Reverse EngineeringBy Brian P. Watson | Posted 06-05-2008
IT's New Detective Game
Web 2.0 has unquestionably opened new doors for advertisers to reach consumers. But it's also unleashed a vicious enemy on business: customer feedback.
Eric K. Clemons, a professor of operations, information management and management at the University of Pennsylvania's Wharton School, and a fellow with the Cutter Consortium's Innovation and Business-IT Strategies practices, has studied the correlation between customer sentiment and business operations. He has seen firsthand how a hotel's operational inefficiencies led to poor online customer reviews, consequently hurting the establishment's financial performance. The world of customer collaboration and interaction holds extreme promise--andpotentially catastrophic dangers--for corporate revenues, Clemons says.
CIOs can play a pivotal role in this world. But will they? As Clemons sees it, IT leaders have a new opportunity to flex their data management and business analysis skills, effectively becoming the central nervous system of their company. Or they can sit back and gather Web data at the request of others--like marketers or quality-assurance pros--who undertake the project of figuring out the correlations between online reviews and operational efficiency.
The choice for IT leaders is clear: Either become a data hunter and gatherer, or remain a simple, lowly shepherd. Clemons spoke recently with CIO Insight online editor Brian P. Watson, who edited and condensed their conversation.
CIO Insight: What are businesses doing to get the most out of Web 2.0?
Eric K. Clemons: There are probably three different opportunities for monetizing the Internet. One is through internal efficiency, internal cost reduction or external cost reduction. Another way involves the platform--whether you're talking about traditional Internet stuff (Web 2.0, Web 3.0, social nets, unlimited bandwidth, virtual reality) or, at the other extreme, mobile text messaging.
The third opportunity is about revenue enhancement. Are you selling insight and intuition? Are you selling products? Are you selling product referrals? Are you selling usage fees on services?
We've been doing bottom-line tweaks through cost reduction for 30 years. There isn't a lot left to clean up in the supply chain, whether internal or company-to-company operations. We've already done that well.
The limitation we have now is in trying to understand perishable inventory with limited information. That's not going to be altered by technology, because we have 10 suppliers while we may have 10 million customers. The place where technology enhancement allows a breakthrough in communications today is with the 10 million customers, not with the 10 suppliers.
People are always focusing on revenue enhancements. How are they doing that with Web 2.0?
Clemons: One way is by selling insight. I look at Second Life, where there are no constraints, and say, "That's how people want to look." I look at YouTube, where everything's free, and say, "That's what people want to post, what people want to watch." So it's using the Net to develop an anthropological insight into social behavior.
A second possibility is selling stuff. Amazon has figured out how to sell lots of books online, but it hasn't really offered me, initially, any breakthrough stuff. I'm looking at products because I'm interested in things you almost couldn't sell if you didn't sell them online. We can sell virtual stuff: There are now shops in Second Life that enable you to buy clothing for your avatar. I don't know how big a market that is, or should be, but at least we know it exists.
There's also money to be made by charging a subscription for information, but that's been remarkably difficult. Very few companies besides Blizzard's World of Warcraft have found a way of getting people to pay for online subscriptions. It's a problem that can be addressed. We need to figure out if it's about educating people, or if there's a fundamental reason why we won't participate in a service that charges us.
Social networking and social network-based advertising can be done badly, like the way Microsoft tried it with Facebook, which was catastrophic. There are some examples of wonderfully crafted marketing efforts that could not exist without social networks-- the natural extension of social network marketing to mobile network-based marketing.
Then there's the Google-based model, which is basically snooping: reading my e-mail and then putting an advertisement in it. Each of these differ enormously in terms of their plausibility, customer acceptance, social acceptability and revenue possibilities.
You say they differ. Does one of these scenarios stand out?
Clemons: We can manufacture anything we want. We can have anything we can describe. A customer doesn't even know he wants it: Somebody will make it; the customer will find it and fall in love with it.
There are probably 5,000 kinds of beers available today, yet at one point, it looked like we were down to three. When I was a kid, it looked like every soft drink would be either Pepsi or Coke. Now Pepsi and Coke are losing sales to products like iced tea. That's not a change in our ability to manufacture--that's a change in our ability to inform consumers.
Some of that comes back to social networking sites, but some of it comes back to learning what consumers want and don't want. We haven't yet figured out how to monitor traffic online to figure out what's hot and what's not. One trend that I think will happen--not because of CIOs, but because of chief marketing officers--is gaining insight from traffic. Companies will have to master social networking marketing.
A friend of mine opened a beer distributorship. If you go to the right-hand side of his Web site and click "What's New," you'll get a list of the stuff that's come in the last two weeks or so. If you click on it, you'll get his description of the beer. Click on the beer name, and you get access to RateBeer.com and hundreds of thousands of reviews. If you get a reviewer and he likes the beer, that doesn't mean you'll like it, and if he hates it, that doesn't mean you'll hate it. A good reviewer will tell you what he likes and doesn't like.
The principal impact of this is that it changes the nature of what my friend sells. He still sells Coors, but he's also one of the highest-volume retailers on the East Coast of the stuff no one else sells. We know he has it, so we go there. He knows we'll find him, so he stocks it. It becomes a self-reinforcing phenomenon.
So a lot of this comes back to search.
Clemons: Social search will become explosively important. The most difficult thing for companies to understand--and this is a huge thing for CIOs--is that you can control what people post about you. You can't control what people post by controlling the posting or the site, but you can control it by not screwing up.
One of my clients had the worst reviews of any hotel in his chain, and I was able to demonstrate that the single, most catastrophic thing for a hotel trying to sell online is the presence of negative reviews. Once we realize that reviews are going to determine what sells and what doesn't sell, there's a whole new role for the CIO: understanding what determines the perception a guest or customer has of the experience and finding out what went wrong to cause those reviews. It's a whole different job.
How does the CIO segue into this new role?
Clemons: I'm giving the CIO a new assignment: Find the causes of a notable negative experience. If something happened five times during five years at one hotel, that's not much of a pattern. So how do you find it? Go to a social networking site, find the reviews that determine your image and reverse-engineer what caused the bad experiences. That's going to be the most high-revenue thing a CIO can do.
If the only thing that determines my ability to sell is the number of negative reviews, I'm no longer managing to percentages, because the negative people are the ones who can have the most impact on my marketability.
So find out what the negative reviews have in common. Find out the approximate cost, the immediate cost. Find out what it was about their experiences that led the customers to write the negative reviews. Find out what it was about your operating conditions at the time that caused the bad experiences and reverse-engineer those conditions. This is a whole new role for the CIO.
How does reverse engineering work?
Clemons: If the customer reviews are not the kind you want, figure out why. Companies have done this for a long time. If you have a batch of defective chips, you figure out why they are defective. That's why there are batch numbers on manufactured products: They are meaningless to consumers, but if there's a defective batch, the numbers help the company figure out what went wrong.
We're talking about reverse engineering not because the company found an error, but because consumers were complaining.
In the hotel's case, it was easy to see what happened: When they overbooked, some people were given inferior rooms. When the hotel was over capacity, the quality of service in the restaurant collapsed. When some people forgot to book a room for a conference, instead of getting the conference rate, they got a nasty rate for the last available rooms. People complained about the price, the service and the quality of the room.
The reasons for negative reviews are not always that obvious. But even when they are, as in this case, the hotel didn't see that customers weren't booking because of the negatives. They didn't know what the problem was, so they couldn't correlate the problem with the dates and what was interesting about those dates.
There are two parts to this: monitoring not only satisfaction surveys, but also what gets posted online, and doing reverse engineering.
The CIO's New Job
What makes this the province of the CIO and not of the chief marketing officer or someone else?
Clemons: This is enormously data-intensive. CIOs have people who know how to do this--how to get all your data off the Internet and into your own internal operations. They know how to look for the relationship between sales and reviews, know what in the reviews drives the sales and knows what in the internal operations drives the reviews.
Finding what matters requires fairly subtle statistics and insight, so the CIO may want to get someone from marketing or QA involved. But it's very data-intensive. It's not like marketing, where you go out and get your data. In this, you don't get your data--you find your data. This isn't about conducting surveys; this is about finding what people actually see and look at and use. Marketing people may be better right now at analyzing the data, but the CIO is probably much better now at getting it.
It's an interesting opportunity. But what if the CIO has too many other responsibilities? What's the downside of not proactively attacking the problem?
Clemons: I always think of the CIO as the chief data officer: the one who knows how to find any information you want. Historically, CIOs have thought of themselves as the people who make sure data is available, make sure you have the right piping, make sure the internal interfaces work and make sure the interfaces to the outside world work. So much of the data you want is external, not internal.
There are two possible roles for the CIO in this arena: He can declare himself the central nervous system for the organization--the one who finds out what people are and aren't looking at. Or he can declare himself the data vacuum--the person who pulls data off the Net for the marketing team at the marketing team's direction.
CIOs can decide whether they want to be at the center of a great detective game or simply be a data vacuum.
You call this a new assignment for CIOs, but are they in step with what you're saying?
Clemons: These events are rare, really complicated to diagnose and, until recently, had no commercial impact. The hotel chain I looked at was blaming corporate for its inability to sell, but it turned out to be the negative reviews. Nobody in the corporation had asked about the relationship between online reviews and online marketability.
CIOs haven't been asked to do this yet. But they will be.
With the explosion of the Internet, why aren't businesses already doing reverse engineering?
Clemons: With operations, we have historically audited for the mean level of performance. We're always trained to measure the middle. But the middle doesn't blog. It's not the mean that affects social networking.
Now we have technology coming into use--social search, social network-based marketing--that has never existed before.
Therefore, we have a new category of event (low probability, high visibility) that determines the ability to market. The data is difficult to gather, and until three years ago, it had no predictive power. Why would anyone be monitoring this?
Now there's a new problem. These are the very things you have to monitor, because these are the very things that affect your ability to sell.
You need to figure out what events affect your customers' perceptions. You need to figure out what caused those events to have their negative impact. It's an interesting detective game that no one cared about before.
Chief marketing officers like Web 2.0. Younger workers do, too. CIOs also are interested. Do CEOs and CFOs get it?
Clemons: A lot of time, when people decide to use Facebook, they don't know what they're going to use it for. The CMO says he'll advertise on Facebook, but we haven't really figured out how to harness community content in a way that doesn't destroy the credibility of the site.