Nostradufus

Our company holds an annual conference where customers from all over the country come to see and hear the latest goings on with our rather large software product. I get up and give a couple of presentations at this event, generally covering our technical evolution and our high level plans for the coming year. Being of the agile persuasion, I have advocated a convention of three releases per year with published guidance on what's likely to be included in each release over the next year. This lets customers into our view of the collective priorities but allows us to remain flexible as we roll along. This guidance gets republished after each release so there is always about a year of product roadmap visibility that customers know is likely to change.

So much for the back story. As I started preparing for this year's conference, I looked back at last year's presentations to remind myself what I'd said. I had one slide for each of the forthcoming three releases, with about four features per release, for a total of 12 feature predictions for 2008. Guess how many have turned out to be true? As I looked at the list I was quite surprised to find that not a single one of my 12 priority features had come to pass. Not one.

I sat back and pondered this for a bit. It's not like we hadn't done our three releases right on time with a slew of very well received initiatives that most customers agree are taking us where they want to go. Our development team has never been so well respected for what it has achieved over the last few years, and the consensus is that the challenge now lies with the customers to keep up with us after a decade of the reverse.

All I can say is that even when you make predictions that you think are safe, even when you give yourself room to move, even when you profess agility and despise those practices in your profession that seem to think the world can bend to the gantt chart rather than the reverse, even when you have the inordinate luxury of making your own deadlines and correctly frame your horizon scans as guidance and not promises, you can still fool yourself into thinking you know something that you don't.

The good news is that taking an agile approach means you talk to customers frequently enough and update them on what's happening and why frequently enough that you have no qualms about having a laugh about this with them because they, too, are focused on this year's progress, not last year's plan. I'm just glad my salary isn't tied to my ability to follow a plan, because I can't even follow one I created myself!