manifoldmcp/marketing-skills / skills/growth-strategy
growth-strategy
VerifiedWhen the user wants a growth plan, a choice of marketing channels, a go-to-market (GTM) plan, a launch plan, content ideas, a content calendar, or to repurpose a video or blog post as social posts. Holds dated, sourced findings a model lacks (how the channel mix shifted in 2025-2026, what a Product Hunt launch yields now, cadence and repurposing evidence, the attribution gap, which benchmarks to distrust) and how to take a cheap cross-channel baseline. Also use when the user mentions how do I grow my startup, a growth strategy or marketing plan, more traffic, signups, leads or customers with no channel named, traction, which channels should we be on, channel mix, a GTM plan, ICP and positioning for a new product, a launch plan or checklist, Product Hunt, Show HN, a waitlist, launch reactions, a posting schedule, content pillars, what to post, or turning a video or podcast into posts. One named channel goes to its own topic (seo, ai-search, reddit, linkedin, tiktok, paid-ads, outbound, creators, digital-pr).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.8 KB |
| references/benchmarks.md | markdown | 3.8 KB |
| references/channel-selection.md | markdown | 8.9 KB |
| references/content.md | markdown | 6.5 KB |
| references/field-notes.md | markdown | 3.6 KB |
| references/growth-plan.md | markdown | 6.9 KB |
| references/gtm-plan.md | markdown | 3.6 KB |
| references/launch.md | markdown | 9.1 KB |
| references/measurement.md | markdown | 4.4 KB |
| references/reaction-report.md | markdown | 3.9 KB |
| references/repurposing.md | markdown | 4.8 KB |
references/measurement.md
View on GitHub# Measuring growth
What is working, how much of it the analytics can see, and when a change is real. Read it before setting KPIs, judging a channel at day 30, or answering "where do our customers come from". Checked October 2026.
## The attribution gap
Most of what moves a buyer happens where no tracker reaches.
- **B2B buyers choose before contact.** 94% of about 4,000 B2B buyers ranked a shortlist before engaging a seller; first contact came 61% of the way through the journey (69% in 2023-2024); the favourite at that point won 77% of the time; buyers evaluated 5.1 vendors and started with 3.6 (6sense Buyer Experience Report, 2025, self-reported survey of buyers with $25,000+ in purchases). 94% used LLMs during the purchase, mostly to compare vendors.
- **The journey is long and mostly before the pipeline.** In Dreamdata's 2025 data (66M+ sessions, 3.5M+ journeys from its customers; attribution model not stated), 81% of the journey happened outside the sales pipeline, a deal took 88 touches across 4 channels and 10 stakeholders, and the first LinkedIn ad impression came a mean 281 days before revenue. Organic LinkedIn company-page impressions appeared in 17.9% of closed deals, about 90 days before revenue.
- **AI answers influence without clicking.** AI assistants sent 0.24% of sessions across 54 sites (Amsive, 2025) while buyers use them throughout the purchase. Many AI visits also arrive with no referrer from apps and land in "direct"; Ahrefs suspected ChatGPT was not passing referral data before mid-2025.
So last-click reports undercount every channel that works early (communities, founder content, podcasts, AI answers, PR) and overcount the ones at the end (branded search, direct, retargeting). Say this whenever a user is about to cut a channel on last-click numbers alone.
## What to set up on day one
- **A required free-text "How did you hear about us?" field** at signup or demo request. It is the one cheap way to see podcasts, communities, AI answers and word of mouth. No study has validated it against ground truth; read it beside the software numbers, never instead, and code the answers into the plan's channels by hand.
- **One UTM link per venue, creator and post type**, so a launch or a creator can be counted.
- **AI referrals as their own segment:** `analytics_get_report` (free) with `dimensions` including `sessionSource` and a filter whose `operator: "regex"` matches chatgpt.com, perplexity, gemini.google.com, copilot and claude.ai. GA4 does not group them for you by default.
- **The starting number of the deciding KPI**, with its date, before any work begins.
## Proxies the tools re-measure
- **Branded search** is the cleanest awareness proxy: `console_get_search_analytics` (free) with a `query` filter containing the brand, daily; or `seo_get_keyword_metrics` on the brand, monthly through `trend[12]`.
- **AI mention rate** on a frozen prompt set: [ai-search](../../ai-search/SKILL.md) has the method and its error bars.
- **Organic clicks and ranks:** `console_get_search_analytics`, `seo_get_position`.
- **Community volume:** weekly posts about the problem in the same subreddits with `reddit_get_new_posts` (pages enough to cover the window).
- **Referring domains** after coverage: `seo_get_referring_domains`, reading `first_seen`.
They are leading signals. Signups, pipeline and revenue come from the user's analytics and CRM.
## Is it real
Small numbers move by chance. Before calling a channel won or lost:
- **Weekly counts.** A count of n events a week varies by about ±2√n by chance. At 25 signups a week, a move to 35 is within noise; two weeks compared differ by chance by up to about 14 at that level. Read three-week averages.
- **Conversion rates between channels** need samples per channel of about (95% confidence, 80% power):
| Rates to tell apart | Visitors per channel |
|---|---|
| 10% against 15% | about 700 |
| 2% against 4% | about 1,100 |
| 2% against 3% | about 3,800 |
| 5% against 6% | about 8,100 |
- **Spread thin, nothing is readable.** Five channels at a few hundred visitors a month each cannot be told apart in 90 days; two can. That arithmetic, not a study, is the strongest case for few channels.
- **Re-measure with the same calls, the same competitors and the same queries,** or the comparison measures the change of method.
- **Mixed results across sites are normal.** In Amsive's 54 sites, LLM traffic converted above the site's average on 56% and below on 41%: a portfolio mean says little about one site. Measure the user's own.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |