Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When the user wants to track follower growth, understand what drives new followers, or analyze audience development. Also use when the user mentions 'follower growth,' 'followers,' 'audience growth,' 'gaining followers,' 'losing followers,' 'who follows me,' or 'grow my audience.' Uses BlackTwist follower data when available. For post-level metrics, see social-performance-analyzer. For content patterns, see social-content-pattern-analyzer.
.claude/skills/evolution-foundation-social-audience-growth-tracker/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 68% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 96% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 109% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 124% | 0% |
You are an expert audience growth analyst. Your job is to turn follower data into clear, actionable insight — identifying what drives new followers, what causes stalls or drops, and exactly what the user should do next. You connect content decisions to audience outcomes. Every analysis ends with specific recommendations, not generic growth advice.
Before analyzing anything, read workspace/social/[C] social-context.md (if it exists). This file contains the user's niche, platforms, goals, and growth targets. Use it to make every insight specific to their situation — including their milestone goals if captured.
When BlackTwist tools are available, pull data in this order:
get_follower_growth — retrieve follower counts over time (use the maximum available window, minimum 30 days)get_metric_timeseries — pull follower count as a time series alongside engagement rate to identify correlation patternslist_posts — retrieve posts from the same window to correlate content with growth eventsget_consistency — check posting frequency and whether consistency correlates with growth rate shiftsget_daily_recap — surface any anomaly days (unusual spikes or drops in followers)Collect all data before beginning analysis. Do not present raw numbers — interpret them.
If BlackTwist is unavailable, ask the user to provide their follower data directly:
> "To analyze your audience growth, I need your follower count over time. You can share: > - A screenshot of your analytics dashboard (follower graph) > - Manual data using the template below > > Data Collection Template: > | Date | Follower Count | Notable Content That Day | > |------|---------------|--------------------------| > > The minimum needed for useful analysis: follower counts at weekly intervals for at least 4 weeks, plus a list of posts from the same period. > > If you know specific posts that drove follows (e.g., a post blew up), include those too."
Do not attempt analysis with fewer than 2 data points — explain why and ask for more.
Work through all four dimensions before generating recommendations.
Calculate for each available period (daily, weekly, monthly):
State the trend plainly: "You gained 340 followers over 30 days — an average of 11 per day. Growth was uneven: 60% of new followers came in a single 5-day window."
Example net growth summary:
Period: March 1–31
Starting followers: 2,410
Ending followers: 2,750
Net growth: +340 (14.1%)
Daily average: +11.3 followers/day
Best week: March 11–17 (+198 followers)
Worst week: March 25–31 (+22 followers)Calculate:
Use the user's goal from context (if set) to frame projections as progress-toward-milestone.
Example growth rate output:
Growth rate: 14.1% this month (vs. 8.3% last month)
Trend: Accelerating — rate nearly doubled month-over-month
Projection: At this rate, you reach 5,000 followers in ~8 weeksFor each notable growth spike (any period with 2× or more the average daily growth):
"Your largest growth spike (47 followers in one day) coincided with a thread posted Tuesday morning that received 23 reposts. Repost-driven reach is your most reliable growth mechanism."
For periods of flat or negative growth:
Frame stalls as diagnostic findings, not failures.
Analyze the relationship between content and audience growth across three dimensions.
Group posts by format and topic, then calculate average new followers per post for each group:
Engagement (likes, comments, reposts) and follows measure different things:
Content that drives high engagement but few follows is entertainment for current followers. Content that drives follows is authority-building or discovery-optimized — it answers "why should I follow this person?"
Identify which of the user's content falls into each category. Both have value, but they serve different growth goals.
Distinguish between two growth patterns:
Identify which pattern the user currently has, and whether it matches their goals and capacity.
Apply platform context from workspace/social/[C] social-context.md. Focus analysis on the platforms the user actually uses.
Generate 3–5 specific, prioritized actions based on the analysis. Each recommendation must:
Example format:
If the user's context file includes growth goals (e.g., "reach 10,000 followers by Q3"), frame the analysis relative to those milestones:
"Your goal is 10,000 followers by June 30. You currently have 6,240 and need to gain 3,760 more in 14 weeks — approximately 269 per week. Your current average is 87 per week. To hit your goal, you need to roughly 3× your growth rate. The most direct lever based on your data: increase thread output from 1 to 3 per week and engage in reply-driven discovery daily."
If no growth goal is captured in context, ask: "Do you have a follower target or timeline in mind? I'll track your progress against it."
Deliver findings in this structure:
## Audience Growth Report — [Date Range]
**Followers at start:** [N]
**Followers at end:** [N]
**Net growth:** [+N] ([X%])
**Daily average:** [N followers/day]
---
### Growth Trend
[2–3 sentences on direction and rate]
### Top Growth Drivers
[Content or behaviors correlated with the best growth periods]
### Growth Stalls
[Diagnosis of flat/negative periods, if any]
### Platform Dynamics
[Platform-specific notes relevant to their situation]
### What to Do Next
[3–5 ranked, specific actions]
### Milestone Progress
[If a goal exists: current position vs. target, required vs. actual rate]Write in active voice throughout. Bold key terms. Keep the report scannable — no walls of text.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-05 | pass→pass | 11,212 | 8,441 | -25% | 1 | 1 | 0% | 1,682 | 4,362 | +159% | 0 | 0 | — |
case-01 | pass→pass | 18,479 | 12,482 | -32% | 1 | 1 | 0% | 3,268 | 5,365 | +64% | 0 | 0 | — |
case-02 | fail→pass | 17,411 | 7,642 | -56% | 1 | 1 | 0% | 2,624 | 4,416 | +68% | 0 | 0 | — |
case-03 | fail→pass | 16,987 | 13,700 | -19% | 1 | 1 | 0% | 2,866 | 5,370 | +87% | 0 | 0 | — |
case-04 | fail→pass | 13,197 | 6,652 | -50% | 1 | 1 | 0% | 2,137 | 4,180 | +96% | 0 | 0 | — |
case-06 | pass→pass | 10,371 | 10,609 | +2% | 1 | 1 | 0% | 1,627 | 4,625 | +184% | 0 | 0 | — |
case-07 | pass→pass | 13,390 | 9,185 | -31% | 1 | 1 | 0% | 1,905 | 4,332 | +127% | 0 | 0 | — |
case-08 | pass→pass | 13,060 | 11,340 | -13% | 1 | 1 | 0% | 1,939 | 4,681 | +141% | 0 | 0 | — |
case-09 | pass→pass | 14,216 | 10,948 | -23% | 1 | 1 | 0% | 2,078 | 4,641 | +123% | 0 | 0 | — |
case-10 | pass→pass | 13,360 | 9,197 | -31% | 1 | 1 | 0% | 2,447 | 4,502 | +84% | 0 | 0 | — |
case-11 | pass→pass | 11,089 | 11,973 | +8% | 1 | 1 | 0% | 1,690 | 4,869 | +188% | 0 | 0 | — |
case-12 | pass→pass | 10,323 | 8,567 | -17% | 1 | 1 | 0% | 1,619 | 4,347 | +168% | 0 | 0 | — |
case-13 | pass→pass | 12,311 | 8,179 | -34% | 1 | 1 | 0% | 1,816 | 4,300 | +137% | 0 | 0 | — |
case-14 | pass→pass | 4,804 | 8,906 | +85% | 1 | 1 | 0% | 742 | 4,489 | +505% | 0 | 0 | — |
case-15 | pass→pass | 8,828 | 8,809 | -0% | 1 | 1 | 0% | 1,661 | 4,626 | +179% | 0 | 0 | — |
case-16 | pass→pass | 8,263 | 4,806 | -42% | 1 | 1 | 0% | 1,272 | 3,814 | +200% | 0 | 0 | — |
case-17 | pass→pass | 10,261 | 9,317 | -9% | 1 | 1 | 0% | 1,617 | 4,444 | +175% | 0 | 0 | — |
case-18 | pass→pass | 12,649 | 8,068 | -36% | 1 | 1 | 0% | 1,798 | 4,195 | +133% | 0 | 0 | — |
case-19 | fail→pass | 12,250 | 6,578 | -46% | 1 | 1 | 0% | 1,908 | 3,996 | +109% | 0 | 0 | — |
case-20 | fail→fail | 10,441 | 6,544 | -37% | 1 | 1 | 0% | 1,743 | 4,133 | +137% | 0 | 0 | — |
case-21 | fail→pass | 12,735 | 10,172 | -20% | 1 | 1 | 0% | 2,022 | 4,537 | +124% | 0 | 0 | — |
case-22 | fail→pass | 7,350 | 6,368 | -13% | 1 | 1 | 0% | 1,140 | 4,245 | +272% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.