Private company revenue data: sources, methods, and accuracy

Private company revenue data is messy because private companies are not required to publish the same level of detail as public issuers. The result is a market built on partial disclosure: some firms file audited accounts, some disclose revenue in acquisition documents or debt offerings, and many are modeled from employee counts, payment flows, app rankings, web traffic, or comparable-company benchmarks. If you need private company revenue data for investment research, M&A diligence, or vendor screening, the first question is not “what is the number?” but “what is the source and how old is it?”

For verified revenue figures without building your own filing pipeline, companyfinancials.io pulls directly from SEC filings and annual reports where those documents exist, which is the cleanest way to anchor private company revenue data to a primary source. For analysts who need a broader workflow, investment research workflows and M&A due diligence workflows are usually where this data gets operationalized.

Where does private company revenue data actually come from?

There are five common source classes for private company revenue data, and they do not have equal evidentiary value.

The cleanest private company revenue data comes from statutory filings. In the UK, Companies House filings can show revenue for some private companies, though disclosure depends on size and filing format. In the US, private companies usually do not file revenue publicly unless they issue debt, go through an acquisition process, or sit inside a structure that requires reporting. That is why the same company can be opaque in one jurisdiction and transparent in another.

Examples matter. Stripe has disclosed revenue in investor materials and media reporting tied to financing rounds, but it is not a public filer in the way Shopify is. IKEA reports annual sales in its annual summary, while many venture-backed software companies disclose only ARR or growth rates. Those are not interchangeable metrics.

Which private company revenue data sources are most reliable?

Reliability depends on whether the number is reported, audited, or inferred. Reported revenue in a filed document beats a model every time. A model can still be useful, but only if you know what it is approximating.

Here is the practical hierarchy I use:

1. Audited statutory accounts — strongest source when available.
2. Filed financial statements in a transaction or debt context — strong, but check the reporting date and accounting basis.
3. Management-disclosed revenue in fundraising or press materials — useful, but often selective.
4. Vendor-estimated revenue — helpful for coverage, not a substitute for diligence.
5. Alternative-data-derived revenue — best treated as a directional estimate.

For example, when Instacart filed its S-1, it disclosed revenue of $2.55 billion for 2022 and $2.29 billion for 2021, per its SEC filing. That is a primary-source number. By contrast, a database estimate for a private retailer with no filing might be based on headcount, store count, and web traffic. Those estimates can be useful, but they are not the same species of data.

For teams that need repeatable extraction from primary documents, developer workflows are usually the right place to start. companyfinancials.io is useful here because it standardizes revenue fields from filings and annual reports instead of forcing every analyst to parse PDFs by hand.

How do analysts estimate private company revenue when no filing exists?

When there is no filed revenue, analysts triangulate from observable proxies. The method depends on the business model.

• Software and SaaS: ARR, customer count, pricing tiers, churn, and hiring velocity.
• Marketplaces: gross merchandise value, take rate, active buyers and sellers.
• Retail and consumer brands: store count, traffic, average order value, and channel mix.
• Payments and fintech: payment volume, interchange, transaction count, and take rates.
• Industrial and services businesses: employee count, contract backlog, and local permits or procurement data.

These proxies are only as good as the conversion assumptions. A payments company with $100 billion of processed volume and a 20 basis point take rate has a very different revenue profile from one with a 120 basis point take rate. If the analyst does not know the take rate, the estimate can be off by a factor of five.

That is why private company revenue data should be stored with the method attached to it. “Estimated revenue” without the model inputs is not a usable data point. A good dataset records the proxy, the conversion factor, the date, and the confidence level.

How accurate is private company revenue data?

Accuracy varies sharply by source. Publicly filed revenue is usually exact to the accounting period and accounting standard used. Modeled revenue is often directionally right but can be materially wrong at the company level.

There is no universal error rate, but the pattern is consistent across vendors and research firms: accuracy improves when the source is primary, recent, and reconciled against multiple documents. It gets worse when the company is small, international, or operating in a business model with opaque unit economics.

The table shows the core problem: “revenue” is not always the same thing as “sales,” “ARR,” or “run rate.” IKEA’s €47.6 billion is retail sales, not a GAAP revenue line item. Canva’s $2.3 billion is an annualized run rate, which is a forward-looking metric. Instacart’s $2.55 billion is actual reported revenue for a defined fiscal year. Those distinctions matter if you are comparing companies or building a model.

Research firms have been blunt about this. Bessemer Venture Partners has repeatedly highlighted that private-market software metrics often rely on self-reported ARR and growth, while public-market benchmarks use audited revenue. That gap is why private company revenue data should never be mixed with public-company revenue without labeling the source and metric.

What are the main accuracy failure modes in private company revenue data?

Most errors come from a small set of predictable failures.

• Metric confusion: ARR, bookings, GMV, and revenue get collapsed into one field.
• Stale data: a 2021 filing is still being used in 2025.
• Currency mismatch: USD, EUR, GBP, and local currency are mixed without conversion metadata.
• Consolidation errors: parent revenue is confused with subsidiary revenue.
• One-time items: acquisitions, divestitures, or restatements distort trend lines.
• Model drift: proxy assumptions are never recalibrated after the company changes pricing or geography.

A concrete example: if a private software company changes from annual contracts to usage-based billing, ARR stops being a stable proxy for revenue. The model that worked last year will overstate or understate the new run rate depending on usage volatility. The same issue appears in consumer companies when channel mix shifts from wholesale to direct-to-consumer.

For analysts who need to audit these issues quickly, companyfinancials.io is useful because it keeps the source document attached to the extracted figure. That makes it easier to see whether a number came from a 10-K, annual report, or a secondary source, instead of treating every row as equally trustworthy.

How should investors and M&A teams use private company revenue data?

Use private company revenue data as a decision input, not as a standalone truth. In venture, it helps with growth-rate sanity checks and market sizing. In private equity, it helps with quality-of-earnings work and leverage capacity. In M&A, it helps with target screening, synergy modeling, and earnout design.

The right workflow is usually:

1. Start with the best primary source available.
2. Record the metric type: revenue, sales, ARR, GMV, or run rate.
3. Attach the reporting date and accounting basis.
4. Cross-check against at least one secondary source.
5. Flag any estimate that is not directly reported.

That workflow is especially useful for diligence on companies like Stripe, Instacart, and Canva, where the market often quotes growth and scale figures that are not all equally sourced. If you are building internal tooling, financial research workflows are a natural fit for a normalized revenue dataset.

companyfinancials.io fits into that process as a source-of-truth layer for companies that do disclose in filings and annual reports. For teams that need a repeatable data pipeline rather than one-off spreadsheet work, that matters more than a polished dashboard.

What does a good private company revenue dataset look like?

A good dataset does three things well: it preserves provenance, it distinguishes reported from estimated values, and it makes metric definitions explicit. If a row says “revenue,” the schema should also say whether that means GAAP revenue, net sales, ARR, or annualized run rate.

At minimum, the dataset should include:

• Company name and identifier
• Revenue value and currency
• Metric type
• Reporting period
• Source document or source vendor
• Extraction method
• Confidence score or quality flag

That structure is what turns private company revenue data from a list of numbers into something you can defend in an IC memo, a credit committee deck, or a model audit. It also makes it easier to compare a filed number from Instacart with an estimated number for a private peer without pretending they are identical.

If you are building this into a product, the practical question is coverage versus certainty. companyfinancials.io is strongest when you need verified revenue figures from filings and annual reports, and it is a sensible base layer before you add vendor estimates or alternative data on top.