type: insight tags: [nrr, data-licensing, health-data, moat-validation, expansion-revenue, compounder-thesis] confidence: medium created: 2026-04-01 source: TEM earnings-review Q4_FY25 persona: bear provenance: legacy source_analysis_path: null source_paragraph_quote: null source_transcript_span: null source_loss_log_path: null

NRR Decline in Data Licensing Businesses Contradicts the "Data Flywheel" Moat Thesis

In data-as-a-service businesses where the bull thesis depends on proprietary data becoming more valuable over time, NRR is a direct measure of that claim. When NRR declines while absolute revenue still grows, the growth is being driven by new customer acquisition — not by existing customers expanding their usage. This distinction matters: a business where growth depends on continuous new customer adds is a linear grower, not a compounder, regardless of how large the underlying data asset is.

Evidence

• TEM Q4 FY25: D&S NRR 140% (Q4 FY24) → 126% (Q4 FY25). Absolute D&S revenue still grew +30.9% YoY ($242M → 316M).At140339M — actual was $316M, meaning the existing base may be contracting or new customer contribution is diluted. Either way, the 14pp NRR decline in a single year challenges the thesis that "38M+ records create a defensible expansion moat."
• The divergence (revenue growth ✓, NRR decline ✗) is the tell: headline revenue growth can mask retention deterioration for 1-3 years before the slowdown becomes visible in absolute numbers.

Implication

For any data licensing business (health data, alternative data, geospatial, financial data) where the bull thesis rests on a network-effects or data-flywheel claim:

1. NRR is more important than revenue growth rate — it measures whether existing customers are finding the data increasingly valuable.
2. A one-year 10pp+ NRR decline is a thesis stress test, not a noise event. Demand a structural explanation before upgrading conviction.
3. If NRR drops below ~115% in a data licensing business, the compounder status is lost; model it as a linear revenue grower and discount the multiple accordingly.
4. Watch for the "growing despite declining NRR" phase — it can persist 2-3 years on new customer adds before manifesting in revenue stagnation.