When is a Website Not a Website?

Setting a New Standard for Self-Service

Not surprisingly, the convergence of customer support with a more analytic/business-process-driven interaction has attracted different breeds of solutions to try for a seat at the table. One of the people I talked with, Todd Paoletti, director of customer self-service solutions at Actuate, explained how a business analytics company wound up talking about Web self-service:

"Enterprise reporting—which is what we do—is essentially about accessing data from a variety of data sources, putting it together and distributing it out to a large number of end users in a very intuitive, digestible way, usually for day-to-day use. That's compared with business analytics and data mining, where you have a lot of crunching of data to help a relatively few members of an organization map and model where to take the business next," began Paoletti.

"With enterprise reporting, you usually deliver information to end users within the organization to help them run the business. But we can also deliver that information outside the organization—clients customers, partners, agents...We're essentially a development platform that companies use to create any type of reporting application for backend business processes— completely horizontal. With that, it has Web-based delivery, good page layout and multipleformat options," said Paoletti. This is the background, he assured me, a history lesson so the next step is clearer.

"We found that a lot of our customers were using our tools for external-facing reporting applications, usually components of larger Web apps, like direct access to account statements, account lookups, etc." And so a new line of business was born: take the normally internally applied analytics tool (which was already accessible via Web interface) and make it visible to external customers! Brilliant!

But why do your customers go to the bother, and expense, to do so? I wondered.

"There are three concentric circles of business value that 10 times out of 10 are present with these types of solutions: One, operational cost reduction; two, customer satisfaction; and three, competitive differentiation. Three years ago, if you were to rank them, the first-move companies would have identified competitive differentiation as the primary driver."

But it takes all kinds... "Oh yes," Paoletti said. "The pragmatist companies of that time were interested in operational efficiencies—cost reduction, less paper, the classic stuff..." He related a story: "One of our customers, a wealth management firm, told us they were sending so many paper reports to their clients that when they went away on a business trip and came back, they couldn't get in their front door...three feet of paper!"

That's hard to tell whether it's a customer satisfaction solution, or a cost saving one, I offered. "That's why I call them concentric circles... it's both. You talk to somebody in the financial group, and they'll say they went with Web delivery for the cost reasons...we had to cut out all that paper. If you ask someone in the customer-service side, they'll say that the guy can't get into his house...we've got to fix this relationship!"

There were other cost-savings drivers, according to Paoletti. "The other major area was cost-center savings, measured in cost-per-interaction. If it costs $10 for an agent to look up an account balance, versus a customer getting it himself from the Web, that was a major factor...and it still is.

"But the area that we're really starting to explore now is that last circle—customer satisfaction slash customer retention slash customer growth... and those are three sides of the same coin!" he laughed. "In the financial services sector—and the telco sector is plugged into this, too—there is a white-hot effort to improve satisfaction within the installed base of customers as a means to achieve organic growth."

He continued: "For the past three or four years in the banking market, for example, growth has been accomplished through merger and acquisition. Some people will argue that there's still plenty of tarmac left for that, and to a degree that's true. But the M&A fervor has dissipated. Banks are now looking at how to grow their client bases organically. To do that, you have to retain that revenue stream, as a starting point, in order to have opportunities to cross-sell and up-sell.

"If you call into a contact center and get bounced around a phone tree, and some of the people you talk to don't have your account information, that's irritating, and can cause a loss of confidence in your organization. I would say that ‘customer sat' is the primary driver for the online customer service experience. And that's just as true on the business-to-business side as it is on the B2C side, because those clients can always take their business someplace else," declared Paoletti.

Speaking of which, I asked whether there were any trends among his customer base leading toward one side or the other—B2B or B2C? He didn't think so, but added quickly that "there's more sharing of ideas and concepts between the two sides of the house. They both want to use the best ideas that the ‘other side' has discovered... the ones that apply, anyway."

What kinds of ideas? "We take our clients through the process of identifying high-value and low-value interactions, and try to help them steer low-value to the Web—account balance, performance of certain stocks...basic ‘what's the status?' questions." But people DO expect more than just the easy stuff "Yes, being able to provide analysis tools, management tools, dashboard views across multiple accounts... these are all part of the new standard for self-service applications."

The Learning Self-Service Tool

It became clear, as these interviews wound down, that there was one consistent theme: the development of a Web channel through which to interact with customers is viewed—at least by this group, and their customers, reportedly—as a serious strategic endeavor with implications all the way up and down the range of business goals.

Davin Yap is a co-founder of the UK-based Transversal Corporation. He is also a Ph.D. from Cambridge University, and the prospect of talking with him frankly scared the hell out of me. It shouldn't have.

"The outward technologies among vendors are all very similar," Yap assured me. "It's a car. We're all selling cars. But the differences are more than just brand; they are the type of transmission, the type of engine... these are the things that are difficult to identify when you're first coming into this market." (Yap was being kind to me.) "The technology doesn't really matter; if they solve the same business need with the same business input...who cares? But there are differences between a station wagon, and a sports car and a van... we don't sell all three, and neither does anyone else."

For instance, I asked, how should "someone just coming into this market" differentiate among solutions? "There are chat bot-type solutions, which are script-driven," he explained. "If someone types in this type of question, it gives them this dialog back. These types of solutions are very high maintenance, compared to others." And, not surprisingly, not necessarily the best choice for every situation. "In some applications, such as financial services, the product is quite complex, and it changes rapidly. It's difficult to maintain a script-driven application. And, it's impossible not to find them tedious in a short period of time, especially if they're not perfect."

So, I wondered, are there other ways to provide content to an inquiry... such as search? "Companies who provide Web self-service do not identify themselves as being in search," Yap insisted. "Search is basically a document-retrieval problem. Search engines work by looking for patterns among hundreds of thousands of documents; the more documents the better, as a matter of fact."

He continued: "The issue of Web self-service is that you have a relatively small knowledgebase of answers to questions, maybe a few thousand answers to a few hundred frequently asked questions. So the search engines don't work; you're not giving them enough information. What you need to do is understand the query itself.