Become a partner
1

Watch what fails

We've spent years cataloguing the operational signatures that precede UK corporate distress, and the cognitive signatures that precede founder regret. The patterns are surprisingly similar — and surprisingly recurring.

2

Encode the patterns

41 stability checks for every iXBRL filing on Companies House. Six dimensions for every business idea. Same methodology: observe the failure, name it, then look for it everywhere it might recur.

3

Surface it before it's too late

For data buyers, the patterns become a sub-5-minute signal feed wired into your stack. For founders, they become a free 5-minute self-assessment. Either way — you see the failure mode before it owns you.

Pick your lens

Two products. Same epistemology.

The Pulse is the production data feed buyers wire into procurement, M&A and quant stacks. The Framework is the free idea-scoring tool that started it all. They share an underlying belief: failure leaves traces, and traces are measurable.

The Stability Engine

Every UK iXBRL filing — ~340,000 a day — runs through 41 named stability checks. Each check is benchmarked against the SIC peer set and auditable back to the source filing reference. Six of the most-used signals below.

01

Current Ratio Trend

Year-on-year movement in current assets / current liabilities. A falling ratio is the single most reliable cliff-edge tell — most defaults show -15% or worse twelve months out. Tracked against the SIC peer-set median so we flag actual decay, not industry-wide compression.

02

Staff-Cost Ratio

Staff costs as a share of turnover, watched for sudden cliffs. A 15%+ headcount cut almost always means the operations team know something the credit bureaus don't — service-delivery failures usually follow inside two quarters.

03

Debt-to-Equity Spike

Sudden D/E expansion — typically >200% YoY — is a covenant-stress signal. Combined with a falling current ratio, it's a textbook pre-administration footprint. Auditable to the exact iXBRL tags driving the calculation.

04

Threshold Crossings

The £10.2M turnover line, £5.1M balance-sheet line and 50-employee line each flip a company between Small / Medium reporting regimes. Each crossing is a same-day signal — and a massive trigger for legal, audit and software upsell.

05

Director Resignation Cluster

Three or more director resignations inside 90 days, weighted by tenure. Catches the classic "rats leaving the ship" pattern weeks before it shows up in the credit file. Cross-referenced against PSC filings for ownership shifts.

06

Cash-Burn Runway

Net cash position divided by trailing 12-month operating loss. Any company below 12 months of runway gets flagged; below 6 months gets escalated. Particularly useful for vendor-risk teams watching loss-making Tier-1 suppliers.

From filing to signal — under four minutes

We sit on the Companies House submission stream. The moment iXBRL lands, the parser, the 41 checks and the signal taxonomy run in sequence — each step auditable, each output traceable to the filing reference.

01

Ingest iXBRL

Direct off the Companies House submission stream. ~14k–22k filings/day, no re-licensed extracts.

02

Extract ratios

Pull the buried tags most credit bureaus throw away — staff costs, current ratio, debt/equity, headcount.

03

Run 41 checks

Each ratio benchmarked against the SIC peer set. Threshold crossings and trend breaks auto-named.

04

Deliver signal

JSON, CSV or webhook. Every record cites the originating Companies House filing reference.

Why AI?

Most people ask friends and family about their ideas. Friends are kind. Family is supportive. Neither is honest. Most data buyers ask credit bureaus about their suppliers. Bureaus are slow. Bureaus aggregate. Neither is fresh.

The same machine that scores ideas across six dimensions parses 340,000 iXBRL filings a day looking for stability signals. Both jobs require the same thing: a system that doesn't have feelings to spare and doesn't have a stake in the outcome.

It just looks at the data, applies the framework, and gives you a straight answer. That's what makes it useful — whether you're stress-testing an idea or stress-testing a vendor.

  • No emotional attachment to your idea — or your supplier
  • Consistent framework, applied identically every time
  • Available instantly — no booking, no quarterly refresh
  • Brutally honest, but never cruel
  • Built from real successes and real failures

Want to go deeper?

Two routes in. Buyers get a sample feed cut to their watchlist within a day. Founders get an instant idea score, no sign-up.