Agile at scale was one of the major themes of the Agile2012 conference. How a team can successfully spread its agile expertise from the few to the many without messing it up was one of the topics discussed in the keynote “Scaling Up Excellence,” given by Robert Sutton, professor of management science and engineering at Stanford University. “Any organization has a center of excellence or pockets of excellence, so in this case it would be one or a few teams or projects that are doing some form of agile,” he said.

When people do scaling of any kind, Sutton said, they have to treat it (to use an analogy) as a ground war and not an air war. “You just need to make one little change at a time, and it’s not anything you do quickly or instantly just by having a conference or getting a speech from the CEO,” he said.

Consistent with that outlook, Sutton said, is a great quote from a book called, “Scaling Software Agility,” by Dean Leffingwell. “Doing agile requires a shift in mindset. You can’t just read a book or attend a class; you will have to try out these new practices, practicing them again and again,” Leffingwell wrote. Sutton agreed that you have to actually live the mindset every day.

Especially critical for all kinds of scaling, Sutton said, is the tradeoff between replication versus localization. “This is about whether or not everything has to be cloned and be exactly the same, or if everything sort of follows from a vague mindset,” he said. Sometimes there are problems with the cloning model, he said. One of the stories he relayed was how Home Depot went to China and opened 12 stores. But more than half of them are now closed because, he said, the company didn’t understand that a do-it-yourself organization doesn’t work in China because they don’t have the do-it-yourself culture. In China, if you’re rich enough to buy something at a Home Depot store, then you’re rich enough to have somebody else assemble it, he said. “So that tradeoff between the degree to which everything has to be done exactly in a pure form versus customizing it for the local setting is always a big design tradeoff for all forms of agile,” he said.

Sutton talked about scaling up excellence, using examples to illustrate his point: from scaling up Facebook, to scaling up startups, to spreading innovation practices throughout Intuit. “From my perspective, the general challenge of how you spread excellence from the few to the many, either as you grow a company or spreading it across a large company or even a network of organizations, there are certain principles that can apply, as well as certain tradeoffs,” he said. “Replication versus local variation, for example, is a classic tradeoff. If you take the case of McDonalds…what things do they keep in the menu from the United States and what stuff do they have to put on or take off when they go to India or China? McDonalds does a lot of local customization.”
Linking hot causes to cool solutions was another idea he introduced. The idea is, if you just make a rational case to people, it’s hard to get them fired up about something. So the most effective thing is to first get them emotionally excited about something, then aim that emotion in the direction you want them to go. “Name the enemy, as Gloria Steinem once said,” he said. “It works for everything. Now, I’m not an agile expert but, as I understand it, one thing that you can do with software development teams is to name the enemy, which, in this case, could be the traditional waterfall method. So that can get the team cranked up for agile.”

Another idea Sutton spoke about is something called “connect and cascade.” This is when, once you get a pocket of excellence in a company, you then start spreading what they’re doing to others. “It’s sort of like this contagious disease type of model,” he said. “You start by training a small team, and then they train the trainers who then train others in agile. This is another example of agile at scale being a ‘ground war, not an air war.’ You don’t just put everyone in a big room and give them a speech and they’re ready to do it. It just doesn’t work that way.”

Sometimes people who write about scaling describe a three-step process, Sutton said. “The steps go from excellence to efficiency to expansion,” he said. “I think this applies to some things better than others, but I think that it applies to agile especially well.”

Sutton’s seven scaling principles:
1. Link hot causes to cool solutions.
2. Live a mindset, don’t just talk about it.
3. When it doubt, cut it out (or cutting cognitive load).
4. Use little things that pack a wallop, and the power of subtle cues.
5. Connect people and cascade excellence.
6. The mindset is the steering wheel and the incentives are the juice.
7. Don’t put up with destructive beliefs and behaviors.