As we’ve written about elsewhere, when people think of a company like Starbucks, they think of it as a single entity. Yet Starbucks is at least 30 legal entities in over 20 countries. It’s this group of legal entities that we call a corporate network of control, or just corporate network for short.
At OpenCorporates, we think mapping out and opening up these corporate networks is one of the most important things we do. Why? Because until we know all about the Starbucks network, we can’t begin to answer all sorts of interesting questions such as: what industries does Starbucks operate in (did you know they operate an insurance company in Vermont?); what territories do they operate in (as well as numerous interesting entities in China they have at least one Cayman Islands company, ‘President Coffee (Cayman) Holdings Ltd.’); what other businesses do they control (something called ‘The New French Bakery Inc’)?
And this information is critical for everyone connected with Starbucks, from their suppliers, to their customers, their employees, shareholders, analysts, regulators, tax inspectors… and so on.
With thanks to a grant from the Alfred P Sloan Foundation, we’ve made a great start on this project, and you have probably seen some of our fascinating visualisations of banks operating in the USA. Now, we’re giving everyone the chance to help in this important task.
How you can help
Whenever possible, we use “scrapers” (software programs which automatically turn bits of web pages into data) to create data. However, scrapers can only work with regular, sensibly formatted information. That’s where you come in.
For example, one of our most important sources is the Securities and Exchange Commission (SEC) in the United States. Large companies that operate in the US must file annual returns with the SEC, and these must include a list of “significant subsidiaries”. Often, the list of significant subsidiaries is presented in a nice, regular format, which we can automatically turn into data:
However, just as often, it’s in a unique format, or just expressed in words. Whenever our SEC scraper encounters a filing that it can’t understand, it saves it to a list of SEC filings to come back to later. It’s in cases like these that we need help:
Over the past few months, we’ve been experimenting with lots of different ways to get help from ordinary users who want to help open up corporate data. It’s really hard to design something that’s simple enough to understand, but nuanced enough to capture sufficient information to be useful.
We’re now ready to release the first of what we’re loosely calling ‘games’. That doesn’t, however, imply flippancy or superficialness (although you do get rewarded with points, and the feedback from our alpha testers has been that they are fun and rewarding).
This first one, the SEC Filings game, tackles the problem we’ve explained above, and while it definitely has a few rough edges, rather than continue to work on it in private, we’re inviting anyone to try it out, and give us feedback.
At the same time, you’ll be creating lots of new open data and helping bring some transparency to the world of corporate networks of control. Have a go, and do send feedback to firstname.lastname@example.org
P.S. The SEC filings are only the start. Starbucks certainly has more than 29 subsidiaries — they’re just the ones designated as “significant” in their 2012 annual return. Email us at email@example.com to become part of our alpha tester programme and help us track down more data or this space!