Dear PM Advisor. August 6, 2012

Dear PM Advisor, How do you best manage R&D projects where the scope and requirements are fuzzy and change is a common-place occurrence?

Stumped Scientist

Dear Stumped Scientist,

As an R&D engineer for fifteen years, I struggled with the same question. Before I fell into Project Management I used to laugh at PMs who tried to predict when we would reach a successful conclusion to our experiments. I even had a Director who's favorite saying was: "You can't predict discovery."

But if a company is going to be successful, it must be able to manage R&D projects in a similar manner to Development or Process or any other project. That means the steering committee must be able to promise a certain amount of time and money to an R&D group in exchange for an agreed upon scope. So how do you do that when you don't know when your experiments are going to yield the results you're searching for?

My take on my old Director's quote is: "You can't predict discovery but you CAN manage it." I'll show you how to manage it.

Let's divide up R&D into the two functions: Research and Development. In pure Research you are given x dollars per year to play around in your field and spit out things that the Development group can turn into products. This is an ongoing operation and does not fall into the realm of Project Management.

Development projects can be further broken into projects that have a defined scope that can be managed like any other project and those that still require a lot of testing. So let's focus on this last group. How do you manage these?

I'll give you an example of a project of my own that I had to manage in this way. I was working for a company that made vascular grafts out of Dacron. DuPont used to sell us the fiber but they wanted to get out of the Medical Device business due to the risks they perceived from silicone breast implant lawsuits. We found a new vendor willing to supply us the raw yarn but we had to prove to the FDA that the grafts we made out of this material were equivalent to those we'd been manufacturing for years. The process of converting raw yarn into complete Medical Devices took about eight weeks.

We would have to run this yarn through the full process, sterilize the products and then test them before we would find out how they did. We didn't expect them to pass the first time. We would test the products, make adjustments to the formula, then run the process again. While we could plan every activity in the eight week process in exquisite detail, we had no idea how many rounds of this process would be required. I talked to the technical expert and he hoped that we might find the answer in about five iterations.

What I decided to do was plan for six iterations, each eight weeks long, with a two week analysis time between each. I presented this to the steering committee, asking for 1.6 million dollars and two years to reach completion. They asked me to promise that I would succeed in six iterations. I couldn't do so. I told them I would keep them informed of the results of each analysis so they could see that the team was making progress. In the end, they had to agree to the terms of the project.

So that's how you manage research projects. Plan all the steps you can and manage the rest. Make sure they understand your assumptions, especially the important one about how many iterations it will take to reach success. Understand that the steering committee may well say no to your project and choose to invest their project dollars elsewhere.

Good luck,

PM Advisor

Send your questions to