The firewall vs. the cloud

I was a little surprised when the Twitter-for-the-Enterprise service Yammer won the TechCrunch50 Best of Show award in 2008. Not because Yammer is a bad app or because I think the idea of a business Twitter is silly (I think it makes a lot of sense, in fact), but because as an enterprise service, Yammer seemed tone deaf in one key area: it was released as a hosted service and didn't give business customers direct control of the records of conversations that their employees might be having.

The hosted model makes financial sense. It's much cheaper to contract with a service like Yammer than to buy and install software inside a business. However, it's not realistic to think that the people inside businesses currently running their own e-mail servers will happily encourage sensitive and timely employee conversations to head outside the firewall to a hosted service. Saving money doesn't trump control.

So it makes sense that Yammer is now available as installable software. Users of this version of the service will pay more--they have to provide their own hardware, backups, and, most importantly, staff to support the app--but they get the control over information that itchy IT types like so much.

Another business Twitter-alike product, Present.ly, launched in 2008 with support for an installed version initially, as well as a hosted service like Yammer. Yoshi Maisami of Intridea, which makes Present.ly, told me his company is making more money from the inside-the-firewall version than from the hosted service. He has more accounts on the hosted app though.

In related news, Egnyte, which offers a hosted file server for businesses, now has a version it calls the "Local Cloud," which puts the data that would otherwise live solely on Egnyte's servers on machines that businesses can install inside their firewalls.

Egnyte is still designed around a the idea of hosted file storage, and the Local Cloud product synchronizes with hosting servers. But the local option does at least give customers access to their data if their Internet connection is disabled.

Not all Web 2.0 start-ups are seeing the need to offer local access to their services, though. I asked Jennifer Zeszut of the market analytics company Scout Labs (review) if any of her customers were asking for a local version of the product. She said they're not.

The difference is the nature of the data being stored, and from where they've traditionally been controlled. Employee communications and file storage come from within a company. Infrastructure, habits, and laws have grown up to encourage that data to stay there. But for online services that undercut traditionally consulting-based services, like measurement analytics, there's little business instinct to pull the entire operation inside the arms of the IT department.

There's also, of course, the issue of perceived reliability. If an e-mail or file storage service goes offline, a business can grind to a halt. While I don't believe that IT departments are inherently able to provide more uptime than hosted services, a business leader at least knows who he or she can complain to if the company's e-mail suffers a glitch. With analytics services, a few minutes or hours of downtime is unlikely to have an immediate material impact on a company's capability to operate.

It's not surprising to hear of start-up companies that provide Web-based services to their customers as being enthusiastic consumers of Web-based services themselves. For my part, I'm happy to build the Webware 100 awards program using services from Webware companies like Wufoo and PollDaddy. But to sell to businesses in the real world, it's not enough to offer a Web-based service that's less expensive than a traditional software app. You've got to understand the habits and laws, the sacred cows, and the fears of customer companies. Local versions of apps that address these issues may be less efficient and more expensive. But they can also sell better.