SaaS growth friction case studies

Five common SaaS-growth scenarios, diagnosed against the AUG v3 framework. Models from a portfolio of 30+ live SaaS products — composites of patterns we see repeatedly. Names + numbers are anonymized; the failure modes are real.

Case 1 of 5

The pricing-page tweaker

SaaS analytics dashboard, 6 months old

AUG Score

0.40

Biggest friction

Acquisition2/10

Acq

2

Act

7

Eng

6

Ret

5

Adv

3

Mon

4

Per

8

Diagnosis

AUG 3.36. Founder spent the last 8 weeks A/B testing pricing-page copy. Acquisition score of 2 (sessions ~400/month) means revenue ceiling per the math floor ≈ €5/month — no pricing tweak beats that ceiling. Activation, Engagement, Performance are already healthy. The bottleneck is the top of funnel; the founder is optimizing the wrong factor.

Recommended fix

Stop pricing experiments. Spend 4 weeks on Acquisition fundamentals: programmatic SEO pages for 3 comparison queries, llms.txt + AI-crawler allowlist, IndexNow on every deploy, one HN Show HN. Re-audit at month 9. Expected AUG lift: ~3.4 → ~7.2.

Case 2 of 5

The signup-celebration trap

B2B utility, 1 year old

AUG Score

0.67

Biggest friction

Retention2/10

Acq

6

Act

8

Eng

5

Ret

2

Adv

4

Mon

5

Per

7

Diagnosis

AUG 3.41. Founder celebrates 5,000 signups in 12 months. But D7 retention is 4% — users sign up, complete onboarding, never return. Activation score 8 is misleading; the activation event is "complete onboarding" which doesn't correspond to "still active in week 2". The product has a one-time intent (filing taxes, generating a one-shot artifact) and was built like a recurring SaaS.

Recommended fix

Either change the product (add weekly-recurring value: progress tracking, weekly digest, version control on past artifacts), or change the business model (one-time purchase instead of subscription). The framework is opinionated: don't build subscription products around one-time intent.

Case 3 of 5

The slow dashboard

SaaS metrics tool, 8 months old

AUG Score

0.03

Biggest friction

Performance1/10

Acq

7

Act

4

Eng

3

Ret

4

Adv

2

Mon

5

Per

1

Diagnosis

AUG 0.84. Founder hired a performance consultant who said "your Lighthouse score is 92 — performance is fine." Wrong score — field data from PageSpeed Insights shows LCP p75 = 7.2s (mobile, 4G). Lab tests run on desktop fiber; real users are on phones with slow connections. Every other factor (Activation, Engagement, Retention, Advocacy) is artificially low because users bail before the dashboard finishes loading.

Recommended fix

Single highest-leverage week of work in this case study: ship static-export landing pages for public surface + lazy-load the dashboard chunks. Re-measure CWV via Google PageSpeed Insights (mobile strategy). Expected lift: LCP 7.2s → 1.4s; AUG 0.84 → ~25 (composite drag from Performance=1 → Performance=8 affects 5 other factors via cascade).

Case 4 of 5

The viral-feature-that-isn't

Generator / share tool, 4 months old

AUG Score

0.34

Biggest friction

Advocacy1/10

Acq

5

Act

7

Eng

6

Ret

5

Adv

1

Mon

4

Per

8

Diagnosis

AUG 1.68. Founder built a generator tool ("Generate your X in 30 seconds"). Users complete the action and leave. No share trigger, no embeddable widget, no public-collection URL. The product has the shape of a viral tool — but zero advocacy mechanics. k-factor measured at 0.02.

Recommended fix

Ship per-result share card (Canvas API → 1200×630 PNG + pre-composed tweet). Add embeddable iframe with attribution link-back. Make result URLs bookmarkable with OG preview cards. Expected: k-factor 0.02 → 0.15-0.25 within 60 days; AUG 1.68 → ~6.5.

Case 5 of 5

The zombie

Reference site, 2 years old

AUG Score

0.00

Biggest friction

Retention1/10

Acq

3

Act

3

Eng

2

Ret

1

Adv

1

Mon

2

Per

5

Diagnosis

AUG 0.04. Two years old; 800 sessions/month; D7 retention 2%; zero shares; €3/month revenue. Founder has spent another 60 hours this year tweaking copy + adding small features. Every factor scored low. This isn't a "fix the weakest factor" situation — it's a 90-day kill-criteria candidate per AUG's built-in floor (composite <5 for 2+ weeks = sunset signal).

Recommended fix

Three options: (1) Kill — redirect domain to portfolio anchor; archive the codebase. (2) Pivot — keep the URL, rebuild concept around a different ICP (probably a different archetype). (3) Accept-as-zombie — keep at zero ongoing investment; revisit in 12 months. Most operators try option 4 (keep tweaking) — the framework explicitly rejects this; it's sunk-cost compounding into more sunk cost.

What does your SaaS look like in this format?

Run your own 7-factor audit. 60 seconds. No signup. Get your AUG score + the one biggest friction to fix this week.