Sentiment Analysis Run Amok

No offense to Glenda from Ocala, FL, but her five-star rating for a plush towel in the latest home catalog I received is meaningless to me. Now don't get me wrong, I think user feedback is extremely helpful. In fact, I wrote about its usefulness in this post Customer Reviews Are a Great Leading Indicator in Warranty Detection.

And reviews are one of the major factors I consider when making online purchases. Whenever I buy new electronics, I make sure to look at the user product reviews on Newegg.com, CNET, and ZDNet. I find that many educated consumers post very helpful feedback. For example, recently I was looking at new monitors when I became fixated on the HP 2509m. It's one of the few monitors in its size and price range with a glossy screen, so I was trying to decide if reflection was going to be a problem. The user reviews were very helpful in helping me figure out it would not, and I went ahead with the purchase. (The monitor, by the way, is absolutely gorgeous.)

But using customer reviews in catalogs and other promotional materials is not a good use of sentiment analysis. Sentiment analysis, sometimes called reputation management, is the tracking of opportunities and threats to a company's brand by examining traditional media and new media and data sources. Customer reviews are just one data source in a more robust sentiment analysis initiative.

Mining your customer information should be used to help make segmentation decisions in terms of which features to offer, look for quality control issues, and much more. Some examples of vendors that can help you mine your customer reviews in no particular order include SAS, Endeca, Nielsen BuzzMetrics, Autonomy, Temis, Jodange, YellowBrix, Factiva, Radian6, and Biz360. Which vendors you gravitate toward depends somewhat on which data sources you want to analyze, whether they are internal or external, tradtional news sources vs. Web 2.0 social media, etc.

This kind of reminds me of the movie review scandal when one major motion picture house was caught making up "Terrific!," "Absolutely Amazing!," and "Best of the Year!" movie reviews. I'm not accusing companies of that, but is it really so hard to stick a fake "Five Stars!" on a catalog page and label it as "anonymous user from Chicago, IL?" That's why I find them useless.

So go ahead and mine that data! But plastering your catalog with random quotes from people that could just be doing it so they can show their sister the newest catalog with their quote in it? Not so much. I don't think that's helping your brand any, which should be the goal of sentiment analysis or reputation management programs.

Gesture Recognition Saves Field Service With the Wave of a Hand

If you've been reading the entries on my blog about augmented reality, you know I've soured a bit on some of the consumer technologies applicability for use in the enterprise. But I came across one that I think makes a lot of sense and it's gesture recognition.

The trend in user interfaces for mobile field service applications is to make the buttons and widgets bigger and easier to read and press. But what if a field service professional didn't have to touch the device at all? Voice-enabled functionality is obviously another solution, but for this post I'm concentrating just on gesture recognition.

Field service technicians in rugged environments face several challenges that make interacting with technology less than ideal. Whether it's because they're wearing gloves, or holding tools, or are working in natural (e.g., sandy, dirty, oily, etc.) environments, entering information into a handheld or laptop is less than appealing. These challenges often inhibit the adoption of mobile devices and field service applications. This is a real pain point.

Gesture recognition is one possible answer. With a few swipes of the hand, a technician could close out an order, pack up, and move on to the next job. All without worrying about damaging the precious little piece of technology their company handed them. I envision that for this to have near-term applicability, it would need to be installed/enabled on a ruggedized or regular laptop. The reasoning is that if you're in a rugged field service environment, you're probably not carrying a lot of the smaller, consumer-grade smart phones. (If you are, however, then yes of course it would still apply.) Many laptops now come with webcams preinstalled (I can't believe I just used that word. Listen to the George Carlin routine on "pre" words and you'll understand why.)

Check out this video from Texas Instruments on SlashGear for their early solution.

Here's another interesting article on Engadget about magnet-based gesture recognition from MagiTact.

I love the potential of this technology and I will continue to dig into this and query the enterprise software and hardware vendors what they're doing on this front. Stay tuned.

By the way, I would like to thank Alan Cheslow at Microsoft for bringing this to my attention (albeit, indirectly through his Twitter feed). If you don't already follow Alan (@acheslow), I highly recommend it. He's plugged into many new technologies and offers insightful analysis on their application. I find his posts to be highly relevant.

What's Up Tech Doc?

Technical documentation takes many forms. The 30,000-page manual Boeing distributes with new planes is at one end of the spectrum. The color fold-out you get with a Lego set is at the other. But regardless of the form, technical documentation needs to be accurate, helpful, and easy to use otherwise it's worthless. If the document fails any of those three litmus tests, customers will turn to alternative sources for information. When that happens, a company loses control. Why is that a problem? Let's take an example from my own experience.

I was playing a game on my home computer one day when it suddenly crashed. I rebooted the system only to be greeted by a black screen with a sentence about how it could not load the OS. Nothing I tried worked. So I logged into another computer and searched Microsoft's website for a fix. For whatever reason, whether it was user error (i.e., my fault) or that of Microsoft, I could not find a ready fix for the error message I was receiving. So I did what most people do and Googled it.

Sure enough I got many hits from technical blogs with answers to my problems. So I picked the most professional looking blog with lots of replies in the affirmative that the fix actually worked and proceeded through the steps.

Prolem solved, right?

Wrong.

Turns out the blog was wrong, and I made matters worse. I had a dual-core processor and the fix only worked for single-cores. I, and Microsoft, lost control of the situation. I'm willing to concede that the proper answer may have actually existed on the Microsoft website, but because I got frustrated quickly and couldn't find it, I went to an alternative source of information and things got worse. That's why companies need to pay more attention to their technical documentation. Combining it with web self-service and frequently asked questions (that are kept up-to-date of course) can resolve a majority of incoming problems. But if you don't pay attention to these resources, customers will go elsewhere. I did and regretted it.

Who Owns the Digital Home?

The goal of this post isn't about finding out who has the most wired-up crib. (For a more detailed tutorial on how to implement home automation, check out this article on CNet.) It's about getting people to think about which manufacturers, retailers, or service providers are best positioned to service the rise of Internet- connected devices in the home. Today at a Gartner Research Offsite on the topic of Connected Mobile Devices, we talked about the future of cellular networks. Which devices would use them? Would smart phones replace all other devices, etc.? As the discussion turned to the home and which devices would be connected and what mode of data transmission they would use, it made me think about the opportunity vendors had to service devices in the home.

Right now I have three laptops, a desktop computer, an HP printer, a Motorola Droid, and an XBox 360 all using my wireless network to connect with the outside world. And if I have a problem with any of them, I need to contact each individual manufacturer for service. Is there an opportunity for one company to step in and own all these devices? I would think Comcast would be in a good position because it's the pipe being used by all these devices. AT&T has been very public about supporting Internet connectivity in the home, and one of my colleagues during the research meeting said AT&T is already doing this (apologies that I can't remember who it was). I think other companies that are reasonably positioned include Best Buy, Sears, and Verizon.

But the question remains: what's the incentive? It's a pretty complex environment, and a service provider would have to familiarize itself with a myriad of devices. There are two scenarios in my mind:

• Opportunity to cross-sell or up-sell devices
• Opportunity to sell additional bandwith

Regarding the first bullet, a diversified manufacturer like a Samsung could service and sell televisions, laptops, phones, and potentially washers, dryers, refrigerators, and other home appliances. They could create a Samsung network in the home.

Regarding the second bullet, a Comcast, Verizon, or AT&T could service a consumer's Internet-connected devices and either sell additional devices or more data plans. The issue of whether each device will have its own data plan or share a data plan is still being worked out, but suffice it to say, carriers would have a vested interest in the resolution of this point.

I think we're still quite a long way from a truly interconnected home where all the devices talk to each other. But we're definitely seeing the rise of devices that reach out to the Internet (or cloud or whatever you want to call it these days) either to receive/deliver content or for remote maintenance. I personally think there won't be one vendor who covers the entire home. I think it will be a combination of a few vendors all specializing on groups of products or data transmission methods.

What are your thoughts?

Augmented Reality or Augmented Hype?

I vote hype. But let me explain. In an earlier blog post "10 Technologies That Won't Have Any Impact in 2010 - and Three That Will," I wrote that while I thought augmented reality applications had a lot of potential for enterprise service organizations and the software and service providers that support them, I ranked the potential for 2010 as "Medium" because I just don't think the market is ready. Since then I've had a few people approach me and politely question me why I didn't rank it higher. So I thought the topic was worthy of further exploration.

One of the people who approached me was colleague, Simon Jacobson, research director for manufacturing, and I asked him to share his thoughts on the potential in manufacturing and service. He wrote:

"The argument I'd make for augmented reality in manufacturing operations is for training and process planning and design functions instead of in the execution and/or performance management functions. Using augmented reality as part of the upfront phase of ideation and design, when production processes are laid out, it can contribute to better understanding of process capabilities of all assets (both worker and physical), which can help identify when and where the points of failure will occur and decrease costs, increase overall quality, increase predictability of manufacturing's ability to respond to demand, etc. As an example, imagine the possibility of being able to see digitized factory layouts and/or complex models of machinery and production assets (especially in distributed environments) where changeover comes at a high cost. There is also the angle to enable a more productive and knowledgeable plant worker. Invensys for example, has created an operator training simulator that uses virtual to create a video game like function to on-board new employees."

Additionally, I found this great blog post in support of augmented reality in service on Clickipedia, the service blog from software vendor, ClickSoftware. (Disclosure: while I will not divulge which companies are clients and which aren't, I will comment that I interact with ClickSoftware as part of my ongoing research on the service supply chain.) Clickipedia highlights some potential uses including:

• An electricity technician can look down a street and see houses affected by a power outage
• A technician can use a heads-up display to overlay a maintenance schedule over an asset

And several more. But it's still all about potential and not, to risk overusing a phrase, reality.

For example, the Gartner Hype Cycle for CRM Customer Service and Field Service, 2009 lists "Rich Information Visualization" at the very initial stages of adoption, which is under the category "Technology Trigger." And I fully admit I'm probably overreacting to ridiculous examples like virtual graffiti, shooting lasers at people, or Shaquille O'Neal dancing on a bottle cap.

So let's keep the discussion going and see what 2010 (but probably 2011) brings.

For more, AMR Research subscribers can read, "The Future of Work Meets Manufacturing 2.0 at Invensys."

Customer Reviews Are a Great Leading Indicator in Warranty Detection

Too many companies start the warranty detection process by examing historical warranty claims. If you're doing that, it's already too late. You've got a problem. Somewhat better are the companies that use analytics to read claims in near real-time to look for patterns as they come in. But there's a step that companies can take that precedes even this, and it's the aggregration, review, and analysis of customer reviews.

At best, about one in ten customers will leave a product review on a website. But it's commonly known that there is a much greater chance a negative experience will be posted versus if the person has a positive experience. There is often valuable clues given in these reviews. Let's take a specific example about an actual review I left about my new refrigerator:

Cheap piece of junk for $1,200

"I've owned this refrigerator for less than a year and already the bottom shelf has cracked under the weight of a few gallons of milk, and the main handle on the outside is loose. I'm sorry, but when I spend $1,200 I expect it to last more than 9 months before things start breaking."

If the OEM was reading this review, they could cross reference it with other reviews to see if other consumers were having the same problems. They could then look at the model number and trace those parts back to a supplier (ideally). The company could then take proactive measures in addressing potential quality issues to mitigate future claims. I think my refrigerator is actually suffering from a quality/warranty issue for two reasons:

1) The replacement shelf the retailer sent is already bending preventing the drawers from closing properly.

2) The service technician told me that he's seen this problem before with this model!

Clearly there is a pattern here that is developing that would not necessarily be readily apparent if the OEM was only looking at submitted warranty claims. So get out ahead of your warranty claims and invest in BI/analytics technology to conduct leading indicator analysis.

10 Technologies That Won't Have Any Impact in 2010—and Three That Will

One of the directives I gave myself for this blog would be to examine popular news stories in technology and examine the applicability to the service supply chain. In a recent post at TechCrunch, Co-Editor Erick Schonfeld wrote about ten technologies that would rock 2010 (his words, not mine). I thought it would be worthwhile to speculate if any of those year-rocking reverberations would be felt in the enterprise world. Here are my thoughts:

1. The Tablet: There are a few anticipated tablet products slated for next year, but the one Schonfeld is simply ga-ga over is the one from Apple. But if we use the penetration of Apple's current products (iMac, iPhone, etc.) in the world of the service supply chain as an indicator, I would have to predict that the Apple Tablet will have little impact despite predictions of selling 10 million units in the first year. I could see some applicability in hospitals where middle-ground devices (i.e., devices between a phone and a laptop) have had some success and perhaps white-collar service, but not in industrial service (e.g., A&D, automotive, heavy equipment, etc.) where devices need to be more rugged and there is already competition from robust hand-helds and now netbooks. Impact: Low

2. Geo: Geolocation is already a big part of service. So I think this one will apply, just not in the way Schonfeld describes. Instead of Facebook, Twitter, Foursquare, etc., GPS chips will feed service applications: field service, fleet management, and asset management. Impact: High

3. Realtime Search: This is about using such search engines as Google and Bing to search real-time data feeds like Twitter and Facebook. Impact: Zero

4. Chrome OS: Microsoft Windows has a near stranglehold on the service enterprise on the desktop and to a lesser extent, mobile devices. Service application vendors are still asking us about the impact of BlackBerry and Symbian in the mobile world, and we never field questions about Linux on the desktop. Impact: Low

5. HTML5: There's a new version of HTML? I hadn't heard. And from the looks of a lot of service companies websites, neither have they. Impact: Low

6. Mobile Video: Video transmitted through mobile devices has the potential to reshape field service. Techs or customers in the field can use video to show remote technicians exactly what they're seeing. This could reduce travel time for technicians (a major cost of service) and increase the number of calls per day. But so far, few companies actually use it. One major limitation has been the lack of support for video on the hardware. And after many conversations with the device manufacturers, I don't see this changing much in 2010. Another challenge has been bandwidth. Despite the growth of 3G mobile networks and the increase in broadband to consumers, transmission clarity and speed has been too erratic to rely on for consistent service support. Impact: Medium

7. Augmented Reality: Probably not what you're thinking, augmented reality is more about layering on data and information to the view of your mobile device's camera and less on holograms. This technology has some potential. Mobile service applications can layer on customer information or the locations of other technicians or better yet, what products and parts would look like properly installed. Even though I think the applicability of this technology is high, I don't think the service market is quite ready. So I'm hedging my bet. Impact: Medium

8. Mobile Transactions: This idea is about turning mobile phones into point of sale devices. Since there are already devices in the enterprise service world that do this, I'm ranking it low. Impact: Low

9. Android: Android will make a big splash in the consumer world. I'm thinking about getting a Droid phone myself, but it's not making a big impact in the world of mobile enterprise service. The Microsoft OS still dominates in North America, and when service application vendors think about new mobile operating systems, they typically add BlackBerry or Symbian. Chances are slightly higher in Europe and Asia-Pacific, but not enough to bump up the ranking. Impact: Low

10. Social CRM: Twitter and Facebook may have taken the consumer world by storm, but they still aren't taken very seriously in the world of enterprise service. There are several consumer-facing brands that are embracing social applications and doing it well, but the general world of industrial service does not and neither do most of the enterprise service software vendors. For example, despite the existence of many real-time communication applications, many field service software vendors still find it necessary to create their own instant messenger. There are obviously exceptions (e.g., Salesforce.com and SAP CRM's integration to Twitter), but it's not prevalent. Impact: Low

And not to seem overly lazy, here are a few technology predictions of my own:

Cross-channel: Cross-channel service applications will gain adoption as companies try to create a seamless customer experience across all points of contact in the enterprise. Whether a customer is accessing a company's website, calling a service center, or stopping by in person, shouldn't matter. Every point of contact should be able to access the same customer information anywhere.

BI/analytics: Using business intelligence and analytical software to help improve service levels, increase profitability, and grow revenue always ranks high in importance in many of the surveys we conduct. I don't see this changing especially as more and more service vendors are either developing these kinds of technologies or are partnering for it.

Remote diagnostics: Machines that can transmit errors automatically to remote service technicians is already a billion dollar market if you include all the machine-to-machine technology available. Cars, photocopiers, computers, hospital equipment, and many more devices can already do this. But there is still plenty of room for growth. Take a look around you right now and identify all the electronic devices that could take advantage of this technology. For me, I see a coffee maker, stove, refrigerator, microwave, and television. And the number of devices that could employ self-fixing technology is even greater.

Agree? Disagree? Please feel free to leave your comments below.

Oh, iTouch, Where Art Thou?

Twas the night before Christmas, and I had yet to pick up my wife's gift. So, in a near panic I headed out to the local big-box electronics retailer looking for an iTouch. I entered the cavernous store and started my hunt. I spied a hanging sign that read "iPods/Mp3 Players," so I headed over to it. The sign was hanging directly above a counter where customers can test out iPods and headphones, but there weren't any actual Mp3 players for sale. Since the counter was smack dab in the middle of three aisles of Mp3 accessories, I figured they had to be close.

I spent the next few minutes walking up and down each aisle, but to no avail. I couldn't find the iPods anywhere. Every section of every aisle was packed with accessories. Lots and lots of accessories. Cases, clips, and armbands were all hung with care in hopes that the consumers would soon be there. Chargers and trasmitters. Headphones and speakers. Radios, docking stations, cables, connectors, batteries, adapters, and shoe sensors all beckoned.

But no iPods.

I spotted a woman carrying an iTouch, to whom I blurted out, "Where did you FIND that?" She was kind enough to point me to a peg (one peg) with two iTouches hanging from it. Somehow I had overlooked this lonely peg in a sea of add-ons. In my defense, the sign above the players read, "iPod/Mp3 Accessories." This confused me: isn't the iTouch the original product on which the accessories are based?

I understand the iPod Economy generates a billion dollars a year in revenue. In 2005, there were 400 add-ons. In 2006, the number of add-ons ballooned to more than 2,000. Today? Amazon lists more than 17,600 Mp3 Accessories for sale (granted there are some for non-iPods in there). But despite these robust numbers, I couldn't help thinking finding the original product was unnecessarily frustrating.

While many companies rely on the service of the original product or the sale of add-ons for a majority of their revenue, they shouldn't forget that not everyone actually owns the product yet. You can't access the service revenue if your customer can't buy the original product in the first place. In enterprise service, this isn't as much of a problem. People generally know where to go to buy a car or a Caterpillar tractor. Or do they? Don't automatically assume people either already own your product or know how to find your product. Make it easy for them to buy the product, even if you're a service company, and the service revenue will follow. Customers will remember that you helped them or, in my case, didn't.

Do You Hear What I Hear? The Biltmore Does

A few years ago I was attending a conference at the Arizona Biltmore, and the hotel manager called me to ask if they could use my room as a conference room. This would, of course, necessitate moving me to a different room. I was unpacking my things and watching the Minnesota Vikings game (I don't remember why; I'm a Patriots fan), but I really didn't mind moving. Even though I wasn't terribly put out, the Biltmore felt the need to make amends, and I was about to learn a lesson in elite customer service.

The first thing I noticed about my new room was that it wasn't a room at all. It was a $1,500-a-night villa. The second thing I noticed as I walked past the kitchen into the living room area bathed in natural sunlight from the skylights above, was that the television was on to the very game I was watching. And there was a bucket of ice with Coronas next to it.

So what's the lesson here? Is it that the hotel manager has terrific hearing and heard the game through the phone? Actually, yes, it is. Before I moved rooms, she said she could hear I was watching the game and asked if they should come later. The lesson for enterprises is that the Biltmore manager put my customer service problem into context.

I had two problems: 1) I had to give up my room and 2) I was watching the game and would miss some of the action. The Biltmore addressed my customer service issues with remedies directly linked to the original problem: 1) a bigger, better room and 2) the game was waiting for me. Too many service organizations have one-size-fits-all service remedies (e.g., handing out gift cards) that don't address the specific problem. If you give a customer a gift card, and their product still isn't fixed, you haven't solved the problem.

But you can't do this unless you have accurate customer information. Hotels like the Biltmore have reputations for knowing their customers well, including their likes and dislikes. One recent customer said it's as if their information "followed them around." In enterprise service, any member of the organization should be able to go to a central location and access the same customer information. It should "follow them around," from sales, to call centers, to field service, to order management, etc. Then, and only then, will you be able to offer personalized and appropriate remedies to any problem that might occur. At that point, you'll truly hear what the customer hears.

JiffyLube Knows How to Shape Demand. Do You?

The other day I drove past my local JiffyLube and two of the employees were standing outside with signs that read “Drive Right Up, No Wait!” What a terrific way to shape demand for their service business. Instead of waiting for it to come to them, they were actively trying to create the demand. Who was their target? The driver who knew they needed an oil change but either a) didn’t like waiting in lines or b) was putting it off for another time. As a father of two young boys who was driving a car that was 2,000 miles overdue, I was extremely tempted. (I fit both A and B scenarios.) But I did have to pass because I was already late for another engagement.

How do you create demand for your service business? Do you wait for it to come to you, or do you go out there with signs that say “No waiting?”

For a more detailed report, AMR Research subscribers can read, “What is Demand Shaping?”