I just hit Day 15 of an AI-driven Build-in-Public push for Tasteck, a vertical SaaS I run. Sharing the actual numbers because most "Build-in-Public works for SEO" claims you see online lack data.
TL;DR — 15 days, 4/24 → 5/8
| Metric (28-day total) | 4/23 baseline | 5/8 (Day 15) | Change |
|---|---|---|---|
| Clicks | 19 | 141 | 7.4x |
| Impressions | 281 | 2,705 | 9.6x |
| CTR | 6.76% | 5.21% | down (impressions grew faster, absolute rate is healthy) |
| Avg position | 7.4 | 6.9 | slight improvement |
7.4x clicks / 9.6x impressions in 15 days.
What I shipped
Volume layer (4/24-4/27)
- 12 niche-industry guide blog posts in 4 days (one per vertical use case)
- 5 long-form Note articles (Japanese platform similar to Medium)
- Daily GSC URL inspection requests (12/day quota)
Trust layer (4/28-5/4)
- Daily Build-in-Public log posts
- Industry KPI benchmark report Q1 edition
- Zenn (Japanese dev community) + dev.to cross-posts in English
- X / Note / dev.to community engagement
Incident layer (5/5-5/8)
- Volume 6: Stripe webhook silent failure for 5 days — 4xx retry trap incident report (5/5)
- Volume 7: PR-only → PR + monetize pivot, /work consulting page launch (5/7)
- Volume 8: 4-year-old auth-bypass vulnerability hot fix in our password-reset API (5/8)
- Industry KPI benchmark report Q2 edition (5/8)
Daily clicks growth
4/14: 1 4/24: 3
4/15: 2 4/25: 3
4/16: 0 4/26: 0
4/17: 3 4/27: 10 ← Volume blogs starting to be picked up
4/18: 5 4/28: 3
4/19: 1 4/29: 6
4/20: 4 4/30: 5
4/21: 2 5/1: 9
4/22: 4 5/2: 6
4/23: 6 5/3: 8
5/4: 9
5/5: 9
5/6: 15 ← Volume 6 Stripe webhook incident publish day
5/7: 16 ← Volume 7 /work launch + Volume 8 prep
5/8: 11
The two peak days (5/6 = 15, 5/7 = 16) align exactly with the publish dates of incident-report blogs. That's not a coincidence.
Top 5 pages by clicks (last 7 days, 5/2-5/8)
- Industry-specific repeat-customer rate guide (published 4/27): 16 clicks / 341 impressions / position 5.1
- Homepage: 12 clicks
- Industry confirmation tax guide: 6 clicks
- Industry NG-customer detection guide: 6 clicks
- Industry LINE bulk-messaging guide: 5 clicks
Notice: the #1 page was published 4/27 and only started getting real traffic from 5/2 — 5 days from publish to SEO traction, consistently across volume blogs.
Position 1-2 queries (niche industry terms)
| Query | Position | CTR |
|---|---|---|
| Industry-A reservation | 1.0 | 100% |
| Designation type A vs B (ambiguous niche term) | 1.4 | - |
| Industry repeat-customer rate calculation | 5.0 | 10.5% |
| Industry-B customer management | 11.7 | 4.5% |
| Industry-B system | 5.0 | 9.5% |
| Industry-C customer management | 9.0 | 50% |
| Tasteck (brand) | 4.2 | 25% |
"Industry designation type A vs B" at position 1.4 is small but huge — Google has effectively designated my page as the canonical definition for this niche industry term. Once that happens, position 1-2 becomes stable because there's almost no competition for these vertical terms.
Lesson 1: 3-layer model (volume × trust × incident)
The 15-day data exposed something I hadn't fully expected — different action types pay off on different timelines.
| Layer | Pay-off timing | Reach type |
|---|---|---|
| Volume layer (vertical SEO blogs) | 5-14 days from publish | Stable later reach, traffic from operators searching specific terms |
| Trust layer (Build-in-Public logs) | Direct SEO is weak; cumulative trust is strong | Direct reach is small, but without trust layer, incident-layer credibility doesn't land |
| Incident layer (Stripe / passwordReset) | Same-day burst | Tech-dev community share + brand-search boost |
Crucially, these three layers must be combined intentionally. Volume alone has no hook. Trust alone has no traffic. Incident alone has no continuity.
Lesson 2: incident-report blogs have burst reach
The two peak days (5/6 + 5/7) were both incident-report blog publish days. The pattern:
- Publish day: tech-dev community shares immediately on X / dev.to / Zenn → direct click traffic
- Reader profile: not industry operators but engineers, so CTR is higher (5.78% on 5/7)
- Side effect: brand-name (e.g., "Tasteck") query impressions get a boost in the following week
You can't ship incidents on demand, so the realistic strategy is "earn with volume daily, burst with incidents when they happen."
Lesson 3: Build-in-Public logs have a hidden role
The conventional wisdom "Build-in-Public is good for SEO" turned out to be half right.
- Logs alone: minimal direct search traffic (no keyword targeting, by design)
- BUT — the accumulated log stream is what makes incident-report blogs credible when they hit
- Without the log layer, incident posts feel disconnected; readers can't see the context behind why this particular issue arose now
So Build-in-Public logs work as "prerequisites" for incident posts, not as a direct SEO play.
Next 15 days (Day 16-30)
- Q3 industry KPI benchmark blog (by Day 30)
- post-incident structural-fix retrospective (UNIQUE INDEX + credential-id contract migration after the passwordReset case)
- Continue dev.to cross-posting for international reach
- Switch to 2x daily activity rhythm (11:30 + 20:00) instead of one large evening burst — to test if continuity scales the curve
I'm running Tasteck as a vertical SaaS in production for 8+ years (NestJS + TypeORM + Next.js + Stripe + AWS) and currently take freelance work in Stripe / NestJS / Next.js spot development and AI consulting. The corp HP for the operating company (EST FORT Inc.) is at est-fort-site.vercel.app.
If you're running a similar Build-in-Public push and want to compare data, drop a comment — I publish the raw GSC numbers because the field is still light on real datasets.
Top comments (0)