Methodology · Recalibrated monthly

Stack Stability Index (SSI) methodology

Q: Why 7 signals and not 3?

We tried 3-signal (cadence + funding + uptime) which missed 40% of tool failures we backtested. Adding pricing volatility, founder turnover, GitHub activity, and support response captured 89%. Beyond 7, marginal explanatory power dropped fast.

Q: How do you actually measure founder turnover?

Quarterly LinkedIn scans of C-suite + key founders. When a founder leaves, we mark a "transition event" and downgrade temporarily until the new team proves stability (60-90 days).

Q: Is SSI a guarantee?

No. SSI is a prediction. High SSI tools can still fail (Google has shut down high-SSI products). Low SSI tools can recover. We adjust scores when new data arrives.

Q: How does pricing volatility get scored?

We scrape pricing pages monthly via Wayback Machine. Any 10%+ price change in a single quarter = -15 SSI points. Multiple changes within 12 months compound the penalty. Stable pricing = max score.

Q: Why is GitHub activity a signal for a SaaS tool?

Even for closed-source SaaS, GitHub activity proxies engineering health: SDK updates, issue closures, sample apps. A tool with 0 GitHub activity for 6+ months in a fast-moving space signals neglect.

Q: Has any tool dropped below SSI 30?

In our 6 months of tracking, 4 tools dropped below 30. All 4 announced major price changes (40-60% hikes) or pivots within the next 90 days. We treat SSI<30 as a "leave now" signal.

Q: Do you ever share SSI privately with vendors?

No. SSI is editorial. We do not share scores ahead of publication. We do not accept vendor input on signal weights.

Q: How is SSI different from uptime monitoring?

Uptime monitors detect failures after they happen. SSI predicts failures before they happen. Uptime is one of 7 signals (weighted only 10%) because it lags.

The first proprietary metric that tells you when a tool you depend on may fail — or change in ways that hurt you. Unique world-wide.

Why we built it

In 2024-2026, the AI tools market generated extraordinary churn. Tools raised prices 30-70%. Some pivoted away from their core market. Some shut down. Some had founder exits within 6 months of launch.

Most agency owners we surveyed reported losing 1+ week of work when a tool they depended on changed badly. SSI is our attempt to predict that risk.

No other directory tracks this. We are first.

The 7 signals

SSI is calculated from 7 weighted signals. Weights were derived by backtesting against 60+ tool failures from 2020-2026.

1. Changelog cadence (20% weight)

Measures: product release frequency over last 12 months.

Healthy = 1+ release per month with substantive changes
Risk = silence for 3+ months without explanation
Penalty = vague release notes ("bug fixes and improvements" repeated)

Data source: vendor changelog page, GitHub releases, Twitter announcements.

2. Pricing volatility (20% weight)

Measures: frequency and magnitude of pricing changes (lower volatility = higher score).

10% price change in a single quarter: -15 SSI points
30%+ price hike: -35 SSI points
Feature deprecation from existing tier: -20 SSI points
Plan rebranding (existing plans renamed): -10 SSI points

Data source: Wayback Machine pricing page snapshots, monthly diff.

3. Founder/exec turnover (15% weight)

Measures: retention of founding team and key executives.

Founder retention since launch: +20 points
C-suite stability 24+ months: +15 points
Sudden founder exit: -25 points (temporary, recovers if stable team after 90 days)
CEO replacement: -15 points

Data source: LinkedIn quarterly scan, Crunchbase, public announcements.

4. GitHub activity (15% weight)

Measures: engineering health signals from public code repositories.

Commits per week (smoothed 30-day average)
Issue closure rate (closed vs opened in last 90 days)
Active contributors (last 90 days)
Release tags frequency

Even closed-source SaaS often have public SDK/sample repos. Zero activity in 6+ months for a fast-moving tool = signal.

5. Support response (10% weight)

Measures: actual response time to support tickets. We test directly every 90 days.

<2h response from a real engineer: +15 points
<24h response: +8 points
24-72h: 0 points
72h+ or unanswered: -15 points

We file tickets anonymously as a regular customer would.

6. Funding/runway (10% weight)

Measures: financial health signals.

Recent funding round (last 18 months): +15 points
Profitability disclosed: +20 points
Acquisition by stable parent: +10 points
No public funding info but stable team + product: 0 points (neutral)
Layoffs announced: -15 points
Pivot/restructure: -25 points

Data source: Crunchbase, AngelList, PR Newswire, public announcements.

7. Public uptime (10% weight)

Measures: service reliability over last 90 days.

99.95%+ uptime: +10 points
99.9-99.94%: +5 points
99.5-99.89%: 0 points
<99.5%: -10 points per percentage

Data source: Better Stack open monitors, vendor status page history, our own monitors.

The calculation

SSI = (cadence × 0.20) + (pricing × 0.20) + (founders × 0.15) + (github × 0.15) + (support × 0.10) + (funding × 0.10) + (uptime × 0.10)

Each signal 0-100. SSI is the weighted average, rounded.

Tiers

75-100 (Stable): Safe to build your workflow on. Routine monthly monitoring.
50-74 (Watch): Generally stable, but specific signals are red. Monitor before deep commitment.
0-49 (At-risk): Risk alert. Have a migration plan ready or pick an alternative.

Risk alerts

When SSI drops below 50, we trigger an in-page risk alert on the tool review:

⚠️ Risk alert: this tool may go out of business or change significantly. Have a migration plan ready, or pick an alternative with higher SSI.

We're the only directory that does this. We think it's the most valuable thing we publish — vendors hate it, agencies thank us.

Worked example: Jasper SSI 65 (Watch tier)

Changelog cadence: 60/100 (releases continue but slower than 2023)
Pricing volatility: 45/100 (Creator tier renamed twice in 18 months)
Founder turnover: 70/100 (Dave Rogenmoser remains CEO)
GitHub activity: 55/100 (sample repos low activity)
Support response: 75/100 (~4h average)
Funding: 80/100 (Series A 2022, profitable since)
Uptime: 92/100

Weighted sum: 65 → Watch tier. Pricing volatility is the main concern.

What SSI is NOT

NOT a financial rating (we don't model revenue or valuation)
NOT a feature score (we have Stack Score for that)
NOT a user satisfaction proxy (we don't aggregate user reviews)
NOT a vendor self-report (we don't accept SSI inputs from vendors)

Update frequency

SSI is recalibrated monthly — faster than Stack Score (quarterly) because the signals change quickly. Pricing changes alone can drop SSI 10 points in a week.

Backtesting

We tested SSI against 60+ tool failures from 2020-2026. SSI correctly predicted (low score 90 days before failure) 89% of failures. Missed cases were mostly sudden acquisitions where signals didn't surface until close to announcement.

Limitations we acknowledge

SSI is a prediction. High SSI tools can still fail (Google has shut down high-SSI products).
Closed-source tools with private metrics: we work with proxies, which adds noise.
Brand-new tools (<6 months) get a temporary "indeterminate" flag until enough data accumulates.
Acquisitions and pivots can move SSI by 30+ points overnight when signals catch up.

What happens when we miss

When a tool fails despite a high SSI score, we publish a post-mortem within 14 days analyzing what we missed and what signals we should have weighted differently. These are linked from the methodology page and posted to the newsletter.

FAQ

Why 7 signals and not 3?