Net Dollar Retention

How net dollar retention works

Net dollar retention answers one question: taking only the customers who were already there at the start of the period, what did they pay by the end of it?

Written in words, net dollar retention is revenue in the current period from the customers who existed at the start of the prior period, divided by what those same customers paid in that prior period.

NDR = Current-period revenue from the starting cohort / Prior-period revenue from that same cohort

Expressed on a monthly recurring revenue basis, the same thing is usually written as an expansion-and-contraction bridge:

NDR = (Starting MRR + Expansion - Contraction - Churn) / Starting MRR

The two forms agree only when the cohort and the revenue base are defined identically. That is the part that goes wrong in practice.

The choices that change the answer

• The cohort. Snowflake's fiscal 2026 annual report on Form 10-K describes a measurement period of the trailing two years, a cohort of customers under capacity contracts who used the platform at any point in the first month of the first year, and a calculation that divides the cohort's product revenue in the second year by its product revenue in the first year, with customers who stopped using the platform remaining in the calculation and contributing zero.
• The revenue base. Recognised revenue, annual recurring revenue, monthly recurring revenue, committed contract value, and consumption revenue all produce different numbers for the same business.
• The treatment of churn. A definition that drops fully churned customers from the denominator produces a flattering figure. A definition that keeps them in, contributing zero, does not.
• The period. Trailing twelve months, trailing twenty-four months, and quarter-annualised all appear in public filings.
• The currency and acquisition adjustments. Snowflake's filing states that its net revenue retention rate is subject to adjustments for acquisitions, consolidations, spin-offs, and other market activity, and that historical periods are restated to reflect them.

Variants

• Gross dollar retention, which excludes expansion and therefore cannot exceed 100 percent. It isolates leakage.
• Logo retention, which counts customers rather than dollars.
• Dollar-based net expansion rate, which is the same construct under a different label.
• Cohort-weighted net retention, which reports the figure separately by customer size or segment rather than as one blended number.

Worked example

Suppose two companies both report 120 percent net dollar retention and both start the year with $20,000,000 from existing customers.

Company A keeps every customer. Nothing churns, nothing downgrades, and the base expands by $4,000,000 through seat growth and upsell. Ending cohort revenue is $24,000,000. Net dollar retention is $24,000,000 divided by $20,000,000, or 120 percent. Gross dollar retention is $20,000,000 divided by $20,000,000, or 100 percent.

Company B loses a third of its base. Churn removes $6,000,000 and downgrades remove $1,000,000, leaving $13,000,000. Its twenty largest accounts, however, expand by $11,000,000. Ending cohort revenue is $24,000,000. Net dollar retention is also 120 percent. Gross dollar retention is ($20,000,000 minus $6,000,000 minus $1,000,000) divided by $20,000,000, or 65 percent.

The two businesses are not remotely alike. Company A compounds. Company B is a concentrated expansion story sitting on a leaking base, and if a single large account stops expanding, the headline number collapses. This is why the pairing matters: net dollar retention without gross dollar retention describes half the picture. All figures are hypothetical.

Where it shows up

In public company filings, net dollar retention appears in the Key Business Metrics section of the annual report on Form 10-K and in the corresponding section of the quarterly report on Form 10-Q, usually in a table alongside customer counts and remaining performance obligations, with a narrative definition immediately beneath. Snowflake reported a net revenue retention rate of 125 percent as of January 31, 2026, 126 percent as of January 31, 2025, and 133 percent as of January 31, 2024, and disclosed in the same filing that it expects the rate to decline over the long term as long-tenured customers become a larger share of the base.

In the Securities and Exchange Commission's 2020 guidance on key performance indicators in management's discussion and analysis, the Commission states that it would generally expect a metric to be accompanied by a clear definition of the metric and how it is calculated, a statement of why the metric provides useful information to investors, and a statement of how management uses it in managing or monitoring the business. The guidance adds that a company changing the method of calculation should consider disclosing the difference and the reasons for the change, and that a metric should not deviate materially from the metrics management actually uses to run the company.

In private company reporting, the figure appears in the monthly or quarterly investor update and in the metrics tab of a diligence data room, typically with the underlying cohort export attached. In a growth equity or late-stage term sheet, the metric shows up indirectly, through the information rights clause: the NVCA model term sheet provides that Major Investors receive annual and quarterly financial statements and other information as determined by the board, which is where a negotiated metrics package usually lands.

In a limited partner update, portfolio-level retention figures appear in the company commentary section rather than as a defined reporting line. The Institutional Limited Partners Association's reporting templates standardise fund-level cash flow and fee reporting, not portfolio company operating metrics.

Common mistakes

• Reporting net dollar retention without gross dollar retention. Expansion from a few accounts can mask severe churn, and the two numbers together identify which business you are looking at.
• Changing the cohort definition between periods. A year-over-year comparison is meaningless if the denominator was built differently, and the SEC's MD&A guidance specifically directs registrants to consider disclosing calculation changes.
• Including new customers. Anyone who signed after the cohort start date belongs in growth, not retention. Including them inflates the figure and destroys its diagnostic value.
• Blending self-serve and enterprise. Small-account churn and large-account expansion net out to a number that describes neither segment.
• Reading a high figure as proof of product quality. Multi-year contracts, usage-based pricing, and annual uplifts all raise net retention mechanically without any change in customer behaviour.
• Comparing across business models. A consumption-priced infrastructure business and a seat-priced application business generate structurally different retention profiles, so the same percentage does not carry the same meaning.

Related terms

Net dollar retention is read alongside rule-of-40 when investors assess whether growth is efficient, and it is one of the metrics that drives pricing at series-b-funding and later rounds, especially in growth-equity deals. It shows up in diligence through the data room and in ongoing reporting to a limited-partner-lp through portfolio commentary. Investors screening a cap-table for follow-on exposure use it to decide whether to exercise pro-rata rights.

Frequently asked questions

What is the net dollar retention formula?

Divide the revenue generated in the current period by the cohort of customers who existed at the start of the prior period, by the revenue that same cohort generated in the prior period, then multiply by 100. The equivalent bridge form is starting recurring revenue plus expansion minus contraction minus churn, all divided by starting recurring revenue. Both exclude customers acquired after the cohort start date.

What is a good net dollar retention rate?

Any figure above 100 percent means the existing base grew. Beyond that, what counts as good depends on the business model, the contract structure, and the segment, and published targets are convention rather than measured benchmarks. The more useful test is the pairing: strong net retention next to strong gross retention indicates a compounding base, while strong net retention next to weak gross retention indicates concentration.

What is the difference between NDR and NRR?

Nothing substantive. Net dollar retention, net revenue retention, dollar-based net revenue retention, and dollar-based net expansion rate are labels for the same construct, and public filers use them interchangeably. What differs between companies is the cohort and revenue-base definition underneath the label, not the label itself.

Can net dollar retention be above 100 percent while the company is shrinking?

Yes, if new customer acquisition has stalled and the existing base is expanding off a smaller starting point, or if a small number of large accounts are growing while many small ones leave. Net dollar retention describes only the existing cohort. Total revenue growth also depends on new logos, which the metric deliberately excludes.

How does net dollar retention differ from gross dollar retention?

Gross dollar retention counts only losses: churn and downgrades, never expansion. It therefore cannot exceed 100 percent and measures how much of the base leaks away. Net dollar retention adds expansion back, so it can exceed 100 percent. Reported together, gross retention tells you about leakage and the gap between the two tells you about expansion.

Do private companies have to define the metric the way public companies do?

There is no obligation, but the discipline is worth adopting early. The SEC's MD&A guidance sets the expectation for registrants that a presented metric carry a clear definition, a reason it is useful, and a statement of how management uses it. Private companies that write that definition down before diligence avoid restating the number midway through a financing.