I visited scruminc. in Boston last September. Unfortunately, he was out in Zurich at the time, but he arranged time (in his hotel room at midnight) for me via Skype with the help of Laura Althoff at scruminc. I’d like to share the interview here.
Kenji: Thank you very much for meeting me, unfortunately via skype. I missed the meeting arrangement last week and you tried to pick me up at the hotel, sorry I mistook the meeting schedule completely one week off !
Jeff: No problem. You did miss a drive in my Tesla Roadster.
Kenji: Oh, no I missed a big one… OK, let me start with this question. First of all, why did you started scrum ? What was the main motivation that drove you?
Jeff: I was first inspired by what Accion was doing in their banking business. “Accion” is a micro financing bank, like “Grameen” Bank in Bangledesh founded by Professor Muhammad Yunus. They loan a little bit of money to a team of poor individuals running small businesses. The money might be used to buy a fruit cart to sell things in the town square or a sewing machine to develop a clothing business. Initially, buying enough food to eat may be the primary goal of developing a small business. After that, clothes are important to the poor because without clothes, their kids they cannot go to school. The loan of a small amount of money to people helps to bootstrap them out of poverty. Give a team of people a little bit and they can change their life dramatically. As I worked in my spare time on the President’s Council of Accion, I noticed in my day job that software developers had similar problems to poor people in South America. They had enough money but they never had enough software. They were almost always late in delivery and the quality was poor. As a result they were under constant pressure from management and felt like they were second class citizens. This was the first awareness that made me try to make a new way of creating software in a better way.
Kenji: When was it that you started thinking about how to make the world a better place?
Jeff: When I was at EASLE Corporation, I was leading a small team to develop a new product that would replace their legacy 4GL products. The development team had always been under pressure, managers were always angry, and customers were never happy. Jeff McKenna was a consultant on the team and we were always talking about why this happens and how to make life for people doing the work. And we found that the organization of the work is the problem. Management is usually hierarchical and driven by command and control pressure. Conway’s law says that the structure of software follows the structure of the organization that created the software. We were concerned that our current large customers like Ford Motor Company would misuse our new product, so we tried to find an “object-oriented” way to organize a team.
Kenji: Did you just start the Scrum way of organizing the team ?
Jeff: Not really. We initially formed a team based on my experience with Bell Labs and the MIT Media Lab. Small teams with no specialized job descriptions where found to be critical to productivity during a decade of research at Bell Labs. At the MIT Medial Lab, delivering really cool software every few weeks was mandatory or your project would be killed. So we set up a small team with one job description, Member of Technical Staff, just like they did at Bell Labs. We created a monthly Sprint with a demo in a Sprint Review at the end of the month. And then we started extensive research. We studied the business and software development literature and cases in which software productivity is more than 5 or 10 times the norm. After the team had read hundreds papers on teamwork and productivity, we encountered one of the most important papers for our new way of producing software — “The New New Product Development Game” by Takeuchi and Nonaka. When we encountered the paper, everyone was convinced that this was the one we were searching for, to describe the way a team should be formed and managed. The idea of new organization which is self-organized rather than hierarchical. We knew that hierarchical organization structure and micro-management makes a team slow, and we decided to use a sports model in which a team self-organizes to win.
Kenji: How did you put it into practice?
Jeff: The development of team with incorporation of the ideans of Takeuchi and Nonaka was completed by the end of 1993. The first Scrum sprint started in January 1994. The team decided to call the team leader a Scrum Master and used a cross-functional team model. But it was not good enough to change our life dramatically.
In the second sprint, in February 1994, we added the “Daily Scrum Meeting” to Scrum and that worked very well. At the time I read Jim Coplien’s paper (reprinted in “Organizational Patterns of Agile Development”, outline slides here) that did a Bell Labs audit of Borland’s QuatroPro software development team. The team consisting of 8 people was 50 times more productive than a Microsoft team. So 8 Borland developers were equivalent to 400 Microsoft developers. From Jim’s paper, I took the “Scrum Daily Meeting” and we fine tuned the meeting in the second sprint, reducing the time to 15 minutes, and developing the questions that each member of the team would talk about.
In the third sprint, the productivity went up 400%, compared to the first two sprints. This surprised everyone when we completed a month’s worth of work in a week. It was caused by the team questioning each others work. Was each person working on the item that would maximize team performance? Could they help one another to finish the work faster? Work that previously took days was reduced to a few hours by pairing.
Kenji: So may I call the Takeuchi and Nonaka’s paper is the main impact on for your work?
Jeff: Yes, I would say that Takeuchi and Nonaka’s paper worked as a catalyst of the whole background theory in our Scrum formalization.
Kenji: And how does Scrum make the world a better place?
Jeff: Current research shows that over 65% of people are unhappy in their work. They feel they are badly managed, have little control over their work, feel uncreative, and this slows down production and produces poor quality product. I though if people could take charge of their lives, and help set their direction, they would find empowerment in their lives, creativity in their work, and be excited about working with others to build great product. The cost of the products would decrease by a factor of 10, allowing more people on the planet to afford things they needed. This has proved to be an accurate assumption.
Kenji: thank you very much. I really enjoyed the interview and this stay in scruminc.
After the talk, Laura showed me a search result of a job search site that said that there were over 400,000 job offers for Scrum engineers… I felt sure that this was changing the world.
** This interview will be included in the book which Prof. Nonaka-san and I are co-writing now. The title will be “Agile and Scrum: Collaborative Software Development That Connects Customers, Engineers, and Management.”
(Edit Oct.4,2013: The book has been published early this year only in Japanese. If any publisher is interested in it, please contact me :-). I spoke about it several time so far and the slide is here.