The mathematician George Box is credited with saying, “all models are wrong; some models are useful.” The so-called “triple constraint” is one of the oldest models in the project management discipline. We all know it is wrong. Is it still useful?
“Pick Two”
There are many versions of the Triple Constraint. The way I first heard it was this: “You can have it good. You can have it cheap. Or you can have it fast. Pick two.” There is a certain amount of intuitive validity to this model. If you want the project done well and quickly, it isn’t going to be too cheap. If you want it done cheap and quickly, it isn’t going to be too good. If you want it done well and cheap … the model falls apart here. No matter how much time you give me, I will not be able to produce a good result at a low cost. This version is not very useful.
The “Iron Triangle”
A second version uses a triangle with the three sides labeled scope, cost, and time. This model suggests that changing the length of one side must be accompanied by a change to one of the other sides in order to maintain the triangle. So an increase in cost means either more time to spend the money, or additional scope in the same amount of time.
But this version of the model allows for some less than useful combinations:
- More time and more cost could produce the same scope. This does happen, but it’s hardly a lesson we want to promote!
- More scope and more time is possible with the same cost. But only if the cost was inflated to begin with. Another lesson to be avoided.
There are variants that replace scope with either quality or performance, but neither of these changes address the logical weaknesses of this version of the model. So in my mind, the Iron Triangle isn’t particularly useful either.
What’s Inside the Triangle?
Some authors label the area inside the triangle as quality. With this model, more time and more cost would deliver the same scope with more quality. That certainly seems reasonable, but quality-in-the-middle also means that more scope can be delivered with a longer schedule at the same cost and with more quality. Nope. Still some fatal errors.
Harold Kerzner, in Project Management: A Systems Approach, shows resources inside the triangle and replaces scope with performance/technology. He also surrounds the whole picture with a ring labeled “within good customer relations.” Kerzner’s inclusion of the customer is, I think, a positive addition to the model. His graphic also contains the thoroughly delightful implication that extending time or cost could puncture the circle of good customer relations. Unfortunately, resources-in-the-middle faces the same problems as quality-in-the-middle: some combinations of changes work; others don’t. So Kerzner’s version of the Triple Constraint isn’t very useful either.
The Three-legged Stool
Among others, Max Wideman has suggested that we should look at the Triple Constraint as a three-legged stool. This variant of the model emphasizes the need to keep things in balance across all three dimensions rather than making an adjustment to one. This idea certainly fits with my experience as a project manager, but many of the combinations of leg-length adjustments don’t work that. For example, a decrease in cost (one shorter leg) would require a reduction in both scope and schedule. The stool analogy falls down.
I’ve seen other attempts to create a more effective model. Some replace the triangle with a rectangle and the extra side is usually risk or quality. I’ve also seen pentagrams and one hexagon, but none were logically consistent once you challenged the model.
A More Effective Representation?
Despite its obvious shortcomings, the Triple Constraint remains a powerful image within the project management community. So I decided to try to come up with a version of the model that would work for me.
I was determined to keep the triangle. As most of you will recall from high school geometry, the triangle is a stable structure: change the length of any side and you must adjust something, either a side or an angle, to keep it a triangle. Whether you are talking about trade-offs or maintaining balance, the triangle works better than any other shape.
The next thing I did was add the word constraints to each of the traditional labels. Since we are talking about the Triple Constraint, this seemed fairly obvious to me, but I’d never seen it elsewhere. I later made several adjustments:
- Time constraints became schedule constraints to recognize that this dimension included more than just the total project duration.
- Scope constraints became product constraints to minimize arguments about the meaning of scope.
- Product constraints got a subtitle. I noted that this side might also be called scope, quality, or performance.
- Cost constraints also got a subtitle: funding, staffing, resources. I did this to recognize that there is more than just money involved along this dimension.
I now asked myself: what exactly is being constrained? After several discussions with colleagues and clients, I came to the conclusion that it was the project’s ability to satisfy the stakeholders’ interests that was being constrained. So I labeled the inside of the triangle stakeholder interests. If the constraints are too tight, the area of the triangle will be smaller, and fewer interests will be satisfied:
- Trying to deliver too much product in too short a time will require the team to work overtime: the team’s interests are not met.
- Providing inadequately skilled resources will produce a less than satisfactory product: the customer’s interests are not met.
- Missing an arbitrary completion date will reduce the benefits realized: management’s interests are not met.
The geometric logic of the model works equally well for other variations. For example, easing the cost constraints by providing more funding or more skilled staff while keeping the schedule and product constraints constant changes the shape of the stakeholders’ interests. Even if the area remains constant, it is likely that some stakeholders will be more satisfied and others less.
Although I was generally satisfied with this version of the model, I made one final adjustment a few years ago … I flipped the triangle so that the point was on the bottom. This allowed me to visually represent the need for balance from the stool model. In my final version, the Triple Constraint (plus one) is not about picking two. It is about establishing and maintaining a balance in order to satisfy the stakeholders’ interests.
And isn’t that what project management is really all about?