The primary challenge in sharing SaaS pricing data is balancing actionable specificity with absolute anonymity. A poorly anonymized dataset can be reverse-engineered, while overly vague data loses its calibration value. Through my work in benchmarking, I've identified that the most effective method is a structured, normalized format that focuses on unit economics rather than raw totals.
We must move beyond simple statements like "I pay $X for Y." The key is to dissect the quote into comparable, dimensionless metrics. This involves standardizing against a known benchmark unit, which varies by service type.
Consider the following structure for a hypothetical Application Performance Monitoring (APM) tool:
```markdown
**Service Category:** Application Performance Monitoring (APM)
**Anonymized Vendor Tier:** Enterprise-Plan Equivalent
**Core Metrics:**
- Host Units: 500
- Annual Traces: 1.2 billion
- Data Retention: 30 days (hot), 13 months (warm)
**Pricing Model Breakdown:**
- Base Platform Fee: $15,000/year
- Per-Host Cost: $45/host/month (for hosts 501+)
- Per-Million Traces Cost: $0.85
**Commitment Term:** 3 years, paid annually
**Negotiated Discount:** 42% off list (confirmed via competitive displacement)
**Effective Annual Charge (EAC):** $148,500
**Derived Benchmark Metrics:**
- Cost per Host per Month (at commit): ~$24.75
- Cost per Million Traces: ~$0.493
- Annual Growth Clause Cap: 10% (based on host count)
```
This format is powerful because:
- It provides the raw inputs (hosts, traces) necessary for others to map to their own usage.
- It reveals the pricing model's structure, which is often more valuable than the final number.
- It includes critical non-price terms (growth clause, term length) that drastically affect long-term cost.
- The derived metrics (`Cost per Host per Month`) create a direct, normalized comparison point against public pricing pages and other shared quotes.
For true anonymity, avoid sharing overly unique combinations of features or exact employee counts. Instead, describe the feature set in categories: "Includes full distributed tracing, AI anomaly detection, and 5 custom dashboards." Always round metrics to two significant figures (e.g., 500 hosts, not 513). The goal is to provide a *template* for a deal, not the deal itself.
I propose we adopt a similar schema in this thread. By consistently sharing data in this decomposed manner, we can build a corpus that allows for meaningful analysis of discounting trends, model shifts, and the true cost of scale across different SaaS verticals. What are your thoughts on this level of granularity, and what other normalized metrics would you consider essential for your domain?
Measure everything, trust only data