Heaven and Companies House
Recently I have been asked about using private company public data information for financial purposes (which could range from credit scoring to potential private equity investment).
In the UK all companies have to register at the Companies House in a process that usually takes minutes, but:
“At the moment, the Companies House register includes almost 4.5m UK businesses but it operates in much the same way it did 150 years ago — meaning criminals have been able to set up seemingly legitimate shell companies without the most basic identity checks.” (Directors to face ID checks in antifraud crackdown, FT)
Just to give you a flavour of the data available, read Cynthia O’Murchu’s virtual address squatting tale in Companies House, my virtual Russian squatter and me.
She found a company (Weight a Minute) that has a director whose country of residence is Heaven and occupation is Creator (see the screenshot below, in case the links stop working once the face ID is finally enforced)

The 4.5 million companies also have to register financial information (company accounts) every year, and since April 2011 the data is supposed to be filed electronically (using XBRL). The deadlines are very flexible, ranging from 9 months to 21 months (for new companies).
My first reaction given all the above background is that cleaning and making sense of this data is a Titanic job (in both senses, as a job fit for a Titan or destined to end like the Titanic), but given that Companies House make it very easy to access the data (API) we gave it a go (in collaboration with Worksheets Systems).
We:
- downloaded XBRL accounts,
- developed a dashboard using published Turnover,

What we immediately noticed is that for such a basic field as Turnover (or Revenue) we have only been able to identify about 70k companies that publish the accounts in a standard (i.e. easy to parse) form (just a little bit less that 2% of the 4.5mn companies — the other 98% seem to happily live under the radar).
Of course, may companies seem to be happy just to send pdfs, and some send just the minimum possible set of data required for the filing. We will keep you posted once we make inroads into this dataset.
Meanwhile, by next year some interesting macro back analysis could be published using the data from the Revenue of companies hit by Covid (big positive or negative change of Turnover). For many investors this analysis will be way too late, but some others might find some nuggets.
Feel free to contact us if there is something special you would like to see.
