The Agile community has long advocated self-organizing teams. However, the emphasis has been on how teams perform work, make technical decisions and the like. Most teams are still operating in the same traditional way when it comes to measuring project performance and the application of controls. If empowerment truly focuses on decentralized decisions and authority, maybe it’s time to re-evaluate how we empower teams from a financial and performance perspective. In too many cases we are still binding them to fixed plans.
My inspiration for this blog comes from meeting Bjarte Bogsnes (Vice President Performance Management Development, Statoil, and author of Implementing Beyond Budgeting) recently in Australia at ThoughtWorks Live events. After talking with Bjarte, I’ve been re-reading his book, which discusses how he helped eliminate budgets in several large companies. One of the insights he gained was thinking about the question, “What do we really use budgets for?” The equivalent question in project terms could be, “What do we use project plans for?” The most obvious answer to that question has three components (you may think of others):
- Co-ordination with other activities
- Financial controls
The insight that Bjarte and others had was that they were using a single number for multiple purposes and that the single number was causing significant problems. By eliminating budgets, monitoring costs and revenue in new ways, and creating a new set of relative performance guides, companies are breaking loose from the budgeting straightjacket, and improving performance. Maybe there is a parallel for projects.
Traditionally, managers look at three project measures—schedule, cost, and scope—and furthermore, they insist on meeting all three planned measures exactly, a virtual impossibility in today’s turbulent business environment. What if we look at three measures for each of these?
- Targets—desired business outcomes (usually a stretch goal)
- Forecasts—best current estimate of outcomes
- Constraints—the limits of the team’s authority
Let’s apply these measures to a project whose traditional cost “budget” was $250,000. We could have a target of $200,000 (might happen if everything went right), a forecast of $240,000 (our current estimate of the total cost, and a constraint of $275,000 (the team was authorized to spend up to this amount). Therefore, if the team delivered the project with the capabilities agreed to (value, not scope), with the appropriate quality of results, within the time “constraint”, then any cost between $200,000 and $275,000 might be considered acceptable. I say might, because only a holistic evaluation of the outcomes can determine performance. Give project teams more leeway with results, a lot more leeway, and performance usually improves.
One key determinant of project success is team motivation, and unfortunately most traditional project controls and measures try hard to “demotivate” teams. So even using the wider limits on the traditional measures above won’t be enough. Traditional measures tend to be of the stick kind (do this or else). Time and again studies, highlighted by Dan Pink’s work Drive: The Surprising Truth About What Motivates Us, show that the best motivators are intrinsic—purpose, mastery, and autonomy—focusing on positives rather than negatives. Better motivators for projects are purpose driven outcomes such as customer value delivery and quality (see the Agile Triangle for more info). Better motivators are relative comparison measures, not absolute numbers. Better motivators are those that allow the team more autonomy, and that includes leeway on things like cost, scope and schedule. Give teams vision, facilitate their self-organization, and provide constraints (loose boundaries), thereby encouraging them to be creative and innovative in their solutions.
Giving people stretch goals may motivate them, but when they are linked to potential punitive actions if not met, their motivation value is lost. Motivation needs to focus on vision and purpose, not penalizing people. When people think they will be judged against plans, they are likely to fudge the plans, or even sometimes the actuals. So break plans into two numbers, each with a specific purpose—a target that was based on business needs and represents a stretch goal, and forecasts (undated regularly) that were the best estimates of outcome.
But, someone always says, what about cost control (it always seems to be about cost control, not value delivery)? The answer for this is 1) to track actual costs carefully and watch trends, and 2) to establish an authority constraint (limit) for each project. This constraint should be generous, not onerous. Projects that are forecasted to exceed their constraints would need further review and possibly additional funding (or termination). The big difference here is the difference between a predicted cost and a cost constraint.
Project teams also need to coordinate with others (teams, projects, departments, etc.), usually about schedules. Forecasts are the numbers used for coordination. If targets are really stretch goals—with say a 50/50 chance of achieving them, then they shouldn’t base their plans on targets, but forecasts—our best estimates of outcomes. By using these two measures in tandem, both coordination and motivation are improved.
For this evaluation system to work two things have to happen. First, everyone, managers and team members have to understand the rationale behind each type of number. Second, everyone needs to evaluate the results holistically. Every number has a purpose, but performance is evaluated by taking all the numbers into account in a holistic manner. No single number, or really even a series of numbers, is adequate to completely evaluate a complex undertaking such as a project. Performance evaluation should focus on dimensions like value and quality, and secondarily on constraints such as cost or scope.
Any metrics system can be gamed. The success, or failure, of any metrics system lies in the intent of the managers and leaders applying the system. As Rob Austin, dean of the business school at the University of New Brunswick, relates in Measuring and Managing Performance in Organizations (1996), “If there is a single message that comes from this book, it is that trust, honesty, and good intentions are more efficient in many social contexts than verification, guile, and self-interest.”