By Sarah Chen, Senior Financial Analyst. What you'll learn: how to read NDR in the post‑AI spending era (July 2026), spot inflation from one‑off consumption or pass‑through billing, convert NDR into revenue durability, and apply an updated, repeatable quality score to prioritize durable SaaS stocks. This guide is for investors who want numbers, not narratives—especially those tracking public SaaS names or late‑stage private candidates with public‑like reporting.

Why this matters now (July 2026)

Through the first half of 2026 the same forces noted in April accelerated: broad generative‑AI adoption, larger model‑training events, and more explicit consumption pricing. At the same time, institutional investors and some governance advisers are pressing for clearer non‑GAAP disclosures. The numbers tell a different story than management narratives: headline NDR remains useful, but the signal is noisier unless you read the footnotes. This update shows what changed in the last quarter and what to do about it.

Prerequisites and context: NDR basics (refresh)

Net dollar retention (NDR) — sometimes called net revenue retention (NRR) — measures how revenue from an existing cohort changes over a period after accounting for expansions, contractions and churn. In short: does the installed base grow in spend, stay flat, or shrink?

  • Basic formula: NDR = (Starting cohort revenue + expansions − downgrades − churn) ÷ starting cohort revenue
  • Interpretation:
    • >100%: the base expands on net
    • =100%: expansions offset churn
    • <100%: the existing base is shrinking

Gross dollar retention (GDR) excludes expansions and shows how much of the base you actually keep before upsells. If a company discloses both, you can separate stickiness from sales‑driven expansion.

Key limitation: NDR is non‑standard. Definitions vary by cohort (e.g., >$100k ARR customers), timeframe (quarter vs. trailing 12 months), and inclusions (usage, services, FX adjustments). The fine print matters: two companies advertising “NDR 120%” can mean very different things.

Step 1: Pull the number correctly — capture the full definition

  1. Collect: get the most recent quarter plus at least four prior quarters or the prior-year comparable quarter. Trajectory matters more than a single point.
  2. Copy the exact definition verbatim: check whether the metric
    • limits to customers above an ARR threshold (e.g., >$100k)
    • uses quarterly, annual, or TTM measurement
    • includes usage‑based revenue, professional services or pass‑through billings
    • excludes acquisitions, FX, refunds or one‑time contract resets
  3. Capture segmentation: does management disclose enterprise vs. SMB NDR, product splits, or region-level NDR? Blended NDR can conceal weakness in a segment.

Why this matters (July 2026): more vendors now explicitly include consumption charges in expansions. If a company counts spikes from model training or high‑volume inference as recurring expansion without disclosure, headline NDR can be inflated temporarily. Copy the definition exactly so you can re‑compute adjusted NDR if needed.

Step 2: Updated benchmarks and valuation context (July 2026)

Benchmarks are similar to earlier in 2026, but market appetite for durable cash flow has increased. Use these bands as working thresholds (adjust by size and model):

  • Elite: ≥128% — usually mission‑critical platforms with sustained land‑and‑expand and recurring consumption that proves sticky
  • Strong: 118%–127% — supports durable ~20%+ growth with healthy new‑logo activity
  • Healthy/mature: 106%–117% — consistent expansions but reliant on new sales for acceleration
  • Watch: 100%–105% — fragile; growth increasingly depends on new customer adds or price changes
  • Warning: <100% — base is shrinking; requires corrective action

Small percentage points compound. A 5‑point NDR gap (e.g., 125% vs. 120%) translates to ≈4.3% difference in base contribution after one year. Given lower multiples for growth with unclear durability in 2026, investors often re‑price stocks after a multi‑point NDR decline.

Step 3: Normalize for July 2026‑specific distortions — the fine‑print checks

Run these checks before anchoring on a headline NDR.

1) Price increases vs. real usage/seat expansion

Companies often combine list price increases with new packaging. Ask: did customers buy more capability or simply pay higher unit prices? Price‑driven expansion can reverse at renewal. Look for wording such as “price realization,” “deal repricing,” or explicit unit‑price increases in MD&A.

2) AI training and inference spikes

Large customers may incur heavy one‑time consumption during model training or data migrations. Distinguish recurring consumption from episodic spikes by comparing quarterly NDR to TTM NDR and checking whether management calls out “training,” “migration,” “fine‑tuning,” or single large‑customer events.

3) Pass‑through cloud and licensing costs

Some firms now pass cloud spend through as a separate line or inflate revenue while offsetting costs in COGS. If pass‑through is material—treat it like grossed‑up revenue. Practical rule of thumb: if pass‑through + third‑party licensing >8%–10% of revenue, weight margins and FCF more heavily.

4) Cohort and reporting resets

Watch for changes in cohort definitions (enterprise‑only, ARR thresholds) or cadence (quarterly → TTM). Methodology shifts often accompany reorgs, M&A or product splits and can create artificial step changes in NDR.

Step 4: Translate NDR into a revenue‑durability model — practical calculations

Convert NDR into expected revenue from the existing base next year — then decide how much new logo revenue you need to meet guidance.

  1. Existing‑base revenue next year ≈ current recurring revenue × NDR when cohort coverage approximates the reported base.
  2. Example (practical): Company A reports $600M ARR and NDR = 112% (TTM). Existing base implied next‑year contribution ≈ $672M. If guidance calls for 20% revenue growth, management needs roughly $120M of new‑customer revenue (20% of $600M is $120M) beyond the existing base expansion—i.e., new sales must deliver meaningful incremental ARR.
  3. Compare to trend: if NDR fell from 130% a year ago to 112% now, you can quantify the incremental new‑logo burden: the base now contributes $78M less next year than it would have under the prior NDR (1.30×600 − 1.12×600 = $78M). That shortfall explains increased S&M hiring or heavier discounting.

Why this matters: the market increasingly prices expected cash flows, not just growth. Translate NDR movement into dollar shortfalls so you can judge whether guidance depends on reasonable sales economics or risky assumptions.

Step 5: Cross‑check NDR with three companion metrics

NDR is necessary but not sufficient. These companion metrics separate accounting effects from economics.

Companion A: Gross margin (and its drivers)

Gross margins for mature SaaS historically sit in the 65%–80% range. In July 2026, AI‑heavy platforms frequently report lower or more volatile gross margins because hosting and inference costs rose. If NDR rises while gross margin falls >5–10 percentage points year‑over‑year, confirm whether expansion is consumption‑driven and whether the company is absorbing costs.

Companion B: Free cash flow (FCF) margin and Rule of 40

Rule of 40 remains a quick filter: revenue growth % + FCF margin % ≈ health check. A 25% growth business with −5% FCF margin scores 20. For companies marketed as durable compounders, investors increasingly expect positive FCF margin within a 2–4 year horizon. Use FCF margin trends, not just GAAP net income, to test balance‑sheet durability.

Companion C: Billings, cRPO and renewal cadence

Because NDR is backward‑looking, billings and contracted revenue (cRPO) are forward indicators. Shorter contract lengths and a higher proportion of month‑to‑month or consumption contracts increase sensitivity to macro and usage swings. Track renewal timing—multi‑year contracts blunt NDR swings; annual or consumption contracts amplify them.

Step 6: Updated NDR quality score (July 2026)

Run this checklist each quarter. I adjusted the weights to reflect increased importance of disclosure and AI/consumption transparency.

  1. NDR level (0–4 points)
    • 0: <100%
    • 1: 100%–105%
    • 2: 106%–115%
    • 3: 116%–127%
    • 4: ≥128%
  2. NDR trend (0–3 points)
    • 0: down ≥10 points YoY or persistent quarterly declines
    • 1: down 1–9 points YoY
    • 2: flat (±1–2 points)
    • 3: up (sustained)
  3. Disclosure quality (0–3 points)
    • 0: vague definition; inconsistent reporting
    • 1: definition provided
    • 2: definition + segmentation (enterprise/SMB or product)
    • 3: definition + segmentation + GDR reported and clear treatment of consumption/pass‑through
  4. Economic translation (0–2 points)
    • 0: negative and worsening FCF margin
    • 1: near break‑even or volatile
    • 2: positive and improving FCF margin with sustainable gross margin
  5. AI/consumption & pass‑through adjustment (0–2 points)
    • 0: material one‑off consumption or pass‑through included in NDR with no nuance
    • 1: disclosed but mixed signal
    • 2: clean recurring consumption growth, or company provides explicit adjusted NDR excluding episodic events

Score range: 0–14. Interpretation: 11–14 = high quality; 7–10 = mixed; <7 = needs valuation cushion or a catalyst. The extra weight for disclosure reflects investor preference for companies that let you model economics reliably.

Common mistakes investors still make

  • Treating NDR as permanent. Expansion engines can and do mean‑revert. A fall from 140% to 120% over a year is common—what matters is cause and pace.
  • Ignoring cohort definitions. Enterprise‑only NDR can mask SMB erosion and vice‑versa.
  • Confusing “no churn” with “high NDR.” Minimal churn plus negligible expansion yields NDR ≈100%—not impressive for a high‑growth multiple.
  • Overlooking revenue quality and pass‑throughs. If NDR is propped by pass‑through charges, margins tell the economic story.
  • Comparing different business models without adjustment. Usage‑based, seat‑based and transaction‑based SaaS behave differently—normalize to ARR or billings where possible.

Pro tips: get more signal from the same disclosures

  • Quantify dollar shortfalls: convert NDR point changes into $‑terms to see how much new‑logo revenue is required to hit guidance.
  • Monitor S&M per ARR added: rising S&M with falling NDR usually precedes margin compression. Target a payback improvement or a falling sales efficiency ratio within 4–6 quarters.
  • Prefer transparency: companies that report NDR and GDR and split consumption from recurring ARR reduce modeling risk.
  • Stress test valuations: if your DCF assumes NDR stays at 120% for five years, run scenarios at 110% and 100%—small shifts compound quickly.
  • Track sub‑cohorts: product-, geography‑ or customer‑size NDR series often show early signals before the blended number moves.

FAQ

What NDR number is “good” for a SaaS stock in July 2026?

As a rule: ≥118% is strong for companies targeting durable compound growth; 106%–117% is solid for mature businesses; <100% is a warning. Adjust by model: usage‑heavy platforms can legitimately run higher but with greater volatility.

Should I prefer companies that report both NDR and GDR?

Yes. Disclosing both is a sign of better financial transparency. GDR isolates retention from expansion and helps you see whether growth comes from product adoption or pricing/sales tactics.

How should I treat NDR when a company shifts to consumption pricing?

Separate recurring ARR growth from consumption spikes. If consumption is included in NDR, weight gross margin and FCF more heavily and ask management for an adjusted NDR that excludes episodic training or migration events.

Can a company with NDR <100% still be a buy?

Yes, but you need a credible path: improving retention/expansion, demonstrable new‑sales economics, or a valuation that already prices in execution risk. Without that, sub‑100% NDR raises the bar for profitable growth.

How often should I update my NDR analysis?

Quarterly for names you follow actively; update immediately when methodology changes. Track 2–4 quarter trends rather than single‑quarter moves and convert NDR changes into dollar impacts on next‑year revenue.

Sources and context: updated through July 2026 based on public earnings commentary in Q1–Q2 2026, investor presentations, and evolving disclosure trends in the SaaS sector. The metric remains non‑standard; the fine print and companion economics (gross margin, billings, FCF) tell the full story. The numbers tell a different story when you triangulate NDR with margins and cash flow.

Bottom line: NDR remains one of the cleaner product‑market fit signals for SaaS, but post‑AI adoption the metric carries new noise. In July 2026, prioritize clear definitions, cohort splits, and the economic translation to gross margin and free cash flow. The numbers tell a different story when you do the math in dollars, not just percentages.