Q4. Tell me about your current venture? A start-up ... that's a big change from SAS.
The startup I'm at has a very interesting idea. It's about creating an extreme level of transparency and then using that information to measure progress. We look at information from the team's source code management, issue tracker, feature tracker and even the IDE.
At the end of the day I can look at a team and see that they're spending too much time in the debugger, so I can talk about continuous integration and test automation as a way to keep people out of the debugger and in code. Another team might be reading code all day, trying to understand it. Then we talk about the peer code review or pair programming to get training levels up.
We automatically create burn up charts and then use them to teach managers why feature creep is bad. They can instantly see how the team can't catch up with an ever growing scope. Or how the team committed to this feature set but only delivered half of that. Then we pull up a "difference from estimation" report to see which features had bad estimates.
And at every step, because we automatically harvest real data in the background, we've got the actual time spent on each feature. Instead of guessing how much development time was spent over three days, we can say it took 12.74 hours to create that feature.
But this level of transparency scares a lot of people. Our industry has been very closed traditionally. Our managers generally don't understand what we do or how we do it. And as a result, most projects fail. Some from bad estimation, others from bad management. Sometimes it's because no-one realizes there's a problem until too late to help. The last numbers I recall seeing were a 75% failure rate for the industry. We've got to step back and see if we can't find a better way to build software. Many of the agile practices do just that. 6th Sense Analytics makes an amazing amount of information available and we're starting to see some pretty amazing insights come out of that data.
A tool like 6th Sense can be abused by a bad manager and it's a fundamental culture shift for many shops. But on the teams with the courage to make the leap, we've found both development and managers come away with a better understanding of the other does.
Why am I at a startup? If you saw my first question and it's answers, you know I like to tinker. And I get to do a lot of that here. We have working software, but there's still a wide variety of work to be done.
Comments