History is rife with examples of companies wandering away from their core expertise and bungling things terribly. Colgate once offered kitchen entrees, Coors decided to sell bottled water alongside its beer, and ESPN tried to sell a mobile phone.

It’s easy to shake our heads at these mistakes. But if the dangers are so obvious, why is diversification so rampant among tech companies? Amazon, Google, and Microsoft often attempt feats outside their core expertise. Now we should add to that list Facebook, which recently announced it would launch Facebook at Work.

Tech companies are no different from other companies, and have seen mixed results when moving away from their expertise. Amazon is well known for its “insatiable diversification push,” and that may be haunting it now; Mashable and others have already connected Amazon’s diversification with the company’s recent stock price drop. In contrast, Apple has mostly stuck to doing a few things very well. Considering Apple’s astronomic stock prices, you’d think other tech companies would be taking note.

Not Facebook. Facebook at Work promises a radical departure for the company. While battling LinkedIn to become the main work-focused social network would make sense for Facebook, that’s not what Facebook is planning. It wants to offer an enterprise collaboration solution.

Facebook at Work would look like a personal Facebook page but be completely separate, for security reasons. It would offer a way for coworkers to get together virtually, discuss projects, and potentially share documents. What could go wrong?

We can see exactly what by looking at Cosmopolitan yogurt. In 1999, the successful women’s magazine brand decided to sell yogurt. For Cosmo, the new market was crowded, as it will be for Facebook. Just as grocery stores have no shortage of yogurt brands or varieties for sale, in the enterprise market chief information officers can choose from collaboration solutions from a host of vendors, in both on-premises and software-as-a-service varieties. Cosmopolitan knew that readers of its magazine enjoyed and purchased a lot of yogurt. Facebook knows that its users are both familiar with and like using the Facebook interface. The next step seems obvious, right?

But Cosmopolitan yogurt was a failure, and had vanished from shelves within 18 months. That’s because a customer’s loyalty to a brand is based on that brand’s expertise. To readers of Cosmo, there’s no better place to get makeup tips or diet advice, but readers knew that had nothing to do with food manufacturing. People are sophisticated enough to know that expertise in one area doesn’t mean expertise in another.

At first glance, you might think that Facebook won’t have this problem. It’s still making software, after all. But just because Facebook can make a platform that lets users share everything doesn’t mean it can make a platform that supports workplace collaboration. In fact, to CIOs and other IT officials, that very share-everything history makes Facebook unsuited for the corporate world.

A strong brand seems to make a company believe it can do anything. But as the history of failed product launches and dim-witted diversification attests, brands aren’t all-powerful symbols. We don’t know what the future will bring for Facebook at Work, but we do know that when companies venture away from their expertise, the market often ends up laughing.

Tom Bice is vice president of product marketing and sales enablement at Attachmate and Novell.