Traditional kickoffs assume predictability, when in reality, large projects are full of complexity. CAPED embraces that complexity—so your project doesn’t fall into the usual estimation traps.
Traditional project kickoffs assume predictability, but large, complex projects—especially in regulated industries—don’t work that way. In this episode, we take a deep dive into Phase 1 of our CAPED approach: Strategic Planning. Learn how to align stakeholders, map complexity, and use Reference Class Forecasting instead of unreliable decomposition-based methods to improve estimates and decision-making.
Episode transcription
Peter Green
In Episode 169, we introduced CAPED, our approach to handling large projects in regulated industries like BioTech, aerospace, telecom, and FinTech. Today, we’re diving deeper into Phase 1 of CAPED: Strategic Planning—and specifically, how Phase 1 sets a project up for success in ways that traditional kickoffs often fail to do.
A typical project kickoff aligns people on goals, reviews documentation, and lays out a plan. That’s fine, but here’s the problem: traditional kickoffs assume predictability, when in reality, large projects are full of complexity. In this episode, we’ll show you how CAPED embraces that complexity—so your project doesn’t fall into the usual estimation traps.
Before we dive in, if you haven’t watched our CAPED overview, check that out first. And visit HumanizingWork.com/CAPED to learn more.
Also, a quick reminder: This show is a free resource sponsored by Humanizing Work, where we help organizations improve leadership, product management, and collaboration. If you’re watching on YouTube, subscribe, like, and drop us a comment. If you’re listening to the podcast, a five-star review helps others discover the show. Thanks for supporting us!
Let’s start with what Phase 1, Strategic Planning, has in common with a traditional project kickoff. There are a few basics—getting everyone aligned on the goals, reviewing any relevant customer or technology constraints, and talking through early solution ideas.
But here’s where CAPED takes a very different approach. Instead of assuming we can analyze our way to a fully predictable plan, we acknowledge that there’s complexity involved—and that means our first step has to be mapping that complexity.
In addition to that complexity map, the other big deliverable in Phase 1 is an estimate—because leadership needs to make decisions about budget, timeline, and ROI. The way most teams try to generate estimates, though, is fundamentally flawed. They start by breaking the work into small pieces, estimating each piece separately, then adding everything up. And, if you’ve been involved in managing big projects, you’ve probably done this. I know I have.
At first glance, that seems pretty logical. The problem is, it’s based on three flawed assumptions.
First, that approach assumes we already know everything that will need to be done. But large projects contain unpredictable elements, and the reality is, we don’t know all the details yet.
Second, even if we had all the right pieces identified, we tend to underestimate how long each one will take. This is something called optimism bias, which causes us to default to the best-case scenario when we estimate. We assume things will go smoothly, that there won’t be unexpected blockers, that everything will fit together neatly. But anyone who’s worked on a big project knows that’s almost never how it actually plays out.
And third, when we see a detailed plan, our brains tend to believe it’s correct, simply because of how much detail is there. It creates an illusion of certainty that isn’t earned. So now we have a plan that’s based on bad assumptions, underestimates how long things will take, and feels incredibly reliable even though it isn’t. That’s a dangerous combination.
So, what’s the alternative? Instead of relying on flawed decomposition-based estimates, we use an approach called Reference Class Forecasting, which was developed by Bent Flyvbjerg.
This method doesn’t try to predict everything from scratch. Instead, it invites us to look at past similar projects and ask, “How long did those actually take?” By using real-world data instead of wishful thinking, we get a far more accurate estimate, and once we’ve built up a good dataset, this process is also much faster than trying to estimate from the ground up.
The other key piece of CAPED Phase 1 is trying to figure out where the big risks are. And, to do that, we systematically identify the assumptions we’re making—especially the ones that could cause real problems if they turn out to be wrong.
When we’re working with clients, we use six categories to brainstorm this. We look at our assumptions about the customer’s outcomes, about customer pains, and frustration, and their willingness to pay for a solution. So, three customer-based assumptions. We also examine assumptions about whether our solution will actually deliver on those customer outcomes and reduce their pains, our ability to deliver, technically and organizationally, and whether we can do that profitably. So, three, really, that are about us on our end; our solution, our ability to deliver, and whether we can build a business case around it.
From there, we sort those assumptions by how important they are—meaning, if we’re wrong about that assumption, how big of a problem does that create for the success of our initiative? That’s on the Y axis. Up and down. Then, on the X axis, left to right, how much evidence do we have to support that assumption, from a lot of evidence to no evidence. The ones that end up in the upper right corner, the ones that are high-risk and low-evidence, become our focus.
Next, we bring in the Cynefin framework to help us determine the best way to gain that evidence. Some assumptions can be figured out through analysis—they’re complicated but knowable. Others are complex, meaning the only way to validate the assumption is through experimentation and learning. This is especially true for anything related to customer assumptions and customer behavior.
To wrap up, CAPED Phase 1, Strategic Planning, does two big things. First, it replaces unreliable, decomposition-based estimates with Reference Class Forecasting, which is backed by real-world data. Second, it systematically identifies and prioritizes the assumptions we need to test, so we don’t discover that our plan is untenable late in the project.
So, if you’re kicking off a big project soon, ask yourself: Are we treating this as predictable? Or are we accounting for complexity? That one question could save you from a lot of costly mistakes.
And, if you found this helpful, let us know in the comments. And if your organization is working on a large, complex project and could use expert guidance, visit HumanizingWork.com to schedule a conversation.
Thanks for watching! In the next episode, we’ll break down how CAPED Phase 2–Active Planning takes a cue from great animation studios and architects, bringing agile iteration into the planning stage of a big initiative. See you then!
Last updated