These days, I view every customer service contact as an experiment.
If you've ever conducted formal experiments, you'll recall the concept of "the null hypothesis." That idea took me a while to grasp back in college but at the end of the day, it comes down to this: experiments don't really yield statistics that prove anything. In fact, if an experiment is successful, its results disprove something: the null hypothesis.
So, the null hypothesis is presumed to be true until the experiment's data show that hypothesis to be incorrect.
And, my null hypothesis for every technical support interaction is as follows: "the customer service experience I am about to have will demonstrate that this company sucks."
Don't ya' just love science?
So, I got a chance to gather some experimental data last night. Let me tell you about it.
Cable TV and high-speed Internet service for our New York City apartment is provided by Time-Warner Cable and their Roadrunner division. Yesterday, I had occasion to interact with both. The DVR in the apartment went kaput last night, endlessly re-booting. A quick conversation with the tech folks confirmed my fear: dead DVR. Solutions: wait for a tech or go pick up a new one ten blocks away. Great, I'll get one myself, thanks a lot.
Null hypothesis, Time-Warner Cable sucks, is disproved by the data. (Hey, it's my experiment, I can set the N anyplace I want!)
Well, that left me with no TV for the evening. Hey, no problemo, it's the Internet Age! There's tons of stuff out there to watch, even the Yankees game.
Ah, but this intense focus on the Internet forced me to come to terms with a nagging suspicion that had been creeping into my thinking recently: "there's something screwy about the Internet connection at the apartment."
Armed with time and interest (two-thirds of the magic triangle of success for any task), I sought to fill in the missing piece, talent. I needed to diagnose the quality of my connection, which isn't too hard thanks to a free service like this one. I ran the test and got these results: Download, 351 kbps; Upload, 96kbps.
Uh oh. Bad juju, bwana.
So, for the second time in two hours, I now attempt a second experiment, this time with the Roadrunner customer service folks. Herein, the lab report for that experiment.
Null hypothesis: Time-Warner Roadrunner sucks.
Data collection: I begin with the usual procedure, calling the customer service number, indicating that my problem is with high speed Internet service. I am immediately channeled to a voice recognition system that will attempt to solve my problem "using the same methods that our technical support staff use."
I think, "Uh oh."
But, I'm a compliant guy. For the most part, I follow the rules and go along with requests. I give it a shot.
First question: "Are you able to connect with the Internet?"
I'm already in trouble. Yes, I think, I can connect with the Internet but at speeds reminiscent of my Compuserve dial-up days.
Faced with this dilemma, I do what any compliant, rule-following customer would do, I say, "no."
This didn't help because the automaton on the line begins to take me through the process of rebooting my modem and my Airport router (yes, I'm one of those deviant Mac users), both of which I'd already done to no avail. In fact, I've taken the Airport out of the equation entirely by hardwiring the modem into the ethernet port.
So, I say, "yes" when asked if I've rebooted the modem.
Now, she says (why do they use women's voices on these systems? Come on, think about it. Most geeks are at least a little less likely to go apeshit ballistic on a female, even a robotic one, than on a guy, right?), "OK, shut down your PC and..."
OK, I think, "this isn't working." So, using a tactic I've learned from other despair-eliciting encounters with voice recognition systems, I say, "agent."
Now, here the Roadrunner system demonstrated some real ingenuity. Other systems I'd used simply complied with my request by saying, "hold on while I connect you with an agent."
Not Roadrunner; no pushover they! Instead, my automatic helper said, "I know you'd like to speak with an agent, but we're almost finished solving your problem."
"No," I think, "we're not even remotely close." So, I just started saying "agent, agent, agent, agent" in an increasingly agitated voice.
Voila! "Hold on while I connect you with an agent."
Now, you've already followed the details of the experiment's methodology pretty closely if you've gotten to this point, so I'll spare you some of the more arcane procedural bits. Rest it to say that the young fellow I spoke with took me through a Ping procedure to determine the number of microseconds it took for signal to make a complete loop from my apartment to some Google servers somewhere. (This after verifying the speed of my service using Speedtest.net.)
After reflecting on the meaning of the Pinged microseconds, this fellow says: "Well, your connection is fine. I suggest you contact Apple to resolve your problem."
[You know those cartoon characters who run into frying pans or doors or anvils? You know that look they have on their faces? OK, so you can now picture me at that moment.]
Stunned, but still conscious, I say, "wait, this configuration works flawlessly on Internet connections all over the country; in my home in Connecticut, in Starbucks all over the place, in hotels. You're telling me that the fact that it only doesn't work here, with your service, is an Apple problem?"
"Well, sir," he replies, "you have a fast connection with the Internet according to our test. The problem must be with your computer."
Now, I've been using computer systems since 1979. For a moment, I'd like you to consider the number of hours I've spent in technical support conversations during that time. Hundreds? For sure. Thousands? Maybe. And one thing I've learned as a result of those torturous hours: "don't lose your cool; it will only come back to bite you in the ass."
Deep breath. Another. Then, "OK, I'd like you to escalate me, please."
[We all know about escalation, right? You talk to first line support people and they follow tightly scripted algorithms to try to solve the run-of-the-mill technical problems that millions of users confront daily. Imagine the kinds of stuff that happens on Level One support hell.
But, I'm not a Level One problem kind of user. I've usually done all the stuff they want you to do on Level One before I called. But, because of the way customer support works, you invariably have to do it all over again in the company of the Level One support person just to prove that a) you're not inept, and, b) the Level One support agent will not get his/her ass kicked for passing you along to Level Two.]
There's a pause on the other end of the line in response to my request. Mr. Level One then asks, "you'd like to speak with a supervisor?" [At this point I get a vague feeling I've walked into one of Plato's Dialogues; this has to be a trick question.]
"Yes. Yes, I would."
"OK," says Level One Man slyly, "but, supervisors aren't technical, so they probably will also say that the problem is with your computer."
[Now, I can say with almost Thomistic certainty, that this was the moment that the data began to strongly favor not rejecting the null hypothesis. I can say that because that was the moment when I really got pissed off. In my experience, there is very strong correlational evidence for getting pissed off and companies sucking.]
"Well, let me speak with a supervisor anyway," I say, as calmly as I can.
Minutes pass.
A youngish sounding woman comes on the line and officiously says to me something like, "I understand that your computer is having difficulty connecting with the Internet."
[First word that pops into my head: "Kafka." "I'm doomed," I think.]
"Well," I say after gathering my wherewithal, "I am getting really lousy throughput and want to speak with someone with a little more technical know-how than the fellow I was just working with."
"Well," she says, "we only have one level of technical support. There is no one else to speak to about this."
[Recall here the image you conjured up earlier of me walking into frying pan. Repeat twice.]
"Wow. Well, how about this? How about you connect me with another agent and we'll take another shot at resolving this?"
"OK," says Ms. Level Two, grudgingly. "Please hold."
Fourteen minutes of irritating muzak later, on comes a new voice, Ms. Level One, saying: "Roadrunner Customer Service, can I help you?"
[Insert frying pan look Number Three here.]
"Um, did that supervisor just decide to not even bother to come back to speak with me?", I ask.
"Well," Ms. Level One says, "my information indicates that the problem is with your computer so she said I should speak with you but there isn't anything else we can do."
[Imagine if you will the terms in which this insane caller had been assigned to poor Ms. Level One? "A crazy man says it's not his computer! You talk to him!!"]
Calling on my decades of experience and now definitely as far down the tech support rabbit hole as it is possible for a modern human being to go, I say, "well, how 'bout you and I take one more shot at diagnosing the problem?"
"OK," sighs Ms. Level One, "but because there weren't any asterisks in the millisecond lines of the Traceroute report, there's nothing else we can do."
This, of course, could have just as well been Klingon rather than English, but it occurred to me that Mr. Level One and I had never, in fact, conducted a Traceroute test. We had only performed a Ping test.
So, I say to Ms. Level One: "Well, we never did Traceroute, only Ping." "OK," she allows, "let's do a Traceroute."
[For the record, I have now been on this call for one hour and fourteen minutes.]
Ever done a Traceroute? Kind of cool, actually, as packets of data scurry through connections trying to get to, from and through routers all over the net, and back. And, it turns out, if the data report from a Traceroute contains asterisks, there's trouble in River City.
I now needed to read Ms. Level One ever character on ten lines of my Traceroute report: "asterisk 169.37.135; 34.123 ms, asterisk"; stuff like that. Finally, I'm finished.
Ms. Level One then announces, "Well, since there are asterisks in your Traceroute report, I can escalate you to the next level of Technical Support."
[Remember those frying pans? OK, this time I want to you recall those rare moments when the cartoon character is standing in the middle of the desert and a nuclear explosion occurs. You can tell by the mushroom cloud. Great billows of smoke rising in that archetypal horrific shape. Eventually the cloud clears. A crispy, pathetic, singed creature stands before you. Eyeballs bulging, the creature blinks twice, in perfect synchronization with a sound effect like <clink, clink>, then disintegrates. OK, you've got a picture of the look on my face.]
"Wait," I say, "so when I was told that there wasn't another level of technical support, it was a lie."
"Well, we say that to everyone," says Ms. Level One, sheepishly.
"Oh, I see, so since you lie to everyone, then that makes it OK."
"I'll pass you along to Level Three now," she answers.
Of course, by this point it is obvious to me that the null hypothesis has been supported at the .001 level of significance.
I'll wrap this up for those still here. Mr. Level Three comes on. We diagnose my connection using Roadrunner's speed test. My download speed is bad an my upload speed hardly qualifies for the conventional use of the word; it's nonexistent. "I'll set you up for a tech appointment," he says. We agree on a date and time. He's all business and spot on. We finish up.
"Oh," I Columbo him, "one more thing. You might want to give anybody who cares a little feedback: it's never a good thing to lie to your customers."
"What do you mean?", he replies.
"Well, I was told that Level One was the only level of Tech Support you offer; that there was nothing anybody could do technically beyond them."
Pause.
"That's insane," he says.
"Yup," I agree, "it sure is."
End.
Results: The results of this experiment make it clear that we cannot reject the null hypothesis that Time-Warner Roadrunner Customer Service sucks.
Further experiments will be necessary to determine the exact cause of that suckage, but these results indicate that the cause is terminal leadership arrogance and stupidity. Lying to customers, especially geeks, is a sure way to ruin any company's reputation. And, come the revolution (i.e., freedom of choice of Internet Service Providers) you will be punished for it.
Recent Comments