Real fleet audits — 10 SaaS products, 1 universal failure-mode
Five worked diagnoses from the AceVault fleet — a 30+ SaaS portfolio run by the operator of this site. Audited 2026-05-16. Sites anonymized but every score is real, computed against the published AUG v3 rubric. The framework gets dogfooded on the operator's own portfolio first.
The universal finding
100% of the audited fleet fails the same way: Acquisition score = 1 AND Monetization score = 1, regardless of how strong other factors are. One site has Advocacy = 10 (51,330 AI crawls/30d). One has Engagement = 8 (7% bounce rate). One has Activation = 6 (4 PV/session). All three still score AUG <0.15 because two factors are bottoming out. This is exactly the failure-mode the funnel-order-discipline principle predicts: until Acquisition is healthy AND Monetization is activated, downstream factors are statistical noise.
Site A · finance reference + 13F data
AUG 0.05 · Criticaldatabase/reference, finance vertical · ~5 weeks
| Acq | Act | Eng | Ret | Adv | Mon | Perf |
|---|---|---|---|---|---|---|
| 1 | 6 | 8 | 4 | 3 | 1 | 8 |
Weakest factor: Monetization (1) — AdSense recrawl pending
Diagnosis: Engagement is exceptional — 7% bounce rate + 3.88 PV/session (both top-decile). Acquisition is the binding constraint at 92 UV/30d. Monetization stuck at $0 because AdSense recrawl click hasn't happened post-compliance-fixes. Math floor: 92 UV × 2 PV × €15 RPM ÷ 1000 = €2.76/mo ceiling. Every Mon optimization is operator-time burn until Acq compounds.
Recommended fix: Acquisition compound (silent SEO + 1 ChatGPT referral measured already — first AI-citation win) + click AdSense recrawl request. Expected: AUG 0.05 → ~3 within 30 days post-recrawl approval. Funnel-order discipline: nothing else changes until Acq + Mon both lift.
Site B · book chapter summaries + viral TikTok
AUG 0.02 · Criticalcontent publisher, education adjacency · ~6 months
| Acq | Act | Eng | Ret | Adv | Mon | Perf |
|---|---|---|---|---|---|---|
| 1 | 3 | 4 | 3 | 8 | 1 | 6 |
Weakest factor: TikTok → site funnel broken (0% CTR despite 100K-view videos)
Diagnosis: Advocacy = 8 is the most asymmetric score in the fleet. TikTok account has 2,924 followers + pinned videos at 31K-108K views = real reach exists. But website Plausible shows 8 UV/30d. Conclusion: latent reach is high, but the bio link doesn't convert to clicks. Every TikTok view → website CTR is currently ~0%. Even 1% conversion = 1,000+ extra UV/mo.
Recommended fix: Fix the TikTok bio CTA. Add a /popular/ landing page that shows the most-viewed content first (matches TikTok intent). Email capture + Bookshop affiliate CTA replace the current footer. Estimated 90 min operator-time for end-to-end fix. Expected: AUG 0.02 → ~0.5 within 30 days. Real signal that Advocacy + Acquisition are independent dimensions — high Advocacy doesn't imply high Acq without a working funnel.
Site C · LLM-context directory + bot magnet
AUG 0.11 · Criticalstatic reference, AI-tooling vertical · ~7 weeks
| Acq | Act | Eng | Ret | Adv | Mon | Perf |
|---|---|---|---|---|---|---|
| 1 | 5 | 6 | 4 | 10 | 1 | 9 |
Weakest factor: Monetization (1) — 51k AI crawls/30d earning $0
Diagnosis: Highest Advocacy score in the fleet (10) — bot:human ratio is 2,700:1 (51,330 AI crawls/30d vs 19 humans/30d). Site is genuinely loved by GPTBot, ClaudeBot, PerplexityBot — but the bot-monetization layer (CF Pay-Per-Crawl + TollBit) is not yet active. Theoretical revenue at $0.02/crawl × 51k × 20% billing rate = ~$200/mo, ~$500/mo theoretical. Currently extracting $0.
Recommended fix: CF Pay-Per-Crawl beta approval (already applied 2026-05-06). Once approved, Mon score jumps from 1 to 5. AUG 0.11 → ~3 within 7 days of activation. Pair with TollBit fallback marketplace. Lesson: bot-rich Advocacy without bot-monetization wiring is leaving demonstrable revenue on the table.
Site D · fermentation calculator
AUG 0.006 · Zombiecalculator, food vertical · ~9 months
| Acq | Act | Eng | Ret | Adv | Mon | Perf |
|---|---|---|---|---|---|---|
| 1 | 4 | 4 | 2 | 2 | 1 | 10 |
Weakest factor: Every factor — flat trajectory after 9 months
Diagnosis: Performance perfect (10) — sub-200ms TTFB, 34KB pages. Every other factor weak. 5 UV/30d after 9 months means trajectory is flat, not compounding. Per I-35 (AUG <5 for 2+ consecutive weeks = 90-day kill-criteria candidate), this site triggers the framework's kill-or-pivot decision.
Recommended fix: Operator-decision required: (a) Kill — sunset, redirect to fleet anchor; (b) Pivot — rebuild around 200+ programmatic vegetable × method pages + Pinterest registration (food vertical compounds well there); (c) Accept-as-zombie — zero ongoing investment, no future review. Recommended: Pivot. Food/fermentation has real Pinterest volume, the niche just hasn't been worked.
Site E · neighborhood data + address tool
AUG 0.03 · Needs focuscalculator/reference, location vertical · ~3 months
| Acq | Act | Eng | Ret | Adv | Mon | Perf |
|---|---|---|---|---|---|---|
| 1 | 6 | 8 | 3 | 2 | 1 | 9 |
Weakest factor: Acquisition (1) — but trajectory is compound, not flat
Diagnosis: Activation = 6 + Engagement = 8 = strong (4 PV/session unusual for first-time visitors). Just migrated infrastructure (CF → Vercel) for reliability. Recent commits show active development (5 commits in last 6 days including WCAG mobile + Microsoft Clarity install). This is what a healthy pre-Acquisition site looks like: technical foundation solid, traffic compound about to kick in.
Recommended fix: Hold. Don't change product. Wait 60-90 days for SEO compound. Apply for AdSense at month 4 (or now if 10+ substantive pages exist). Expected trajectory: AUG 0.03 → 0.3-0.5 within 90 days based on silent-SEO compound rate. The only thing this site needs is time — every other lever is correctly tuned.
Methodology + caveats
- Performance: measured via real-time curl probe (TTFB + page-weight + render-blocking script count). ~70% of full Core Web Vitals signal; Google PSI API quota was exhausted at audit time. Confidence: 0.7.
- Acquisition + Monetization: confirmed from site analytics (Plausible / Cloudflare Web Analytics) + monetization-stack state files. Confidence: 0.85.
- Activation + Engagement: bounce rate and PV/session measured where available; remaining sub-components estimated against the rubric median. Confidence: 0.6.
- Retention: 8 of 10 fleet sites lack cohort tracking. Estimated against archetype defaults. Confidence: 0.4.
- Advocacy: AI-citation count + TikTok follower data + measured cross-site referrals where present. Confidence: 0.75.
Composite formula: AUG = 100 × Acq × Act × Eng × Ret × Adv × Mon × Perf ÷ 10⁷ — multiplicative so a zero in any factor near-zeros the whole. See full scoring transparency.
Next audit: 2026-05-23 (weekly Sunday cadence via the operator's fleet-audit cron). Drift >5 points week-over-week auto-flags as a rollback-candidate per I-22.
Audit your own SaaS
60-second 7-factor scan. Free. No signup. Same rubric used on the fleet audits above.