It’s easy to break stuff. Just ask the parent of any 3-year old and they’ll confirm it for you. Fortunately, a child’s toys are relatively inexpensive compared to, say, a Porsche. The hottest fastest cars in the world have a redline on the tachometer. It’s there for a reason. It says “Sure, you can act like a 3-year old & wind this baby up. But if you overdo it, you’ll have junk on your hands.”
Which is why it mystifies me when I hear from test clients that set out to break their contact center systems; or maybe they say it a little softer like “See where it breaks…”
Our experience has been that this attitude gets things off on the wrong foot. Building anything is about building something that works. Testing its operation by intending to break it is a self-fulfilling prophecy. The problem is that it is so easy to break stuff that if you focus on breaking you’re likely to overlook what should be the objective of any test endeavor – demonstrating that it works doing what you designed it to do.
At IQ Services, we like to talk about our StressTest™ load & performance testing as a demonstration of performance, i.e., let’s figure out what the system’s supposed do and how it’s supposed to do it before we think about how to break it. If the usual load on the infrastructure that supports claim status inquiry is only 10% of all calls received, it may be interesting to run 100% of the test calls against that application, but is it useful? Why not set out to try it in the context of where it’s expected to be – with 4 other applications contributing 90% of the traffic and consuming the bulk of the resources? Does it still work? Or does the load eligibility put on back-end web services crowd out claim status when eligibility hits its target of 50% of all traffic? Wouldn’t that be more meaningful?
Demonstrating the performance of your contact center solution starts with knowing its performance objectives. Whenever we enter a StressTest™ engagement, we start out by asking a whole bunch of questions about intentions and expectations, about the business rules that influence handling customer transactions, and about the channels through which you intend to handle those transactions. This helps us focus our efforts so you efficiently & effectively use the budget you’ve allocated for testing.
Now don’t get me wrong – you don’t want to have blinders on and never push up against design margins. In fact you should. It’s just that you don’t want to approach testing with the attitude of a 3-year old – banging away at stuff until it breaks. Focus on the normal operating range & demonstrate to yourself and to the businesses that the solutions you’ve built work given the rules they’ve imposed. And then give it a little nudge.
Stay focused on what makes sense - demonstrate the performance of your contact center solution the way you intend & expect it to be used. You’ll uncover all sorts of meaningful issues and challenges, and you won’t be hammering away like a 3-year old.
Mike Burke
http://www.iq-services.com/
6601 Lyndale Ave South, #330
Minneapolis, MN 55423
Monday, September 28, 2009
Friday, September 18, 2009
HeartBeat™ by the numbers
I was fortunate enough to have an idea for a webinar accepted by the Contact Center Performance Forum (CCPF) a couple of months ago & the event came off without a hitch on August 6th. The topic was “Great Customer Experiences Start with Consistently High Performing Technology,” the expected audience call center managers – the people managers, not the IT wonks. I’m used to talking more about the performance of the machines in the contact center than the impact on the agents, but the hook for this webinar was you’re not going to have even half a chance at a positive experience if the technology required to handle & deliver the calls to agents falls down on the job. When callers finally do get through they’ve already been preconditioned with a lousy experience, and who will they take it out on? The CSRs of course! The moderator worked me over a bit and kept saying “Mike, you’ve got to make this real for people. Give us some real numbers. And don’t forget, agents are people too!”
So I did some digging and I was more than a little surprised.
One of the services we offer is surveillance for self-service solutions in production, the ones you often deal with before 0-ing out to get to an agent. This is our HeartBeat™ service. HeartBeat™ generates test calls one-at-a-time around the clock to access systems via the PSTN to ensure they are available & working as intended; if not, an automated notification is generated.
During a typical month we generate anywhere from half a million to 600,000 HeartBeat interactions. Would you believe that month after month, anywhere from 4% to upwards of 6% of those interactions encounter some kind of availability or performance issue? I know the HeartBeat value prop really well, but even I was surprised… 5% issues on average? Really??
Yes, really. Who knew?
Shan did. Shan manages our HeartBeat team. He lives this every day, along with Mike2 & Evan. They make up the team that defines the test cases, configures the servers, figures out what’s ok and not ok, who the system should call when there’s a ring-no-answer vs. a host down. It’s a whole lot more than just making a phone call and checking for answer.
So I asked Shan to tell me how things break down - literally – here’s what he told me:
Correcting for repeat issues (ones that last for a while and are therefore detected over & over again), here’s the distribution:
40% - Issue with answer – busy, ring-no-answer, silence or click
40% - Caller-requested information unavailable – host issue
20% - Caller disconnected prematurely
So out of 600,000 test calls in a typical month, 12,000 are answered incorrectly, or not at all. Another 12,000 are customers being led on a wild goose chase all the way to the point of finally being able to retrieve the info they need only to find out it wasn’t actually available. And another 6,000 callers are just getting cut off – they get to start from scratch.
But back to the webinar for a minute…
So here you are a call center manager working hard to keep your people and your customers and your business units all happy. You’re up-selling and cross-selling while cutting costs & keeping attendance high & training agents to deliver the best possible customer experience. And all the while you don’t know if the technology you’re counting on to take care of your customers & offload your agents is doing its job or not. A batting average of 350 gets you noticed in the majors and 94% is probably an A when grading on a curve. But how does that rate in the contact center?! Can you really afford 5% of your customer interactions going sideways?
Mike Burke
http://www.iq-services.com/
6601 Lyndale Ave South, #330
Minneapolis, MN 55423
So I did some digging and I was more than a little surprised.
One of the services we offer is surveillance for self-service solutions in production, the ones you often deal with before 0-ing out to get to an agent. This is our HeartBeat™ service. HeartBeat™ generates test calls one-at-a-time around the clock to access systems via the PSTN to ensure they are available & working as intended; if not, an automated notification is generated.
During a typical month we generate anywhere from half a million to 600,000 HeartBeat interactions. Would you believe that month after month, anywhere from 4% to upwards of 6% of those interactions encounter some kind of availability or performance issue? I know the HeartBeat value prop really well, but even I was surprised… 5% issues on average? Really??
Yes, really. Who knew?
Shan did. Shan manages our HeartBeat team. He lives this every day, along with Mike2 & Evan. They make up the team that defines the test cases, configures the servers, figures out what’s ok and not ok, who the system should call when there’s a ring-no-answer vs. a host down. It’s a whole lot more than just making a phone call and checking for answer.
So I asked Shan to tell me how things break down - literally – here’s what he told me:
Correcting for repeat issues (ones that last for a while and are therefore detected over & over again), here’s the distribution:
40% - Issue with answer – busy, ring-no-answer, silence or click
40% - Caller-requested information unavailable – host issue
20% - Caller disconnected prematurely
So out of 600,000 test calls in a typical month, 12,000 are answered incorrectly, or not at all. Another 12,000 are customers being led on a wild goose chase all the way to the point of finally being able to retrieve the info they need only to find out it wasn’t actually available. And another 6,000 callers are just getting cut off – they get to start from scratch.
But back to the webinar for a minute…
So here you are a call center manager working hard to keep your people and your customers and your business units all happy. You’re up-selling and cross-selling while cutting costs & keeping attendance high & training agents to deliver the best possible customer experience. And all the while you don’t know if the technology you’re counting on to take care of your customers & offload your agents is doing its job or not. A batting average of 350 gets you noticed in the majors and 94% is probably an A when grading on a curve. But how does that rate in the contact center?! Can you really afford 5% of your customer interactions going sideways?
Mike Burke
http://www.iq-services.com/
6601 Lyndale Ave South, #330
Minneapolis, MN 55423
Wednesday, September 2, 2009
Positive customer experiences
One of the things that’s really cool about working for IQ Services is the customer experience we get.
Because we test systems remotely – from the outside in, just like your customers use them – we very rarely go on-site. Not that we haven’t. But the vast majority of the time we just don’t do it because we don’t have to. The upside to this methodology is our customers don’t have to do anything to their systems for us to be able to test or monitor. They don’t have to get us security credentials to be onsite. And because travel & shipping aren’t required, it’s all very cost-effective.
The downside is we don’t meet our customers face-to-face. Not that that’s required in today’s virtual world. But it’s just friendlier when you can look someone in the eye, read their body language, and get to know them on a more personal level. Especially when you go through the wringer with them like we often do during performance testing engagements at 3am on a Saturday.
Now the exception to all this is trade shows. IQ Services exhibits at something like 14 or 15 trade shows each year. It’s our chance to find out what’s on customers’ minds, the new trends in the industry and of course generate new leads. But then there’s this other thing that happens almost all the time – I’ll be going through my pitch with a new prospect and somebody will stroll up, badge flipped so you can’t read the name, the way they always are. And as I get to the punch line or value prop, the “stranger” pipes in with something like “You should use these guys – they really help out a lot. Let me tell you what I learned…”
And then they go into their story, and start selling for us.
Like I said, it’s really cool! This happens at almost every trade show, sometimes more than once. Let’s face it…we have a very narrow niche and a really obscure service. When someone comes over to meet us & gets rolling with our sales pitch, it’s really neat. One year a very grateful customer wandered over, dropped to his knees & bowed down in front of our booth. That got some real attention! We had no idea who he was until he spoke - now that’s a customer experience worth savoring!
Mike Burke
http://www.iq-services.com/
6601 Lyndale Ave South, #330
Minneapolis, MN 55423
Because we test systems remotely – from the outside in, just like your customers use them – we very rarely go on-site. Not that we haven’t. But the vast majority of the time we just don’t do it because we don’t have to. The upside to this methodology is our customers don’t have to do anything to their systems for us to be able to test or monitor. They don’t have to get us security credentials to be onsite. And because travel & shipping aren’t required, it’s all very cost-effective.
The downside is we don’t meet our customers face-to-face. Not that that’s required in today’s virtual world. But it’s just friendlier when you can look someone in the eye, read their body language, and get to know them on a more personal level. Especially when you go through the wringer with them like we often do during performance testing engagements at 3am on a Saturday.
Now the exception to all this is trade shows. IQ Services exhibits at something like 14 or 15 trade shows each year. It’s our chance to find out what’s on customers’ minds, the new trends in the industry and of course generate new leads. But then there’s this other thing that happens almost all the time – I’ll be going through my pitch with a new prospect and somebody will stroll up, badge flipped so you can’t read the name, the way they always are. And as I get to the punch line or value prop, the “stranger” pipes in with something like “You should use these guys – they really help out a lot. Let me tell you what I learned…”
And then they go into their story, and start selling for us.
Like I said, it’s really cool! This happens at almost every trade show, sometimes more than once. Let’s face it…we have a very narrow niche and a really obscure service. When someone comes over to meet us & gets rolling with our sales pitch, it’s really neat. One year a very grateful customer wandered over, dropped to his knees & bowed down in front of our booth. That got some real attention! We had no idea who he was until he spoke - now that’s a customer experience worth savoring!
Mike Burke
http://www.iq-services.com/
6601 Lyndale Ave South, #330
Minneapolis, MN 55423
Subscribe to:
Posts (Atom)
