Thursday, December 22, 2011

Known ways of managing ET #03 - The Gamble

tl;dr – exploratory testing can be a gamble. And what's the problem with that?

As I write this, it’s coming up to Christmas*. Christmas is a hard deadline**, but there are clear gift-related requirements from one’s nearest and dearest***. As weeks turn to days and days turn to hours, some people find they have no gifts, yet no appetite either for handing over socks or gift vouchers. So they gamble. They give themselves some time, and head off to hunt for presents. Sometimes, not everyone gets a present, sometimes the presents are junk. Sometimes, they’re inspired. Generally, after spending roughly the allotted time and just a touch more than the allotted budget, the giver has something approaching the right number and selection of gift-ish things. We’re all familiar with a gamble, we know that some are more comfortable with a gamble than others, and we know that others positively delight in the last minute sprint-and-bundle.

Exploratory Testing is in great part about discovery. If you’re looking for real surprises, it’s pretty pointless to say how long**** you think it might take you to find them. It is more rational to set limits on how long you’re going to spend looking. This is a gamble. Your budget isn’t necessarily set to somehow match the value of the stuff you’ll find, but may be rather more influenced by what you have available to let you look. If you’re comfortable with a gamble, you may be comfortable with managing ET by lobbying for a budget from the project, and working to find a great way of spending that budget for the project.

Time for a couple of examples: “We want to spend 40 hours in the first week after delivery looking for trouble” is a gamble. “We’ll need someone for three days to prepare the exploratory environments, data, tools and ideas” is an investment in a known deliverable. These aren’t opposites. You will have goals for both. You can schedule both. Nonetheless, some test managers refuse to take gambles to their project managers.

This is not always because the project manager is genuinely uncomfortable with a gamble. Testing is a Wicked Problem, and no one thanks you for dropping a Wicked Problem into their basket of responsibilities. However, many project managers of my acquaintance are rather good gamblers, instinctively and analytically weighing up risk and return. Their management talents extend well beyond the business of juggling durations and dependencies. If you’re comfortable with a gamble, candid about the realities of ET and convincing about your skills, you may well see the PM’s eyes light up. The game is on. You’ll get some proportion of what you’ve asked for, and you’ll go looking for problems.

Of course, managing ET doesn’t end here. But – crucially – this is where it can start.

* or whatever you call it in your tribe.
** although some years I have good reason to send New Year cards.
*** and from those who are remote and unfamiliar, come to that.
**** I’m using long and budget here to mean the aggregate of time, money, people and so on. I don’t simply mean time or money. And we’re all aware that there isn’t a simple equation to switch back and forth between money and people. P ≠ mc2

No comments:

Post a Comment