Saturday, February 6, 2010

magicJack CEO Dan Borislow Interview Part II

Each year at CES, queues for taxis, coffee, and lunch always seem to take a month, particularly during the first two days of the show.

Pretend it’s still CES…it’s been a long month. You don’t read me for snippets, you read me for analysis. Maybe even insight, at least occasionally.

When we left off four weeks ago, I’d just broken the news that magicJack was stirring up the telecom market once again, with the announcement of their forthcoming femtocell, which I (and now most others) call the “femtojack”. After Dan’s team published their release on the femtojack, I was inundated with a ton of technical and regulatory questions. I’m not going to attempt to answer them here—not only is doing so magicJack’s job, but I’d rather focus on a topic I consider more compelling—customer service, particularly in the context of making things right.

What’s the secret to success for just about any company? Ideas. Innovation. Passion. Execution. The magicJackers are short on none of these. Borislow and his team exude passion, whether discussing hardware, software, networking, saving consumers money, or their competition. I have to laugh at some outlets taking Dan to task for swearing when describing a competitor. You know what? Good for him—this walking-on-eggshells, politically correct world in which we live today can be stifling. I enjoy hearing someone call it like it is. Precious few entrepreneurs are willing to publicly call out their competition, out of fear that doing so might hurt fundraising efforts. Less need to raise capital means less need to do or say the politically correct thing. Free the mind, and the mouth will follow. Or something like that. Exhibit A: Mark Cuban. He’s having fun.

(Speaking of having fun, just for goofing around sake, we conducted a fairly unique magicJack call—Dan’s laptop using a magicJack connected to the world via a Sprint 4G modem to my Mac using a magicJack connected to the world via my a Verizon MiFi. And it worked!)

While most people think “magicJack” (which is how I refer to them for simplicity’s sake) as the company itself, the actual parent company is named YMAX. YMAX owns magicJack, but they also own and operate a voice network, and own companies dedicated to software and hardware development, among other things. What makes this particularly cool is that Dan has assembled the entire end-to-end ecosystem to minimize risk in his business equation. While owning each component by no means ensures success, ownership enables integration on a scale all but impossible for the typical system developer—and rest assured, magicJack is a system, rather than a couple of best-effort Internet clients common in most VOIP offerings.

Case in point—magicJack’s silicon subsidiary has developed a chip integrating unheard-of capability for the cost. Don’t believe me? The femtojack’s core silicon contains technology for echo cancellation; serial flash; a CD-ROM emulator; controllers for GSM and the femtocell; audio codecs; and subscriber line (SLIC) and USB interfaces—and probably some other stuff I’m forgetting. When you do the math on the bill of materials to bring this type of device to market at a price of $40-$50, you begin to understand just how large-scale this integration actually is.

Right then. Back to customer service.

A year ago, I had no experience with magicJack. When I showed up at the It Won’t Stay in Vegas CES party, one of the pieces of swag was magicJack. I grabbed a couple, both for testing use in the lab and for use from overseas. During the sign-up process, I was dismayed to see a really lousy effort on privacy, with the user required to give up a bunch of personally identifiable information (PII) before even seeing magicJack’s privacy policy. That was bogus, and I called out magicJack, challenging them to resolve this.

Shortly after my post, I received e-mails from both magicJack’s PR person and from Borislow himself. I couldn’t quite figure out if I’d lightly struck a nerve or thoroughly pissed him off, but we ended up parrying in a conversation both on the blog and privately on e-mail. Dan stressed how much progress they’d made in improving their customer service ratings over the previous year, as well as their commitment to ongoing customer care. As most readers know, all too often this type of talk isn’t worth the paper it’s not written on—basically, it’s bullshit with a capital BULL.

Except, in this case, it wasn’t. Dan made the commitment that they were going to get better, and they did, going from an F from the Better Business Bureau to an A- today. That’s snatching victory from the jaws of defeat. Yow. You don’t make a transition like this overnight…so how did they do it?

Dan realized that he had a company-defining issue, one requiring an innovative solution that a typical call center couldn’t handle. As part of reinventing magicJack’s approach to customer service, he set some very aggressive goals—first and foremost, establishing customer communication in less than five seconds. How could a company with this type of growth rate achieve such a high customer satisfaction rate while keeping costs low enough to maintain margins and operational efficiency? Off-shoring.

But, not your typical off-shoring.

Most readers are familiar with the debacle so many companies have encountered when off-shoring technical support, leading to lost customers, negative press, and eventually re-on-shoring (perhaps coining a term…maybe “repatriating” would be better.). magicJack decided to offshore to countries with a wealth of competent English writers/readers—but to only offer support via online chat. Why? Economies of scale. An operator taking a voice call can address a single issue at a time. An operator conducting an online chat can handle up to five incidents simultaneously, maintaining the efficiency that magicJack must have to deliver quality service at a cut-rate price. Beyond that, the percentage of English-fluent readers/writers is (in Dan’s estimation) as much as ten times higher than the pool of English-fluent speakers—meaning, someone who has learned written English is a great tech support employee for magicJack, regardless of how well that person actually speaks English, or how strong a spoken accent might be. A bigger pool of applicants means better-qualified customer service agents, a more competitive internal support environment, and a better overall solution.

Dan had one of his team members walk me through their back-end support setup, allowing me to watch typed chat exchanges between real, live customers scattered across the globe. I must say, magicJack’s system is pretty sweet—efficient, effective, and extremely beneficial from a cost standpoint.

I mentioned that agents service multiple customer support chats simultaneously. Eavesdropping on the screen of an agent half a world away who’s conducting three, four, five simultaneous chats made me think of two things—instant messaging and intentional latency. Watching an agent move from chat window to chat window to chat window reminded me why I don’t use traditional IM clients, and why I only log into Skype when I actually need to chat with someone. If a bunch of chat windows were my job, I think I’d be great at it, but since it’s not, I’ll remain largely IM-free. “Intentional latency” may not be the exact term, but it’s a concept that’s stuck with me since taking a class from Jim Carlini at Northwestern two decades ago—the concept that an automated call distribution (ACD) system is as valuable for routing calls to the proper operator as it is in training customers to not expect the phone to be picked up on the first ring. Outside of Southwest Airlines, I can’t think of the last time I called any service provider who didn’t make me go through a phone chain to reach the proper person. (Thanks, Southwest.) I’m conditioned to deal with some amount of delay (and angst) before reaching somebody who can actually help me out. With magicJack’s agents supporting multiple chats at once, customers can see a delay of 10-20 seconds or more from the time they end up typing a response until the agent begins his or her own response. But, since that’s set as the norm, I don’t think customers do too much teeth-gnashing. Chats are initiated quickly, usually in less than five seconds; agents keep the customers engaged; and most importantly, customers anonymously grade the agent at the end of the chat.

Grades. Not just for schoolwork any more.

I love the grading concept. I wish that more customer service interactions could be graded immediately upon completion. A couple of weeks ago, United had a big fail at ORD, when their inability to get a jetway to a gate-arrived airplane for 11 minutes caused four of us on a flight to SFO to misconnect, even through the SFO airplane was still on the gate; based on the number of us who were ready to burst out of the starting gates to get off of our plane, I know that we weren’t the only four who ended up missing connections we should’ve easily made, but didn’t. United, you earned an F. A few days earlier, I’d pulled into a Marriott property with an unplowed parking lot, even though it hadn’t snowed in days. I don’t typically mention stuff like this to the hotel, but this was egregious, particularly since I ruined a new pair of shoes trudging in from the lot. The front desk agent shrugged off my comment about the lot not being plowed; even when I pressed her on the issue of why it wasn’t plowed, she showed no interest in providing an explanation. Marriott, you earned an F—but you raised it to a B courtesy of the property’s general manager dropping me a note within 48 hours of my website complaint, offering to pick up the cost of my shoes and put some extra points in my account. Here at the Marriott Wardman Park where Shmoocon attendees are currently enduring (?) the snowpocalypse, Marriott gets an A+, having kept employees on-property last night to ensure that they could meet the expected level of service for the ~1500 of us in attendance.

Anyway, back to magicJack support. I think they’re light years ahead of what most companies are doing—not all, but most. They’re using tools to not only manage customer service, but to measure customer service, then act upon that data. The idea of tracking agents’ output for purposes of reward, penalty, or compensation certainly isn’t new. But, the ability to combine the server-side monitoring of agents with customer-side innovation (such as setting a 24-hour tracking cookie on a client’s computer, and immediately escalating the chat to a “superagent” if the client initiates a second call in a 24-hour period) is compelling. Tying that together with the ability to remotely take over a customer’s machine via LivePerson to actually show them how to resolve the issue, or simply fixing it for them, makes for a further compelling solution—one which ideally leads to a customer grading an agent as an A.

The three call regions are staffed 24 hours a day, seven days a week, with an average of about 120 agents in service at any given moment; roughly ten or so of the highest-rated agents (superagents) are on duty at any time, ensuring escalation coverage no matter the hour. Average chat length is about 16 minutes; with agents supporting up to five chats at a time, the efficiency factor works out well. Sure, you might say that single-tasking a chat could result in a faster time-to-resolution; but, when you think about the time that troubleshooting takes (plug in your magicJack, make sure your phone cord is tightly seated, tell me what’s on your screen now, etc.), multitasking becomes much more attractive.

Even more interestingly, agents are compensated based on their call volume—and a grade multiplier. Meaning, an agent can’t simply rush through support calls to ring up a high number of calls complete, regardless of quality. The grade multiplier ensures that agents will provide due care to each caller (or at least, to the vast majority); sloppiness leads to low grades, which leads to less compensation, which leads to no longer being a customer service agent for magicJack.

I’m intrigued by this story and this model not because I got a couple of freebie units last year. I’m intrigued by this story because to me, this is really a case of “ain’t necessity a mother”. Or something like that. magicJack had a serious, serious problem a couple of years ago, when their Better Business Bureau rating was an F. Folks, it don’t get much worse than that. Memories are long, hearsay even longer. I had dinner at CES with a buddy of mine who’s in the VOIP space. He mentioned that he’d seen my femtojack post, and asked how I could respect a company whose BBB rating was an F. I told him that I couldn’t respect a company whose BBB rating was an F—and asked him if he’s bothered to check magicJack’s BBB rating recently. When I told them that magicJack had improved to an A-, he mumbled something along the lines of “they suck”, then ordered another drink.


Well, yeah. magicJack did suck. And, some portion of geekdom and the phone-calling public will continue to think they suck—but for a pretty high percentage of their user base, magicJack must be doing something right. Shelf space in Wal-Mart and Best Buy isn’t cheap to obtain, and is even tougher to hold onto if product return rates are too high. Units are obviously selling in, selling through, and being used with a decent level of customer satisfaction.

I’m hopeful that the femtojack will achieve similar or even more widespread success. No, I don’t know whether YMAX is going to be able to overcome the wrath of mobile carriers, and whether they’re on the verge of an avalanche of grief from the FCC.

But I can’t wait to find out.