Some projects are not able to be planned, because they depend on innovation not yet available. As an example; the race to the moon. Kennedy said they would bring man to the moon, but at that moment in time, it was not clear what would be needed to do that.

But the US did it, and did it in the timeframe Kennedy was giving. How can you make such estimations, with so much unclear yet? Are there some ways to plan the unplannable?

5 Answers 5


You could look at rolling wave planning.

Under rolling wave planning, you only have a detailed plan for a small block of time, perhaps as little as two weeks. You also maintain a "cloud" of tasks that you know that you eventually need to do, but you don't know how to do them, when they need to be done, or dependencies between them. Tasks can be added or removed to this cloud throughout the project. You also have milestones and a final goal that you keep moving toward.

On a continuous basis, you revise your detailed schedule using known information. How you maintain that detailed schedule is up to you - it could be a WBS, a Gantt Chart, or something else entirely. However, it only contains a small segment of the project, that which is known at the time. As you gain more information and it becomes possible to do new things, you can move tasks out of the cloud into the detailed schedule, add tasks to the cloud, or delete tasks that aren't actually necessary from the cloud.

You can couple this with a stage-gate approach. A stage could be a time boxed interval or at the completion of a milestone, in which you review progress and determine if it's feasible to continue the project.

There are some blog posts about rolling wave planning: Johanna Rothman, PM Crunch, and Project Management Knowledge.

  • Interesting. But how do you estimate the cloud?
    – Mnementh
    Commented Nov 23, 2011 at 0:25
  • @Mnementh You don't. You just track tasks and milestones. There might be a larger schedule or deadline to add value. You might also want to track the current money spent versus projected value of the product to make sure you aren't spending more than the project will earn you. Conventional risk management techniques to track project viability and health can be used.
    – Thomas Owens
    Commented Nov 23, 2011 at 0:39
  • Sounds a lot like agile ;-) Commented Nov 23, 2011 at 12:17

When the project depends on innovation, it sounds like impossible to manage schedules. This is usually a well observed phenomenon. However, what is even more important to realize is that you should not even try to manage schedule!

Now, that sounds counter to any project management practice. But i will elaborate. On a normal day when you commute, you iterate over a few possible paths initially that tries to minimize the time it takes from reaching your source to destination. But when you are on a trekking, much more important than reaching home back in a predictable time, than to really worry about the fact that you don't carry any unwanted injuries; and definitely you want to bring predictability in schedule at the cost of such injuries.

This is not to say that you allow arbitrary lack of discipline for crucial projects; but rather you re-define what is important. The most important things you try to control and track is risk and total cost. Eventually you will see that while you could not predict schedule, you would still end up with what would have been your best case.

Here are some pointers to think about:

  1. Measure and manage risk - in research projects, we track what is the next biggest bottleneck which could potentially turnout to to be a show stoppers. Sometimes, even if the risk is not going to be a show stopper, it can be potentially be a cause for making a huge section of work irrelevant time and money to put is a waste.

  2. Experiment and learn - when uncertainties are abound one must try out somethings with a small elements and validate the proposition. Do not put any (or put least possible) effort and money which may or may be relevant in the light of above validation.

  3. Track total cost that schedule - time wasted is sometimes new learning so it is either cost or investment. But what is important is that efforts which has been made which is neither productive nor result in any learning, usually means money and time both wasted. Tracking only time against results in an unpredictable results often misses this point. which is very crucial.

  4. Plan often and plan to throw away plan - since new learning will transform how you think about the whole thing, iteratively keep planning again. In fact, most often, you should only plan for the next milestone. Yes, in many cases, customers won't be happy but you will realize that your plan needs to be very accurate till next milestone, where as next milestones are a bit hazy because they depend on the outcomes of the first one. So it is best as that you can provide gradually reducing list of milestone.

  5. Over communicate - when things are at stake, and world is uncertain, the worst night mare could be that you tie up yourself in the room alone. Many decisions needs to be taken, and as much as you invite all others in making decisions, and provide them the potential consequences of each decisions, it would benefit to digest (and recover from it) the potential disasters as and when happen. Also, adding more perspectives in decision making and communication helps over looking potentially disastrous problems.


Plans, like work, have layers of abstraction. If you have some grand, huge idea that is so complex and/or depends on things that have not yet been invented, it is as plan-able as planning a small landscaping project, just in a more abstract way. As you progress, you decompose those layers, driving your plan to a more tactical, more precise level of abstraction, to the point where you can execute. Using progressive elaboration as discussed earlier, your making pieces of your plan more tactical as you become more in the know, while leaving other portions of your plan still abstract.

You can set an abstract schedule, as well, just like Kennedy did. It is called a Rough Order of Magnitude target, one that has a known plus or minus variance. That schedule becomes more precise as you move along, moving to a notional estimate, then to a tactical target against which you would measure your performance.


High uncertainty projects

What you are talking about are ‘high uncertainty projects’. All projects are unknown to some degree, some to a high degree. Medical research projects are a good example.

Dealing with uncertainty is different from dealing with risks. And you still can and should plan.

Basically, there are three different types of high uncertainty projects:

  1. Where the end of the project is not clear. The first stages are clear the later stages are not. In this type of projects you still can plan the stages but there might be some major re-planning at a stage gate. An option in this type of projects is to make two projects out of one.
  2. Research projects. These projects normally deal with the ultimate uncertainty. In such projects you want to drop the project plan and only work with stage plans. This obviously impacts on the requirements for the stage gates. Key here is to manage expectations as it is not unusual that 9 out of 10 projects fail. But the one successful project is worth all the other ones.
  3. Projects with a feasibility study. There are two types of feasibility studies. A) Feasibility studies to confirm that you are heading in the right direction. This is similar to projects where the end is not clear. B) Feasibility studies to identify different options. In this case making a specific feasibility project can be helpful.

You need to be careful in chosing your project method. Some well known methods fail with high uncertainty, others force you to complicated work-arounds.

PRIME® offers you an add-on PowerPack to deal with high uncertainty projects. For details:



It depends, if the uncertainty is part of your project. If you can influence it, you have a lot of good advice already.

If you depend on inventions or progress of unrelated persons/companies, don't do a project based on this. Do a project with what is available now, and if something new becomes available, improve your project.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.