band (all plans) shows posts per week, average engagement rate, and the best post of the window. The deep dive (Starter and above) breaks the creator’s recent posts down further:

Posting cadence — a weekly posting histogram, the share of weeks with at least one post, the average gap between posts, and a trend direction (ramping up / steady / slowing down) based on whether the second half of the window out-posts the first half.
Engagement quality — a weekly engagement-rate trend plus median, best, and worst post ER. Here ER includes saves: (likes + comments + shares + saves) / plays.
Top posts — the top five by plays or by ER. A Breakout badge marks posts with at least 3× the creator’s median plays.
Best time & format — a posting heatmap (average ER by hour and weekday, in the creator’s regional timezone where it is unambiguous, otherwise UTC), the format mix, and the creator’s most-used hashtags and sounds.

The default window is the last 90 days, capped by your plan’s history window. With fewer than 8 posts in the window, distribution-based views (like the heatmap) are hidden and a small-sample note is shown instead — averages over a handful of posts would mislead.

The Analyze report

Above the deep dive, an Insights panel turns the same cadence, engagement, breakout, best-time, format, and top-hashtag/sound data into a plain-language read. Pick an angle and select Analyze:

What’s working — a read of the creator’s current strengths.
How to improve — concrete room for improvement, worded so it reads naturally on your own profiles or a competitor’s.
What changed — the biggest shifts in this creator’s activity over the selected window.
Full read — a complete summary across all of the above.

Each report includes a short written summary you can copy to share or reuse. Analyze uses your monthly analyze credits — the same shared pool as hashtag, audio, and portfolio insights — and a report that comes back empty (not enough activity in the window) doesn’t cost a credit.

Audio Hook Score

Question: Is this sound about to take off?

The Audio Hook Score (Professional and above) is a 0–100 composite shown on audio pages, blending five signals:

Signal	Weight	What it measures
Velocity	0.40	Growth in the number of creators using the sound, last 7 days vs the prior 7
Acceleration	0.30	Whether that growth is itself speeding up day over day, with a bonus for several consecutive rising days
Chart proximity	0.15	Room left to climb — a sound that hasn’t charted yet scores full marks; the higher it already charts, the less upside remains
Hook clarity	0.10	Whether the sound has a chorus of usable length (roughly 7–20 seconds) — the raw material of a hook
Geographic breadth	0.05	How many regional charts the sound has appeared in over the last 30 days

The composite maps to a label: Climbing (70+), Peaking (50+), Stable (30+), Fading (below 30).

Scoreboard velocity

Question: Who is accelerating, not just who is big?

The portfolio scoreboard’s velocity columns (Starter and above) compare each profile against its own recent past over the period you select:

Δ Followers — percentage follower change across the period, measured from the follower count at the period’s start.
ER drift — the change in engagement rate, in percentage points, versus the previous period of equal length. Scoreboard ER is (likes + comments + shares) / plays.

Velocity needs history on both sides of the comparison. Until a profile has been tracked long enough, its cell shows a “gathering — 3/7d”-style progress note (days collected out of days needed) rather than a misleading number. Snapshot columns (Followers, ER, Posts/wk, Avg views) always show current values on every plan.

Portfolio pulse

Question: What actually moved across the whole roster?

The pulse card lists the biggest movers among everything you track — profiles, hashtags and sounds — each with the change and, where a meaningful baseline exists, the percentage change.

Mover signals are measured over at least 14 days, and the card names the span it used. That floor is deliberate: follower and sound-usage counts aren’t sampled every day, so a narrower window would silently drop quieter entities rather than report them as flat. On a portfolio’s analytics page the card follows the period you select, and never shrinks below the 14-day floor. Only entities that actually changed are listed; a window with no notable movement says so instead of padding the card.

The KPI tiles are a separate, shorter view — the last 7 days across all tracked profiles, or the period you select where a period picker is available:

Followers — the sum of each profile’s latest follower count.
Growth / week — each profile’s follower growth over the window, normalized to a weekly rate, averaged across profiles.
Median ER — the median engagement rate of posts published in the window, where ER is (likes + comments + shares) / plays.
Posts (7d) — the number of posts your tracked profiles published in the window.

The best post highlighted under the pulse card is also a 7-day pick, so it can name a post the 14-day mover list doesn’t mention.

Pulse and scoreboard ER deliberately exclude saves, while Creator Insights ER includes them — the same posts can show slightly different ER in the two places.

Weekly AI insights

Question: Why is this hashtag suddenly everywhere?

The weekly trends digest ranks breakout hashtags by week-over-week growth in new posts (this week’s new posts versus the prior week’s; hashtags with no prior baseline are labeled NEW). Each digest issue covers one region at a time, so the ranking reflects that region’s trending data.

For each breakout, busypipe’s AI reads the week’s stats and a handful of the hashtag’s top posts, then writes a one-to-two-sentence explanation of why it’s moving. The AI is constrained to what the provided posts and stats support — it may not invent events, people, brands, or dates. Every insight ships with provenance links to the hashtag page and the top post it drew on, so you can check the receipts. When a grounded explanation isn’t possible, the digest falls back to a plain factual line instead.

Plan-gated access

Post Intelligence is available on your plan:

Feature	Free	Starter	Professional	Enterprise
Intelligence on demo posts	✓	✓	✓	✓
Intelligence on your tracked posts	—	✓	✓	✓
Brand mention extraction	—	—	✓	✓
Scheduled intelligence reports	—	—	✓	✓
Custom enrichment prompts	—	—	—	✓
Multi-language support	—	—	—	✓

Free plan: see Intelligence on one demo post per workspace.

Starter plan and above: Intelligence on every post from your tracked profiles.

Professional plan and above: Brand mention extraction and scheduled intelligence reports.

Enterprise: Custom enrichment prompts, multi-language support, and dedicated rate limits.

Upgrade your plan to unlock more scores on more posts.

When a score shows “Not enough data”

Some scores need specific measurements that aren’t available for every post:

Reach Efficiency compares a post’s first-day views against expectations, so it needs an engagement measurement from roughly 24 hours after publishing. Posts that started being tracked later than that can’t be scored.
Topic Momentum measures whether a post’s hashtags were rising or cooling around posting time. If we haven’t seen the post’s hashtags on enough posts across multiple days, there is no signal to score.

Click the Reach or Topic “Not enough data” tile on the post page to see exactly which measurement is missing for that post.