Friday, April 29, 2011

Stumbling, falling, getting back up

The Tribal Knowledge project.
Oh yeah.

Late last year I was given the charge to manage a project to document troubleshooting processes that exist only in the minds of a handful of tech support gurus. It will help the support group to train new support people more effectively, and it will take the pressure off the gurus, who currently spend a lot of time solving the same problems over and over.

We put the project aside at the start of the year, because things were getting busy. After the really busy period, we met in February and agreed to cut down the scope of the pilot phase. People were still too busy to meet regularly to create the half-dozen flowcharts we envisioned as the pilot phase of the project; a lot had been postponed during the big thrash in January. We would do just three flowcharts. We met again a couple weeks later and cut the pilot phase to the smallest possible effort: a single troubleshooting flowchart, to be shared and evaluated, tweaked and published on the BlobCo collaboration site. I could almost hear an imaginary sportscaster exclaim, "Oh, that was a nasty stumble - let's see if they can recover."

A very skilled tech support person volunteered to build the flowchart, and emailed it out for comments a few days later. Someone provided constructive technical review comments. The flowchart owner got very busy and didn't have time to make the changes. I urged her to post it in its unfinished form and let others collaborate on completing it.

Nothing happened.
Imaginary sportscaster: "And we've got a player down on the field - it looked like she was going to recover from that stumble, but she's down."
What went wrong?

The concepts of working laterally, collaborating, and doing stuff because it's the next step in the plan instead of because the boss said "Do it" are central to this project. They're also markers of a mind-set that was once strongly discouraged at BlobCo. I could see from early on that the organization was recovering from a nasty bout of 1980s-style command-and-control, but I didn't realize how deeply it had scarred people. Everybody is too busy with their regular work to participate in this subversive, risky project, because "too busy" is the perfect reason — it's true, and it shows that people's priorities are right: We can't do this strictly internal project because we've got customers who need our help.
Because it takes us a long time to help each customer.
Because we've never documented and shared our troubleshooting knowledge, so we have to go ask somebody instead of looking up a troubleshooting flowchart in some central place.

We are like lumberjacks too busy cutting down trees to stop and sharpen our axes.

I asked the VP sponsoring the project to help me figure out how to get it back on track. He proposed a way: Authority would be exercised. People would be given assignments. There would be rewards at the end.
This didn't feel right.

Six of the couple dozen participants showed up at the next meeting. Several people had mysteriously taken the day off. I was frustrated. But the passive resistance and my frustration both validated my sense that we were taking the wrong approach. My sponsor bowed out and told me I needed to enlist a new project sponsor, suggesting two people who were well-placed to do this.

Time to seek out a Jedi master. I talked to Cat Herder.

"You've got too many people on this project. You need to drop the ones who aren't passionate about it," said Cat Herder, who went on to give advice that resonated strongly with my original sense that the project can only succeed if people WANT to participate. In just a few minutes, I heard several ideas about how to foster the creative, collaborative environment that will set people up for success; and we roughed out a plan for reshaping the Tribal Knowledge project and getting it moving again.
I am listening.

You have to fall down before you understand what it means to get up, and it helps if someone offers you a steadying hand.

I may have just connected with the next great teacher in my life.

Friday, April 1, 2011

You just might find...You get what you need

Sorry, I've had the Rolling Stones rattling around in my head all evening; hence the title.

A lot has been said about the Society for Technical Communication lately, so I trust you to forgive me if I don't beat that dead horse. Instead, I'd prefer to flog one of my own favorite equine corpses, Getting What We Need From The Software Guys.

Late last week I started what I thought would be a minor update to the product help for MumbleCo's flagship product, Big Software. One of the changes was that things had been streamlined. They had eliminated all cases of multiple paths to one task. Excellent! I slammed through all the task topics in the help and updated the navigation. Easy-peasy.

Except...

As I updated the top-level navigation in each task topic, I realized that in addition to the top-level changes in the software - tabs moving from one place to another - the entire user interface had been streamlined. Just about every gol-dang screen. Gone were the buttons that hid drop-down menus; now we just had clickable links right out in front of God and everybody. Extra clicks had been removed from nearly every task. This was a huge improvement in usability, but suddenly the help update stopped looking minor.

I have worked with many technical writers who would have stopped right there and gone up in a tight spiral, yelling variations on "How could they do this to us?" You know the routine.
"Why didn't they tell us they were making this change? Are they TRYING to mess with us?"
"OMG, they never tell us anything!"
"Code guys are so self-absorbed!"
And my favorite, "They have no respect for WOMEN!"
I skipped all that. Here's why:

Taking all these gripes in reverse:
I'm a woman, and no developer has ever shown me disrespect. So I have doubts about the whole misogyny thing. Next!
Developers have their own priorities and their own way of seeing the world, just as technical communicators do. Next!
They never tell us anything - OK, do we ever tell THEM anything? Such as how we work, what we need, what our schedules look like? Generally, not until we fail to get what we want. Next!
They didn't tell us about this stuff because it looked pretty trivial to them. Why should they bother us over something as trivial as taking a pull-down menu and making all the choices visible all the time? They don't know how we work, so they don't realize that's important to us.

So I skipped the yelling step and went straight to the next one - solving the problem.

The head of the software development team walked by my desk at an opportune moment. I hollered his name: "Code Jedi!"
"I didn't do it!" he yelled back, speeding up.
That's my line, but I didn't call him on it. "I know you didn't, and that's the problem," I hollered back.
That stopped him. "What's up?"

I explained that I needed information I hadn't been getting, and that I knew the reason I wasn't getting it was because Code Jedi's team didn't know that I needed it - I hadn't ever asked explicitly. As a consequence, I was trawling every single task topic in the help for Big Software to check it against the Big Software user interface, because a lot of things had changed while I wasn't looking.

I explained that while the changes streamline the tasks that people do using the software, those changes make the help inaccurate. People who are comfortable with software in general won't be troubled in the least by the changes, but those are not the people who resort to the help. If I don't know about a change in the sequence of things that people have to click, then the people who rely on the help to build their confidence will have that confidence shattered when the help doesn't match what they see on the screen. So I need to know about even tiny changes in the user interface. If it's connected to a reported issue, I'll find out, because I work from the issues tracking system just as the software development team does. But if it's so minor that someone fixes it on the fly, I may not find out.

Code Jedi got it. All of it.
"First of all," he explained, "We don't do ANYTHING unless we're told to do it. So it doesn't happen unless it's connected to a reported issue. But you're right that we're not giving you the information you need. What we can do is put a check box in the issue reporting form, to flag whether the fix affects the user interface."

"That sounds good, but what I'd really like is before-and-after screen shots."

Code Jedi thought about that. "That would be hard. The developer would have to have access to before-and-after environments. Or they'd have to remember to do the screen shot before they started."

"OK, I could probably get what I need if I could just get an 'after' screen shot reliably."

"We could do that. Make the issue tracking form include the checkbox that flags a UI change, and attach a screen shot if it's checked. That would not add a lot to the guys' workload. We could do that. And QA could check whether the UI changes and bounce the issue back if one of my guys forgets to attach the screen shot you need."

OK. That would give me everything I need.
It took five minutes of friendly conversation to devise a process to ensure that I always get the information I need. "Thanks, Code Jedi. There will be chocolate-chip cookies in the near future."

"Awesome."

Problem solved.
And I've got all the ingredients for chocolate-chip cookies on hand.