Datadog's S-1 defined its own growth yardstick

The record

Datadog, Inc. filed its Form S-1 on August 23, 2019 ahead of its IPO. The S-1 discloses that its infrastructure monitoring and application-performance-monitoring products are priced per host, a term the filing defines as a server, cloud or on-premise, while its logs product is priced primarily by log events indexed. The filing reports a self-defined metric, dollar-based net retention rate, calculated by comparing annual recurring revenue from the same cohort of customers across a trailing 12 months: 141% as of December 31, 2017, 151% as of December 31, 2018, and 146% as of both June 30, 2018 and June 30, 2019. The filing also discloses a separate dollar-based gross retention rate that it states was in the "low-to-mid 90% range" across each of the last eight quarters before filing.

What the documents establish

The S-1 is Datadog's own account of a usage-based pricing model paired with an issuer-defined expansion metric; the retention figures are self-reported and calculated under a methodology the company itself sets out in the filing, not an independently audited measure. A final amendment, the S-1/A filed September 17, 2019, restates the identical 141%, 151% and 146% figures and sets an assumed IPO price of $25.00 per share, the midpoint of its range, confirming the retention figures were unchanged at the filing closest to pricing. Neither document discloses a dollar-based net retention figure for any period after June 30, 2019.

The operating read

A usage-priced infrastructure company disclosing net retention above 140% is describing existing customers expanding their spend faster than any customer churn reduces it, which is a materially different growth driver than new-logo acquisition. Editorially, a reader comparing this metric across companies should confirm each company defines dollar-based net retention the same way Datadog does, since the calculation window, the treatment of new customers, and whether the figure is point-in-time or a trailing average all vary by issuer, and Datadog's own filing spells out its specific method rather than relying on an industry-standard definition.

What to check before you decide

Before comparing a usage-based company's retention metric to a peer's, check the following.

• Does the filing define its retention metric's calculation window and cohort the same way as the company being compared?
• Is the metric point-in-time, as Datadog's June 2018 and June 2019 figures are, or a trailing weighted average?
• Has the company disclosed a more recent figure in a subsequent periodic filing?

Datadog's own disclosed figures describe a specific 2017-2019 window under a stated method; extending that pattern to a later period or a different company requires checking the newer filing's own definition, not assuming it matches.