It’s harder than you would think to know how many people work at a company.

I was giving a speech a few years ago and I wanted to know how many engineers worked at Bloomberg LP. I called HR to try and get the number, but it wasn’t easily accessible.

In part this is political. Headcount data at most companies is considered sensitive. It’s also a technology problem because the data has to be stored in a database somewhere.

In the end, I cited data from LinkedIn. At the time Bloomberg had about 22,000 employees globally. Now, LinkedIn says there are 25,209 employees, a 13 percent increase.

The number 25,209 is mind blowingly precise. And LinkedIn breaks it down into a few broad categories. There are 6,526 engineers, 2,958 editorial employees and 2,429 in sales.

Now, for a number of reasons that data will never be 100% accurate and it will differ from the counts companies maintain.(People don’t update LinkedIn profiles in real time and LinkedIn may count someone as an engineer based on their profile while a company may not.)

But worrying about the exact counts misses the big picture.

The data provides a broad directional signal and is probably within spitting distance.
Until LinkedIn began to crowdsource and display it, people relied on quarterly announcements from public companies. Private companies disclosed whenever they wanted.

LinkedIn employment data lifts a veil while at the same time making it harder for companies’ to mislead investors and others. Companies cannot fudge data they don’t control.

The data offers the basis for some incredible insights. It allows you to compare growth over time for a specific company or group of companies in an industry. It allows you to see trends.

For example, Goldman Sachs has promoted itself heavily as a technology company. But according to LinkedIn it currently has fewer engineers (4,774) than Morgan Stanley (4,884.)

It’s possible to leverage those datasets to track in almost real time which firm is expanding and in which area. It’s possible to develop predictive algorithms to anticipate growth.

The data could answer questions companies avoid, such as the gender breakdown overall and by role. There would be no doubt as to whether a company that says it is hiring more women is doing so. (Currently, only about 6% of large U.S. companies disclose gender breakdowns.)

The data could show which companies hire younger versus older workers and where they are located.

It COULD do all of that.

So far, however, LinkedIn only makes the employment insights available as part of its premium subscription and it limits the data to two years of history that cannot be downloaded.

One opportunity LinkedIn has to leverage and capitalize on the data might lay in its expanding news operation run by Daniel Roth.

LinkedIn as a brand is associated with employment. The data could be used to break news about major trends. (Obviously, guidelines would have to be set to avoid abusing the information.)

No one else has this data. And it’s hard to imagine anyone else building it.

I don’t know how LinkedIn manages the database, but I would assume it would require a non-trivial investment to scrub the data and make it more accessible.

It should be possible.

According to LinkedIn, the company LinkedIn is a “technology company” that has 6,489 engineers.