S-1 KPI Disclosure Requirements: Definitions and Evidence
For a pre-IPO CFO or reporting lead, S-1 KPI disclosure requirements raise a practical question: can someone outside the business understand and reproduce the metrics management relies on? Choose the metrics, classify their calculations and build evidence supporting the number and its story.
Key takeaway: Give each proposed KPI an approved definition, comparable history and named evidence owner before the prospectus turns it into an investor expectation.
Which private-company metrics belong in an S-1?
Start with information material to understanding the business. A fundraising deck isn't an automatic prospectus checklist. Regulation S-K Item 303 requires MD&A to explain material financial information and underlying drivers, including statistical data that helps readers understand performance.
Inventory board decks, operating reviews, forecasts and proposed prospectus charts. Ask which decisions each metric informs and what investors would miss without it.
A marketplace might use bookings to explain demand before revenue appears. A fleet operator might use utilization to explain capacity spending. An advertising platform might use engagement to explain monetization opportunities.
Record the inclusion or exclusion judgment. Resolve weak evidence for material metrics; prominence in a VC pitch alone doesn't establish materiality in MD&A.
What do S-1 KPI disclosure requirements expect you to explain?
Explain the meaning, calculation, investor usefulness and management use. The SEC's Release 33-10751, issued January 30, 2020 and effective February 25, 2020, generally expects that context based on the circumstances. Material estimates and assumptions also need attention.
Use this suggested definition card to prepare the disclosure:
- Population and identity: People, accounts, legal entities, transactions or assets? How do you handle duplicates and related entities?
- Qualifying event and window: A visit, purchase, booking or completed service? Daily activity, a trailing period or a closing balance?
- Exclusions and aggregation: Cancellations, test accounts, bots, downtime and averages. State the cutoff and relevant currency treatment.
- Limitations: Estimated inputs, unavailable history and measurement dependencies that affect interpretation.
- Business purpose: The investor question the metric answers and the management decision it informs.
Reddit's February 22, 2024 S-1 illustrates the specificity needed. It defined daily active uniques, or DAUq, using an identifiable unique identifier and a website visit or application opening during 24 hours, including logged-in and logged-out traffic.
Its period average summed daily DAUq and divided by days in the period. That doesn't establish a count of unique natural people; identity, activity and averaging each need scrutiny.
Is an operating KPI automatically a non-GAAP measure?
No. Classify the actual calculation. Item 10(e) excludes operating and statistical measures, and certain ratios composed exclusively of GAAP measures or measures that aren't non-GAAP. Being outside the financial statements, or being expressed in dollars, doesn't settle the classification.
Work through three questions: What are the inputs? Does the formula adjust a comparable GAAP financial measure? Does an operating or statistical exclusion apply?
| Proposed disclosure | Calculation to inspect | Review route |
|---|---|---|
| Active customers | Distinct customers meeting a stated activity threshold | Operating KPI definition and evidence review |
| Transaction value | Transactions, included charges and reversals | Examine the formula and relationship to revenue |
| GAAP revenue per active customer | GAAP revenue divided by an operating customer count | Assess the ratio exclusion |
| Adjusted net income | GAAP net income with specified amounts removed or added | Non-GAAP review |
For a historical non-GAAP financial measure in an S-1, Item 10(e) requires a quantitative reconciliation, the comparable GAAP measure with equal or greater prominence, and the applicable usefulness and management-use explanations. The non-GAAP compliance guide covers that separate review in more depth.
Keep the classification memo with the definition card.
What do Airbnb's marketplace metrics teach about definitions?
Transaction timing and ordinary words can hide different meanings. Airbnb's November 16, 2020 S-1 defined Nights and Experiences Booked as booked nights plus experience seats, net of cancellations and alterations in the period. Gross Booking Value included host earnings, service fees, cleaning fees and taxes, also net of cancellations and alterations.
The filing illustrated a February booking counted in the March quarter and a May cancellation reducing the June quarter. It distinguished that booking timing from revenue recognized upon check-in. A bookings chart therefore needs a timing explanation before readers use it to interpret revenue.
The December 1, 2020 S-1/A glossary treated a listing as active when it was viewable and had previously been booked at least once, excluding Hotel Tonight. Readers shouldn't assume that “active” means a booking within the latest month or year.
These are historical IPO definitions. Airbnb's June 30, 2026 Form 10-Q, signed August 6, uses Nights and Seats Booked and includes services as well as experiences. Peer research should retain the filing date and definition version, not just the metric label.
How should you handle a change in KPI methodology?
Separate operating performance from the measurement change. The SEC's KPI guidance calls for considering material differences, reasons and effects, including effects on previously reported information. Whether prior metrics need recasting depends on the change's significance; the guidance doesn't impose an unconditional recast for every change.
Consider this illustrative scenario, with counts in thousands:
| Comparison | Prior period | Current period | Reported growth |
|---|---|---|---|
| Original definition for both periods | 100 | 120 | 20% |
| Revised definition for both periods | 90 | 99 | 10% |
| Original prior period versus revised current period | 100 | 99 | -1% |
The mixed comparison creates a decline. Preserve both calculations, quantify the bridge and document the comparison supporting the narrative.
Reddit faced a concrete version of this question. In a February 16, 2024 staff letter, the SEC asked about material trend effects of removing detected automated agents prospectively without recalculating earlier user metrics.
Reddit's February 22 response estimated that approximately 0.9 million suspected agents were included in 57.3 million Q3 2022 DAUq. It estimated an impact below 1.6% in each affected quarter, but couldn't accurately measure the historical weekly metric's effect because sufficient telemetry wasn't available.
That was an issuer-specific disclosure-review exchange, not an enforcement finding or a 1.6% safe harbor. The preparation lesson is to preserve the data needed to assess a future filter change, including information supporting weekly aggregation rather than daily totals alone.
What evidence lets a reviewer reproduce each KPI?
Retain the population, calculation version and period-specific inputs. A dashboard screenshot records a result. A useful preparation file also lets a second reviewer trace how that result arose and identify differences between periods.
Use this suggested KPI-to-evidence matrix across business models:
| KPI and management decision | Definition choices to approve | Evidence to retain | Suggested reviewer check |
|---|---|---|---|
| Active customers for account growth | Entity versus account, spend threshold, lookback and parent grouping | Customer-ID mapping, invoice extract, exclusions and historical rerun | Controller checks duplicates and threshold effects |
| Active users for engagement investment | Identifier, qualifying event, daily window and bot filter | Event specification, query version, filter log and historical extracts | Reporting lead reproduces sample days and filter effects |
| Bookings for marketplace expansion | Booking versus completion, fees, taxes, reversal cutoff and currency | Transaction ledger, status history, currency policy and cancellation bridge | Finance checks timing against the revenue explanation |
| Utilization for capacity spending | Available capacity, downtime and averaging | Asset register, usage extract and denominator bridge | Finance checks idle and out-of-service treatment |
| Adjusted profitability for allocation | GAAP starting point and each adjustment | Accounting linkage, adjustment support and reconciliation | Accounting and counsel complete non-GAAP review |
In a completed illustrative customer record, approve “distinct parent customer IDs with at least one paid invoice in the trailing 90 days, excluding internal accounts.” Preserve the invoice snapshot, parent-ID mapping, exclusion list and saved query under metric AC-01, version 2, for September 30, 2026.
The preparer records 1,240 customers. A reviewer reruns the query and inspects duplicate-parent handling; both record names and dates. Add comparative counts, disclosure locations and open exceptions.
A completed file lets a reviewer reproduce the value and understand the comparison. That preparation standard isn't an audit opinion. The evidence and traceability workflow can help organize supporting records.
How should finance, data, counsel and IR review the disclosure?
Review the number and narrative together. Give data or operations responsibility for source completeness and calculation changes, finance responsibility for comparability, and counsel and reporting responsibility for the disclosure treatment. IR should check that proposed investor materials use the approved meaning.
Use a repeatable preparation sequence:
- Approve the definition and classification before drafting charts.
- Freeze period inputs and rerun historical comparisons.
- Resolve exceptions and document methodology changes.
- Review the number, usefulness explanation and limitations together.
- Check MD&A, business, risk factors and investor materials for conflicting definitions.
- Preserve the approved package and require approval for subsequent changes.
This suggested workflow isn't an SEC timetable. Rule 13a-15 addresses disclosure controls for covered Section 12 registrants, with exclusions; its periodic evaluations shouldn't be described as an existing obligation for every private startup.
The distinction still matters for readiness. Disclosure controls address information needed for reports, beyond preparation of GAAP financial statements. A well-controlled ledger doesn't by itself establish the completeness of user telemetry.
Frequently asked questions
Does the financial-statement audit cover every operating KPI?
Don't assume it does. Establish the procedures for each offering statistic with auditors and underwriters; an audit opinion on financial statements isn't a blanket opinion on all prospectus metrics.
Can the comfort letter cover active-user counts?
PCAOB AS 6101, paragraph .55 limits the information accountants should comment on, emphasizing financial information derived from controlled accounting records, with specified exceptions. Ask about the particular metric and procedures early rather than promising comfort on operational logs.
What if historical data cannot reproduce the new definition?
Document what's missing, assess the effect on interpretation and agree the disclosure with reporting and counsel. Don't imply that unreproducible history is comparable; Reddit's correspondence illustrates why the limitation itself matters.
Can we copy a peer's KPI definition?
Use it as a comparison, then test it against your population, systems and management decisions. A peer's filing explains its metric; your evidence file must support yours.
When is a KPI ready for the prospectus?
When the team has resolved materiality, meaning, classification, comparability and evidence, and reviewed the explanation. Put the definition and the reproducible calculation in front of the same reviewer.
