Would you deliberately redline a Porsche?

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