Error In Your Favor — continued

User experience
It's not just about the money, though.

The research Barefruit did with consumers also showed that they were generally frustrated with error messages, and often blamed ISPs for them. This combination of a bad user experience and mistaken assignment of blame could obviously have an adverse impact on ISPs.

Even when the Barefruit error pages don't generate any additional revenue, they at least deliver a better user experience and that in itself is worth something, Roberts argues.

Barefruit does press ISPs to give users the choice of opting out. "But not many opt out," he says. "Why would they? Do you really want error pages?"

The key to delivering both the better user experience and the revenue is the Barefruit relevance engine, which resides on its servers, where user traffic is diverted in error situations.

"Relevance equals revenue," Roberts says. "If we can drive relevance, you'll get the revenue following."

The technology is patented worldwide and uses algorithmic and language analysis techniques to derive relevance from user input and text on previously visited pages.

The process is largely automated, but Barefruit also employs "natural language speakers"—aka human beings—in each linguistic market where it has deployments (UK and U.S. English, German and Brazilian Portuguese so far).

They analyze errors received every day, including those where the Barefruit technology had poor success in deriving relevance, and use the information to fine-tune taxonomies embedded in the engine.

"When we first kick off with an ISP in a new territory, on 100 errors, we might get relevance on 40 percent or 50 percent. After a few months, [through fine tuning], it gets to 60 percent to 70 percent."

When the Barefruit technology cannot derive relevance, it presents the user with an error notification and a search box.

The background
So why isn't every ISP in the world using the Barefruit service by now?

The company has been in operation since 2004. Roberts says it has fewer than 20 deployments at this point—that includes trials it expects to go commercial.

Part of the seeming slow start is that it took time for Barefruit to develop and refine its technology and business model, secure funding, establish relationships and build infrastructure.

"But it's also that ISPs, like any large business, take awhile to move forward and make a decision on something like this," Roberts says. "They're quite rightly cautious."

The fastest sale, he says, took 30 days. The longest? Barefruit has been in negotiations with one ISP since 2005.

What often takes time is the legals and negotiation of terms among multiple principals in the deal—Barefruit, the ISP, plus ad search companies.

Deploying the solution is not technologically complex. The work required to implement DNS error interception (where the user mistypes a URL) is very straightforward. "If we had engineers available, we could have it up in a day," Roberts says.

Barefruit recommends ISPs start with DNS error interception. Intercepting HTTP errors—in situations where the URL resolves but is not available—may involve installing hardware. But ISPs don't have to invest any capital in either case, Roberts stresses.

"It's a compelling proposition," he argues. "The revenue comes at no cost to the ISP. Any other way they do business, they'd have to spend $10 to generate $1."

"The other thing is that it's at no risk to them—for two reasons. First, there's a fail-over server, and second, toggling [the service] on or off is under their engineers' control."

No risk. No capital outlay. Free money. Sounds like a no-brainer. But if that's the case, why is Barefruit still in negotiations with one ISP after three years?

— End

< Back to page one