Interview with Jim Highsmith
Israel Gat: By all accounts, the first edition of your book Agile Project Management: Creating Innovative Products (APM) was very well received. In describing it, Harvard Business School professor Rob Austin stated, "Jim Highsmith is one of the few modern writers who are helping us understand the new nature of work in the knowledge economy." I myself often characterized the first edition as "a thinking man's guide to Agile." What transpired over the past five years that made you write a new edition?
Jim Highsmith: Primarily the increased interest in Agile at higher levels of management. When the first edition of the book came out, most Agile initiatives were being undertaken by "rogue" teams in organizationsteams that were trying Agile in spite of all the organizational roadblocks and antibodies. Over the past 34 years, Agile transitions are increasingly being driven by VPs, CIOs, and CTOs in big companies.
A second big factor that encouraged me to write a second edition was that organizations are constantly telling teams, "Be agile, be flexible, be adaptable," but also, "Conform to the scope, schedule, and cost plan." My contention for several years now is that for organizational agility to prosper, we have to change the way we measure successto a way that encourages rather than discourages agility. This thinking has evolved into something I call the "Agile TriangleValue, Quality, and Constraints" (schedule, scope, and cost). The goals aren't schedule and costthe goals are value and quality. The others are constraints. Even scope isn't the goal, at least not scope in terms of detail stories or requirements. At a strategic level, "value" is whether or not the team produces a releasable product that has enough capabilities to meet the customer's needs.
Israel: Noticeably, the second edition has specific chapters on advanced release planning, scaling, governance, and performance measures. Let's start with the first of these important topicsadvanced release planning. Various Agilists regard release planning as "Iteration 0." In contrast, you emphasize the importance of anchoring release planning as an integral part of the software development lifecycle. How do you reconcile the two approaches? How does Agile serve as an antidote to uncertainty when it's tied to the software development lifecycle?
Jim: This is part of dealing with larger projects, and projects within larger organizations. I have talked with too many managers and executives in Europe, the U.S., and Australia who are frustrated with their Agile teams who seemed to be completely focused on the next iteration and have little regard for determining the feasibility of the entire project. Release planning gives everyonemanagers and team membersa vision of what the entire project looks like. It's not detailed, but has enough information to provide an idea of what the releasable product will contain, when it will be completed, and how much it will cost. Estimating stories or features on a backlog isn't enough.
Israel: Moving on to the second new topic, scaling. You describe two dimensions of scaling: "scale up" to carry out bigger Agile projects, and "scale out" to implement distributed Agile projects. The analogy to computing styles is intriguing. A Wall Street firm might install a bigger mainframe to satisfy growing needs, or it might opt to distribute the workload across various servers. The demands on the firm's computing infrastructure, programming style, and skill set are quite different in these two routes. Would a similar phenomenon hold for "scaling up Agile" versus "scaling out Agile"?
Jim: I'm not sure how far this analogy goes. While there are some technical and architectural issues in both types of scaling, the emphasis in my book is on organizational scalinghow to grow large, self-organizing teams. There are three main design issues:
- Organizational design. A hybrid-network structure.
- Communications/collaboration design. Impacts distributed projects.
- Decision-making design. Maybe the most important, and an area virtually ignored by most.
Israel: You put special emphasis in your book on "portfolio governance" (as distinct from "project governance"). This emphasis nicely ties the operational with the strategic. In particular, Agile can provide solid operational data to the strategic process much more frequently than more traditional software methods can. Do you envision portfolio governance evolving to the point that we could metaphorically consider it as turning the Agile "dial" in real time?
Jim: Certainly, I use the analogy of projects with short iterations delivering chunks of work. A portfolio management process that uses an Agile concept should focus on shorter projects to enhance the flexibility of the entire portfolio, just like project flexibility. For example, it might be that phase 1 of a CRM system had very high value, but the next chunk of value comes from phase 1 of a new financial management system. The third-highest chunk of value comes from phase 2 of the CRM system. In a traditional portfolio management system, the entire CRM system might be implemented over, say, 24 months, rather than breaking it into three- to six-month chunks. By interspersing phases from higher-valued applications, the overall company ROI could be enhanced. Furthermore, the last phases of the CRM system might never be implemented, because they keep being supplanted by higher-valued applications.
Israel: Governance is typically carried out with a certain mindset about performance measurement. The one you recommend in your book is captured in the quip, "How many people go to work each Monday morning determined to perform poor-quality work?" How do we ensure proper oversight without losing the wisdom of this quip?
Jim: The approach to governance and performance measurement in Agile Project Management: Creating Innovative Products, Second Edition is in some ways tilting at windmills, because what I propose goes counter to what's done in so many organizations. Interestingly, in workshops and conference presentations I haven't really had people push back, as in, "I don't think you're right"; it's been more, "I'm not sure this will fly in my company." I really think people are searching for new measures of performance and governance that fit with agility, flexibility, and adaptabilitythey just haven't had another model to consider. If nothing else, I hope the models I've presented contribute to the growing dialogue about measuring success in a different way.
Israel: In a 2008 InformIT interview, Donald Knuth reflected on the assimilation of new methods as follows: "[T]he traditional methods stay with us. Since they're understood by a vast community of programmers, most people have no big incentive to change, just as I'm not motivated to learn Esperanto even though it might be preferable to English and German and French and Russian [if everybody switched]." While Knuth specifically discusses programming languages, his thoughts about adoption of "old" versus "new" are applicable to other domains as well. Do you see Agile adoption following the dynamics Knuth highlights, or is Agile adoption likely to follow a different path?
Jim: The part of the quote I would disagree with is that we have no incentive to change. From a management perspective, a March 2009 BusinessWeek article says it all: "There is no more normal." It's like the difference between Newtonian and quantum physics: One was adequate for its time, but not for the future. In this same article is a quote by John Chambers, CEO of Cisco, that the biggest mistake he made was not moving fast enough.
I think most changes follow a similar path, but the key to organizational change is persistence, which comes from real incentives at all levels. Management sees the incentive of greater value, shorter schedules, etc. Developers and testers see the incentive of higher-quality workthey see a better way to do their jobs. Very few developers and testers who begin using Agile methods want to go back to the old ways. This is in stark contrast with the heavy methodologies of the 1980s and early 1990s (in which I was heavily involved). Most developers and testers didn't like the heavy methodologies because they actually put up barriers to progress, and they didn't reflect how people really worked. Most developers and testers like Agile because the technical practices are geared to themto improving their ability to produce results, not impeding them.
Israel: In APM, you discuss the Agile revolution. To quote you: "In industry after industrypharmaceuticals, software, automobiles, integrated circuitscustomer demands for continuous innovation and the plunging cost of experimentation are signaling a massive switch from anticipatory to adaptive styles of development." Revolutions, as we all know, might or might not succeed. With respect to the Agile revolution, sustainability is a hotly debated issue. Ken Schwaber (co-creator of Scrum) recently stated: "I estimate that 75% of those organizations using Scrum will not succeed in getting the benefits that they hope for from it. […]The intention of Scrum is to make [their dysfunctions] transparent so the organization can fix them. Unfortunately, many organizations change Scrum to accommodate the inadequacies or dysfunctions instead of solving them." Obviously, the Agile champion needs to be very thoughtful about change management. What would your recommendation to the Agile champion be? How could he or she enhance the prospects to achieve a long-lasting Agile transformation?
Jim: Well, I have to admit that I disagree with Ken on a number of issues, this being one of them. Not everyone will get all the benefits they might like, but a large number will get significant benefits. What might seem to be an inadequacy or dysfunction in one organization might be beneficial in another. I've been around organizational management long enough to know that there is no silver bullet, and there may be multiple solutions to an issue, given the context. How a company in China deals with an issue may be very different from one in New Zealand or Germany. In any Agile transition there must be a give-and-take, sometimes adapting the organization to Agile practices, and sometimes adapting Agile practices to the organizationthere's nothing wrong with that. I've seen too many instances of companies tying themselves in knots over some consultant's "religious" approach to an Agile methodology.
I've always said that the "art" of Agile management is having the judgment to balance conflicting issues, and the good Agile champion has to have good judgment to know when to push and when to back off.
Israel: We are really getting into deep cultural issues here. William Schneider discusses control, collaboration, competence, and cultivation as four core cultures of an organization. Obviously, the Agile champion usually operates within the norms of the single core culture that his or her organization has developed. How could/should an Agile champion anchor the revolution in each of these four cultures?
Jim: I deal with this subject at length and in depth in my book Agile Software Development Ecosystems. In short, my opinion is that more methodology implementations flounder from failing to account adequately for an organization's cultural factors than for any other reason. A particular culture isn't necessarily change-tolerant or change-resistant, but it may resist certain types of changes and embrace others. Culture provides both the strength to persist in the face of adversity and the weakness to persist in the face of folly. Any attempt to implement a methodology requires that an organization analyze its own fundamental cultureits core values. Many methodology failures are caused by a problem definition followed by a solution design, with little analysis of whether the solution design fits the company or the project team's culture.
Cultural relativism means that, again, cultures are not good or bad per se; they just are. However, some cultures may be better or worse at addressing a particular opportunity or problem type. They may also be better or worse at attracting talented developers. The critical issue for methodology selection is understanding one's own culture, the ecosystem in which a methodology must be implemented. Without that understanding, whether one uses Schneider's model or some other, methodology implementation faces hurdlesones that are present but often nearly invisible.
Israel: The quantitative studies cited in your book indicate impressive gains in software quality, productivity, and time-to-market through Agile methods. What are the risks that accompany such gains? What risk-mitigation strategies would you recommend to the executive who embarks on an Agile initiative on a large scale?
Jim: There are two pieces of the Agile equationmanagement agility and technical agility (per talking with Josh Kerievsky). The big gains come from companies that implement both sidesself-organizing teams, release planning, iteration planning, and so on; and unit testing, simple design, test-driven development, etc. I think the risks are both in trying to do too much too soon, or doing too little. Some try superficial Agile, just doing a couple of easy practices like daily standup meetings, and then wonder why they don't get the benefits. Others try too much without adequate training and coaching, and flounder in the chaos of change. Balancing between too much and too little is part of the transition process that doesn't get enough attention.
The other big problem with large-scale transitions is not allowing for the learning curve, and the fact that performance may get worse before it gets better. Managers often expect too much too soon and don't have the persistence to reap the really big benefits of Agile. The learning or change curve also needs to be taken into account when planning a large transitionhow much change can you afford, and how much do you want to spread out the cost?
At another level, done the right way, an Agile transition can be free, to paraphrase an old quality book, Quality Is Free, by Phillip Crosby. The "free" part comes from using management agility to closely connect to customer value and reduce marginally valued features. Since Agile can deliver value incrementally, the very important question becomes, "What do you want100% of the value for 100% of the cost, or 93% of the value for 75% of the cost?" When you have the discipline to alter the value-cost ratio, you can pay for both an Agile transition and putting higher quality into the product.
Israel: You are one of the people who conceived the Agile Manifesto in 1991. May I ask you to do a quick "retrospective" of the period since you signed the Manifesto? What have we as a movement learned over the past eight years?
Jim: First, I think the Agile movement has been a tremendous success and brought significant benefits to companies worldwide. While there is inevitably some dilution of Agile practices as more organizations jump on the Agile bandwagon, as a whole the transition has been positive.
That said, one of my concerns about the Agile movement is whether or not we ourselves have learned to adapt to changeone of the basic tenets of agility. In any movement like this, there is a tendency toward a religious-like fervor to do things the right way, the Agile way, but it can morph into "my" Agile way. For example, in the earlier years of the Agile movement there were questions about whether feature-driven development was Agile, and today there are questions from some about Kanban. To me, if I look at the underlying concepts of Agile, both are clearly compatible. If the Agile movement is to continue, we have to better understand what the core Agile principles really are, and not just our personal interpretation, and then find ways to incorporate thoughts and ideas that may seem in conflict with our own ideas. Just because some Agile camps may have a more widespread audience, that doesn't make them the source for all things Agile. The essence of change is tolerance for new ideas that conflict with our own.