Customer Service

Humanizing the Online Customer Experience

The Internet has seen massive adoption. Online retailing has continued to dramatically build its customer base and more and more companies are sending their offline customers to the Internet for service and support. To service both these old and new customers, the Internet has already had an abundant history of “self-service” tools that have each seen their boom and bust cycle.

Active forms, eCRM and analytics have all tried to make the Internet an equal, but more cost-effective, participant in the world of customer communications along with Mr. Bell’s great invention, the telephone.

A problem remains. In spite of the Internet’s massive adoption, in spite of its remarkable power to make (and unmake) entire new industries, online customer service remains the ugly stepchild of corporate America. We say that we want to satisfy customers online. We say that we want them to feel just as well-served online as in the high touch bricks and mortar store or when they are on a telephone speaking with a human sales representative.

Why does customer service lag so far behind on the Internet, even as so many are adopting the Web as their favored method of finding goods, services and information?

The Inhuman Internet

If you walked into a store and asked where to find black penny loafers, would a salesperson hand you a catalog and tell you to find it yourself? And if that did occur, would you stay, or go to another store?

You go to the store to buy something and get advice and help from a human being. If you called a company with the intention of signing up for a customer loyalty program, would you expect to have to sign up through a voice recognition unit? The reason that the Internet doesn’t seem human is simple: For the most part, it isn’t. Humans find out what you actually want — and then try to give it to you.

Instead of being a channel for human-quality service, the Internet has become the curtain behind which companies hide, telling themselves that by giving the customer the tools for eventual “self-help,” they have, in fact, communicated with their customers and given them what they need — or at the very least provided what they think will be OK for their customers at a lower cost than complete service.

That curtain also hides our customers from us. We can’t tell what they really want as they click through our Web pages, as they search for information, or as they get halfway through filling out an online form. While we can gather statistics and analyze, draw conclusions, and decide what we think our customers want, we have not really engaged them. They do not fully understand what we are offering, and we cannot tell if they are satisfied or simply frustrated and confused. More importantly, we haven’t taken full advantage of the opportunity to interact with them when they are there, at our site, just waiting for us to really engage.

Give the People What They Want

There is a famous retelling of the death of Samuel Goldwyn, the great Hollywood mogul and tyrant. Goldwyn’s funeral was mobbed. Everyone in Hollywood seemed to be there. Milton Berle and Jack Benny stood on the steps of the building where the services were being held.

Surveying the teeming thousands before them, Berle said, “Who’d have thought so many people would come to his funeral?”

“It’s the same old story,” replied Benny. “Give the people what they want . . . “

What do the people want? Are they longing for the chance to attend your Web site’s funeral? The gold standard is an interaction with another human being. We know what we expect from a sales assistant who is good at the job. If we use that as the standard, let’s look at some technologies that are commonly deployed online and what their limitations are.

Three Clicks and You’re Out Isn’t Good Enough

Web pages are the foundation of a good Web site — but are they customer service? What would you think of a sales assistant who handed you a catalog when you asked about a product? What if there were no sales assistants — only signs that gave you progressively more detail as you got deeper into the store, like a menu bar?

What if, every time you guessed your direction incorrectly, you had to start all over at the front of the store? You’d probably end up with about a 3 percent conversion rate on visitors to your store.

That’s what the average online retail site gets.

The Allure of the Algorithm

Search is my favorite online tool. I use it daily. But search doesn’t tell you things, it only finds things — lots of things. And brings them all back to you like a mad, overachieving Labrador Retriever, leaving you to sift through the pile, looking for what you really want. That’s when search is operating as designed. However, if you’re not sure or clear enough, if you’re not using the right words or phrases when searching, then you’ll be left with a potentially never-ending task looking for that desired needle in the digital haystack.

We want to believe that there is a magical algorithm out there that will provide all customers what they want and, in so doing, save us from actually having to find out what they want; that this magical algorithm will somehow match up human minds with bits of information from our Web sites, leaving both sides of the sales equation completely satisfied and untouched by our own human hands.

When customers are seeking every red sweater in your store, it’s a great model for interaction. When customers want customer service and all they get is 5,263 responses to the question “How do I return the sweater I bought?” along with an at-the-ready search engine looking at them like a slobbering Lab, dying to be asked to go fetch is simply not a great model. Your customers aren’t being served, and they know it.

80/20 Automation = 20 Percent Failure, Over and Over

The coming tool for customer service is conversational, in the form of the automated agent. The language engines that support these deployments have improved dramatically from the old “bot” types of deployments. They understand what users are asking for and answer their questions with much more power and agility than a search engine. Anything that can be scripted in Java or HTML can accompany an answer: graphics, forms, Web navigation, links and more, extending the conversation beyond words to pictures, symbols and pages. Advanced systems can selectively “take control” and lead the conversation, whether briefly to clarify ambiguous responses, or more extensively to “accompany” a visitor through a business process, such as signing up for a promotion or obtaining an RMA.

It sounds like the most human-like capability yet. Where’s the catch?

The catch is, nothing can be automated completely. When you reach the edge of this tool’s automated universe, it doesn’t know the answer. Poorly deployed examples fall off abruptly, either coming up with not-on-the-mark default answers or, just as bad, trying to match every question with the answers that it has, whether the question fits or not. The results are not human, or even humane. And how many times a day can you afford to alienate 20 percent of your customers?

Be sure to demand a sophisticated strategy for the other 20 percent. Advanced systems can even be configured to move the conversation to a customer support representative and back to automation again without impacting the customer’s experience.

Human Beings Are Only Human

The most human form of online customer service is, of course, provided by human beings. Presently, this takes the form of online chat, with some sites moving in the direction of VoIP. There is still a dilemma, however, inherent in these products. Depending upon the type of deployment, they can easily be just as expensive as the telephone.

In order to reduce this expense, companies typically have customer service reps monitoring several chats simultaneously. In many deployments, this has degraded the customer experience to a point where the real value of including human beings in the service process is completely lost.

Recently, a friend called me up and told me that there was a new tool out there that provided automation of Web chat. He had experienced this tool at a certain hardware manufacturer’s Web site. Since my company automates online chat, I was intrigued and went on the site. I found that there actually was no automation at all. The human beings chatting on the site were simply so badly trained and overtaxed that many of their answers had only a “keyword” sort of connection to what was really being asked of them. The company was paying for real human support but was giving the impression of cheap, poorly deployed automation.

Given enough challenges, even human beings can’t humanize the online customer experience.

What People Really Want

So what do the people want? They want it all! Just as no one hammer can drive every nail, no one strategy or technology will serve customers well in every situation.

After all, the most important aspect of being served by another human being is that they find out quickly what we want and then build the experience around our need. That’s the human response, and that’s the kind of customer service that the people really want.


Robert Williams is CEO ofConversive, Inc.


1 Comment

  • Excellent article. I agree wholeheartedly. Online customer service is even more important than brick and mortar customer service because the customer cannot see you. So online chat or emails can be misconstrued when it comes to tone. Humans still want to be humans and talk about what they are looking for or some just want to talk. Give them what they want and your business will succeed.

Leave a Comment

Please sign in to post or reply to a comment. New users create a free account.

CRM Buyer Channels