TL;DR

• Estimation evolves through four stages as your project gains definition
• ROM estimates suit early alignment; detailed estimates require a full planning cycle
• Assumptions are as important as the numbers — always document and communicate them
• A Step 3 estimate higher than Step 2 is a common pitfall due to increased scope visibility
• Rolling wave planning works well for medical devices, but other approaches can too

From the stone age past to a sci-fi future, humans estimate things constantly: Do I have enough food to get through the winter? How hard should I toss this spear at that deer? Does my AI powered robo-dragon need an oil change? We do this to inform many of the 2,000 decisions we make per hour [1] (fewer if you have a pre-schooler who tells you what to do all the time) [2]. Without estimates, we simply can’t make informed decisions.

The same is true for product development: we need an idea of how much time or money various options might take in order to choose the right pathway for our situation. An experienced developer can make some of these decisions in a second, leaning on their past failures and successes to guide them. However, even the most experienced folks can only utilize their gut instinct to a certain extent. As projects get more complex, more time and more tools are needed to provide accurate estimates (for example, the James Webb Space Telescope took years just to estimate).

This blog provides an overview of the major steps involved in estimating cost and schedule. Though this is specifically based on medical devices, similar projects across a variety of industries will follow similar steps.

Step 1: Align on Scope and Surface Risks Early

The first step is to quickly surface any reasons the program might not work: cost, schedule, resource constraints, or significant risks.

It takes a fair amount of instinct and experience to know what questions to ask and what details to focus on here, but as a general rule, at the early stage you want to start with deliverables to help define scope (what are we producing exactly?), which will quickly lead to a conversation about resources (which people are we using? Do we have the right equipment?). Lastly, at this stage we should be able to understand major risks (what could derail the project, and how long will it take to realize or retire that risk?).

From this, we can use our estimating toolkit (see below) to provide a ROM [3] cost and schedule range (experience is telling me this program might be 9-18 months and $50,000 – $150,000, with the key risk being whether or not we can find an off-the-shelf motor). With that early ballpark estimate in hand, we can align with stakeholders on whether it’s worth spending any more time estimating. If so, proceed to add more detail.

Step 2: Frame a Range to Get a Decision

Here, we are looking to get a tentative “go” decision on moving forward with detailed planning. Often, decision makers will be looking to decide between options, and thus this “go” decision is to say “assuming your initial estimates hold, and nothing else changes, we have chosen this project and will commit the resources.”

Generally, a range that’s about 1x-1.5x should be appropriate to support this. A significant number of assumptions will need to be made to frame the estimate as well. Depending on expectations from stakeholders, a good strategy is usually to position the nominal estimate near the 1x side, with maximum contingency and risk accounting for the 1.5x side. Where possible, specifically naming the factors that could drive the estimate toward the maximum helps build confidence and prevents the awkward situation where approvers expect the 1x outcome and executors expect the 1.5x.

Realistically, getting a 1x-1.5x estimates is only possible if you’ve done a similar program before (see the toolkit). If you don’t have that, you’re unlikely to be able to get the accuracy needed without getting into detailed planning, per Step 3. Either way, with approval from stakeholders it’s time to continue to Step 3.

What if my Step 2 estimate is way higher than my Step 1 estimate? It happens! Remember Step 1 was based on a gut feeling and a meeting or two, so it’s reasonable to expect it will change. Before this point, no real commitments had been made, so the impact is very low.

Step 3: Detail the Plan Decision by Decision

Once we get the go ahead, we can use a planning-estimating cycle to build the detailed plan. This is where we start to add details. Each time we do, we can estimate its cost using our toolkit (see below), and then decide whether or not to include it in our overall plan: we could include wheels, but that would be an extra 3 months of effort, so let’s plan for rigid feet. After a dozen of these little decisions, a fleshed-out plan will form that hopefully meets your needs. The first few times you do this, you’ll notice that the estimate has grown considerably from your initial range. With experience, this gap will slowly disappear or even reverse.

After detailing the plan, a final review should be held with all stakeholders. Ideally, there will be no surprises here. The discussion will then turn to logistics: when do we start, how to we transfer funding, what’s our first milestone, etc.

What if my Step 3 estimate is way higher than my Step 2 estimate? This is unfortunate, but it is often the reality that as we dig into details, the true scope of work reveals itself. A good way to communicate this is to explain which assumptions are proving false: we assumed we would be able to find a low-cost partner for our software, but after digging into it, we found the vendor can’t support us, so we now plan to go in-house.

Step 4: Lock the Near-Term and Start Moving

The next step is to slice off the closest chunk of your plan and add the specific logistics: who will do what, in what order, using what equipment, where, how often will we meet to talk about it, who’s on vacation next week, etc.? After that, what happens next depends on your organizational structure: some organizations will approve the whole thing with regular check ins. Others will want to keep the leash tighter, releasing fundings in short term chunks only. In the latter case, how much to bite off is a balance: too much and your risk increases as you try to plan farther out, but too little and you will constantly be going back to your stakeholders for approval in small slices.

Once all that is in place, it’s finally time to start. Make sure everyone knows that their role is, how best to work together, and other kickoff best practices, and then get going! The world isn’t going to change itself.

What if my Step 4 estimate is way higher than my Step 3 estimate? This shouldn’t happen, because Step 4 is only supposed to be adding logistics and firming up short term plans. That being said, a living estimate should be kept throughout the program, and discoveries made underway that impact and refine the living estimate are common. Make sure you understand and communicate these as you go.

References

[1]: How Many Decisions Do We Make Each Day? | Psychology Today

[2]: Not a real fact

[3]: Rough Order of Magnitude. -25% to +75% as defined by PMBOK, but in this author’s experience different people use the term ROM in different ways. In fact, I’m doing that right now!

Related Resources