Skip to content
MarketPrior

Company & founder diligence

Company due diligence, cited line by line.

Name a company or founder. MarketPrior screens sanctions, litigation and adverse media, reads financials, ownership and traction, and returns one risk × upside verdict. Every finding links to the document it came from.

Run a free screenSee report pricing

The verdict

Risk and upside, scored separately, then fused.

The risk score and the upside score are computed from different evidence, then combined into one of four verdicts. The drivers behind each score are listed, and a report with missing inputs is marked partial.

  • STRONG & CLEAN

    Low risk, strong upside signals.

  • PROMISING BUT RISKY

    Real upside, with findings you need to read first.

  • CLEAN BUT QUIET

    Nothing adverse, and little evidence of momentum.

  • AVOID

    Risk outweighs upside. Shown as “AVOID — SANCTIONED” only when a sanctions hit matches on an identifier.

Risk

Screening that knows the difference between a name and a match.

A fuzzy name hit treated like a confirmed one either buries a real sanctioned entity in noise or brands an innocent company. So every hit is tiered, and weak ones are shown for review instead of scored.

01

Sanctions, PEP and disqualified directors

OFAC, EU, UN and UK lists, the UK disqualified-directors register and a PEP screen, run on the company, its officers and its owners. Every hit is tiered confirmed, probable or possible against the identity we hold.

02

Litigation, weighted by role

Enforcement releases, agency actions and court dockets. The subject’s party role is read from the caption: suing to enforce its own contracts is not a red flag, being a defendant is. Nature and status set the weight.

03

Adverse media, scored for materiality

News is classified for whether it is about the subject, its sentiment and its materiality. Only material, negative items about the subject count as risk; positive coverage feeds upside instead.

Profile & upside

Financials, ownership and traction from primary sources.

Filings and registries first, then public signals of momentum. Competitors are listed alongside, and each section shows how fresh it is.

01

Financials

Reported figures from SEC XBRL company facts, ESEF filings and national registries such as Companies House accounts.

02

Ownership & control

GLEIF parent and subsidiary links, UK persons with significant control and SEC 13D/13G holders above 5%. Investors are inferred from SEC Form D director and promoter seats and named participants. That is public record, not a cap table.

03

Traction signals

Open roles on public job boards (Greenhouse, Lever, Ashby and others), Tranco rank and Wayback history for the website, detected tech stack, GitHub activity, Hacker News mentions and App Store listings.

Every source, with its licence and last health check, is on the data sources page.

Cited memo

A written memo with no uncited sentences.

The memo is checked sentence by sentence before you see it. A sentence without a valid citation is dropped rather than shown.

  • Each claim links to the filing, register entry, docket or article behind it
  • Export as PDF, HTML or Markdown
  • Share a read-only link that expires and can be revoked
  • The same report is available over the REST API and MCP

Example

What the top of a report looks like.

An illustrative layout with placeholder values. It is not a real company or a real result.

Example Co.illustrative
sections refreshed · today

Verdict

PROMISING BUT RISKY

Risk 58 / 100 · Upside 71 / 100

Top risk

Defendant in an open contract dispute

[1] court docket

Top signal

Open roles up over 90 days

[2] public job board

ScreenResultBasis
SanctionsNo confirmed or probable hitOFAC · EU · UN · UK
Officer screen1 possible match, for review (not scored)name only
ForecastFollow-on within 18 months, with an 80% rangeForm D model · backtested

Outcome forecast

Will it raise again, or go quiet?

Two calibrated probabilities for companies with SEC Form D history: new money reported on a Form D within 18 months, and no further Form D within 24 months. Each comes with its top drivers and an 80% range.

0.83

AUC, follow-on within 18 months

Backtest · 2015–2024

0.20

Brier skill vs base rate

Backtest · 2015–2024

≈0.01

Expected calibration error

Backtest · 2015–2024

3.3×

Follow-on rate in the top decile vs average

Backtest · 2015–2024

How the backtest was run, and its limits

Rolling yearly out-of-time test years 2015–2024 on the complete public SEC Form D data sets (US operating issuers, pooled investment funds excluded; about 301,000 held-out company-year observations). A model on time since the last filing alone reaches AUC 0.69, so recency does much of the work. On startup-like companies alone the AUC is 0.73. The goes-quiet model scores AUC 0.83 on the same protocol.

  • Outcomes are Form D filings, not company success: a company can raise without filing a Form D, or stay healthy without raising.
  • Companies funded only through instruments that never file a Form D are under-represented.
  • Hiring, web and tech-stack signals are not in this model yet.
  • Probabilities are calibrated on the most recent fully observed year; if the funding climate shifts, calibration can drift.

Free screen, full report

Screen any company for free. Pay for the reports you need.

The verdict, risk score and upside score are always free. The full report unlocks per company, and re-running a company you have unlocked costs nothing.

  • A full run usually finishes in well under a minute (median about 25 s in our internal benchmark runs)
  • Each section can be re-run on its own when it goes stale
  • Add the company to a watchlist to keep it current
Compare diligence plans

Coverage limits

What a report cannot tell you.

MarketPrior reads public records and public web signals. Where they are thin, the report says so.

Private-company people and owners

Officers and ownership come from what registries publish. For private companies outside the UK, this is often partial.

Some EU registries

A few national registries charge for data or block automated access. Those gaps show as missing inputs, not guesses.

Decision support, not a decision

Probable and possible matches need a human check. The report shows the evidence so you can make that check.

Related

Product

Continuous monitoring

Re-screen a watchlist on a cadence and get alerts on new material events.

See monitoring

Product

Trigger-based leads

Find companies from new raises, awards, registrations and officer changes.

See lead generation

Data

Data sources

Every registry, list and feed the report reads, with licence and status.

Browse sources

Developers

Diligence API & MCP

Run diligence and read every section from your own code or agent.

Read the API docs

FAQ

Due diligence report questions

What is in a company due diligence report?
A fused risk × upside verdict, then the evidence behind it: sanctions, PEP and director-disqualification screening; litigation with the subject’s party role; adverse media; financials; ownership and control; investors inferred from Form D; hiring, web, tech-stack and developer signals; competitors; an outcome forecast; and a memo in which every sentence cites a source.
How do you avoid false sanctions matches?
Each watch-list hit is tiered against the identity we hold for the subject. Confirmed means an identifier match (LEI, company number, CIK or list id, or for a person date of birth plus nationality or exact name). Probable means an exact name or listed alias plus a corroborating attribute. Possible is anything weaker: it is shown for review and never scored. The verdict reads “AVOID — SANCTIONED” only on a confirmed match.
What does the free screen include?
The verdict, the risk score and the upside score, for any company. The remaining sections (litigation detail, financials, ownership, investors, traction, forecast, memo and exports) unlock per company. Re-running a company you have already unlocked is free.
How accurate is the forecast?
On a rolling out-of-time backtest over 2015–2024 using the complete public SEC Form D data sets, the follow-on model scored AUC 0.83, Brier skill 0.20, calibration error about 0.01 and 3.3× lift in the top decile. These are backtest figures, and the outcome is a Form D filing, not company success. On startup-like companies alone the AUC is lower (0.73).
Does it cover private companies and non-US companies?
Yes, with thinner coverage. Officers and ownership for private companies depend on what registries publish, so they can be incomplete. Some EU registries are paid or block automated access, and those gaps are shown in the report rather than filled with guesses. The data sources page lists every source and its status.
How long does a report take?
A full run usually finishes in well under a minute. In our internal benchmark runs the median was about 25 seconds per company.

Run due diligence on your next company.

Start with a free screen: verdict, risk score and upside score, each backed by its sources.

Run a free screenSee pricing