Reverse-engineer how a competitor grew — Wayback teardown, propagation mapping, 6-stage flywheel scoring.
By Iris Wei (生姜) · ex-COO of AFFiNE (60K+ GitHub stars) · 30× Product Hunt #1
When a competitor blows up, the instinct is to copy their latest move — usually the wrong one. This playbook reverse-engineers the whole flywheel: how their website evolved, how the launch propagated, and which growth stages are actually strong — so you copy the system, not the symptom.
| AI startups analyzed | 150+ |
| Website versions compared | 3 (V1/Beta/Launch) |
| Flywheel stages scored | 6 |
| Lovable launch-day impressions | 229K |
| Channels in AppFlowy's zero-budget launch (self-reported) | ~40 |
| Stars in week 1 after replicating the SOP (AFFiNE, self-reported) | ~6,000 |
| Replicable result for other teams (self-reported) | 1,000-3,000 stars / 3 months |
1 · Website teardown (3 versions)
Pull V1, Beta and Launch snapshots from the Wayback Machine and record, per version: one-line positioning, page length/structure, use cases, CTA/pricing, SEO changes, and which community channels launched when.
2 · Social media teardown
Use SimilarWeb to find their main traffic source, then deep-dive the right channel: official vs founder vs DevRel vs KOL accounts, posting cadence, and content-type mix.
3 · Map the propagation chain
Reconstruct the launch as 4 phases — internal ignition (hours) → ecosystem catalysis (1–2 days) → KOL detonation (2–3 days) → broad penetration (weeks) — and identify the traffic kings, topic starters, and mindset-shifters.
4 · Score the growth flywheel
Rate all six stages (Activation, Referral, Acquisition, Retention, Revenue, Product) strong-vs-weak using signals like Discord DAU, systematic user-story propagation, and clear free-to-paid triggers.
When mapping a competitor's launch, weight the content types that actually drove spread.
| Rank | Content type | Why it travels |
|---|---|---|
| 🥇 | Visual demo (video/GIF) | Strongest immediate impact |
| 🥈 | Real user success story | Most credible |
| 🥉 | Competitor comparison | Sparks discussion |
| 4–5 | Stats / feature announcements | Social currency / narrow reach |
How do I reverse-engineer a competitor's growth?
Run the 4 steps in order — website teardown (3 Wayback versions), social-media channel analysis, traffic-source mapping, ad-spend — then score their 6-stage growth flywheel to find where the strength actually is.
Which content types drive the most spread?
Validated ranking: visual demos (video/GIF) > real user success stories > competitor comparisons > stats/feature announcements. Weight your teardown — and your own launch — toward the top of that list.
What tools do I need?
Wayback Machine for historical website snapshots, SimilarWeb for traffic-source distribution, and Semrush for SEO and ad-spend analysis.
Who built this?
Iris Wei (生姜) — ex-COO of AFFiNE (60K+ GitHub stars), from 150+ AI-startup competitive analyses including a full Lovable launch teardown.
How do I research what a competitor did before they got big?
Rewind with Google's date-range filter (e.g. to 2016/2018), compare their taglines across eras, and dig archived early user cases — ClickUp reportedly started by serving ~10 marketing consultancies to crack one vertical before expanding.
How do I reverse-engineer a competitor's launch day?
Look up their open-source date on star-history, then date-restrict Google searches to that day/week. AppFlowy's launch reportedly used only ~40 channels with zero budget; replicating that SOP helped AFFiNE reach ~6,000 stars in week one (self-reported).