One thing many project managers learn early on is that deadlines are rarely met and more often than not, it’s because of inaccurate estimates.
Estimation accuracy is influenced by many factors: missing information, limited experience, high complexity, shifting scope, dependencies outside your control. All worth solving for on their own. But even when every one of those is absent, when the task is simple, familiar, and entirely within your control, estimates can still miss the mark. In this short course we’ll focus on that layer specifically: the mental shortcuts involved in estimating, and some techniques for reducing their impact.
I believe that becoming aware of these shortcuts, and of how universal they are, regardless of expertise, culture, or familiarity with the context, makes us a little more compassionate toward the imperfect decisions we all make sometimes.
By the end of this course:
- Given a planning scenario, you will be able to immediately recognize some of the most common mental shortcuts you or your team members may fall into, affecting the accuracy of your estimates.
- Given a planning scenario, you will be able to apply the appropriate techniques to reduce the impact of these mental shortcuts, producing an estimate that is measurably closer to the actual outcome.
Let’s dive in and meet the little tricksters living in our minds!
Before We Begin: A Quick Exercise
Your best friend’s anniversary is next month, and you’ve decided to organize a surprise party at your place. Your friend loves surprises, and this will be your gift to them. As you start planning, you identify a few things you’ll need to take care of:
- Make sure everyone invited shows up (roughly 20 people)
- Get the house ready for the event and some party props for attendees to wear
- Provide enough food and drinks for everyone, including the birthday cake
- Find a way to make sure your friend is available for the party, without spoiling the surprise
You have exactly one month (30 days).
If you were organizing this by yourself, how many days would you give each activity? Fill in below the Initial estimations based on your first instinct. Then, once you’ve finished the course, come back and fill in the Revised section and see how your thinking shifted.
Mental Traps
Before we get into estimation accuracy specifically, let’s look at the types of mental shortcuts we fall into when trying to make rational decisions: heuristics, cognitive biases, and logical fallacies.
Heuristics were first studied by psychologists Daniel Kahneman and Amos Tversky, starting in the 1970s. In simple terms, a heuristic is our tendency to simplify things in order to solve complex problems. When faced with a complex problem, our brain substitutes it with a simpler one, solves that instead, and lets us believe we solved the original. And we’re rarely aware it’s happening.
We live in a world full of stimuli, constantly needing to make quick decisions, so our brains rely on shortcuts built from past experience. These shortcuts help us decide quickly in similar (not identical) contexts, but in most cases, they’re not actually the best decisions we could make.
This kind of decision-making developed over thousands of years of evolution. Our world used to be dangerous, and the speed of our decisions kept us alive. Even though today’s world holds fewer life-threatening dangers, our brains still lean heavily on these shortcuts, mainly because, once learned, they reduce cognitive load.
In project management, the clearest example of heuristic in action is relying solely on lessons learned from past projects, without accounting for the specific context of the current one. Say a client hires us to build a website. Based on initial discussions and our experience building similar sites, we estimate three months. Yet the estimate turns out inaccurate for any number of reasons: the client is more thorough than others we’ve worked with, an unplanned risk materializes, we need to learn a new technology, and so on.
Another mental trap we fall into is cognitive bias. If heuristics are about simplifying reality to act faster, a cognitive bias is “the tendency to think in one way when other options are just as good, if not better” (McRaney, 2013, p. 11). Have you noticed how, once you start thinking about buying something, you suddenly see it everywhere? That’s confirmation bias in action: we simply pay more attention to information that confirms what we already believe or want.
In a project, once implementation is underway, we might become reluctant to change tactics, even when doing so would lead to better results, due to the sunk cost bias that makes us focus too much on the resources already invested.
When our biases get challenged in everyday life, we often turn to logical fallacies — starting with a conclusion already in mind and working backward to justify it. Superstitions, for instance, are born from the post-hoc fallacy: if one event follows another, we assume the first caused the second.
Say a project manager cuts code review from the process right before a tight deadline, just to save time. The launch goes smoothly, no issues in the first few days, so the conclusion feels obvious: “we skipped code review and everything worked out fine, so it was never really necessary”. In reality, the two events were never properly connected: the bugs that review would have caught often surface weeks later, as production issues or maintenance challenges, by which point no one thinks to trace them back to the step that got cut.
Now that we’ve covered the basics, let’s look at three of these mental shortcuts that affect the accuracy of our estimates.
Trap #1: The Planning Fallacy
The number one trap we fall into is the planning fallacy: we tend to underestimate how long a project will take, while overestimating our own ability to execute it. That second part is also known as the overconfidence bias.
You might ask: “isn’t this something past experience should fix?” Instead of answering directly, let me ask you this: how often, when recalling a difficult or unpleasant experience, have you remembered it more kindly than you actually felt while living through it? That’s the brain’s memory-protection mechanism at work. To help us survive, it tends to soften unpleasant memories or reframe them more optimistically. As psychologist David Gilbert puts it: memory is a representation, not a replica. The more we reflect on a past event, the less accurate our memory of it becomes. That’s why the best practice for documenting lessons learned is to do it as things happen, not at the end of the project.
Still, even when we do have accurate knowledge of how long something takes, we often remain overconfident in our ability to get it done. So the planning fallacy can catch us even when we have relevant, well-documented past experience, and it shows up even more clearly with new projects or tasks we’ve never tackled before. We tend to be far too optimistic. And the further away the actual work is from the moment we make the estimate, the stronger this effect becomes.
So what can we do about it?
Click or tap on each point for more details.
Simplify the project or task.
One reason the planning fallacy shows up is that our minds struggle to reason about combined probabilities. The more activities involved, the harder it is to accurately estimate their combined duration. The more complex the task, the more our brain loses track of the odds that something unexpected will happen. Simplifying tasks is one of the most reliable ways to improve estimate accuracy.
Take time for a risk analysis.
And plan how you’d respond to those risks, especially when the schedule is tight. It’s easy to skip this step when we’re short on time or feeling optimistic, but it’s exactly what helps you account for the unforeseen.
Consider all the factors.
When estimating duration, factor in the effort required, resource availability, and how the task depends on, or is depended on by, other tasks.
Back to our exercise: by now you’ve made a rough estimate for each phase. Let’s zoom into the first activity: inviting guests and confirming their attendance. At first glance, it looks simple: just send everyone the same message with the date, time, and location, and ask them to confirm. Shouldn’t take more than an hour, right? But if you break down the task further, you might realise that not all the effort is in your control. Making the guest list, writing the invitation and sending it out to all the guests are quick steps and depend only on you. But obtaining confirmations depends on your friends’ schedules, and might take a few days. Some invitations might even turn into ongoing conversations you’ll need to keep up with until the event.
Use reference class forecasting.
Instead of estimating a task purely from how it “feels,” look at how long similar tasks actually took in the past, yours or someone else’s. This is different from the “lessons learned” heuristic we discussed earlier, which is more about general experience. Here, you’re deliberately using historical data as a corrective anchor: if your last three website redesigns were each scoped at “6 weeks” but all took closer to 10, that gap is information that should be factored into your next estimate.
The use of artificial intelligence can also prove helpful, if the model is provided with accurate past project data.
Try three-point estimation.
For each task, estimate three numbers: an optimistic duration (if everything goes right), a most-likely duration, and a pessimistic one (if things go wrong). A simple version of this – the PERT formula – averages them with extra weight on the most-likely value: (optimistic + 4 × most-likely + pessimistic) / 6. This gives you a single number that already accounts for some uncertainty, and the spread between optimistic and pessimistic doubles as your estimated range.
As you can see, there’s a lot to weigh in when estimating a task’s duration or costs: breaking it into smaller pieces, accounting for effort, resource availability, task dependencies, and even uncertain events that might come up. Doing this reduces the impact of the planning fallacy and gets you closer to an accurate estimate, even when the deadline itself is fixed. In that case, we can always resort to reverse engineering: distributing task durations backward from a fixed timebox.
Trap #2: The Dunning-Kruger Effect
“The Dunning-Kruger effect occurs when a person’s lack of knowledge and skills in a certain area cause them to overestimate their own competence. By contrast, this effect also causes those who excel in a given area to think the task is simple for everyone, and underestimate their relative abilities as well.” – thedecisionlab.com
When we know nothing about a subject, most of us have no trouble admitting it. But as we start learning, our confidence grows too, sometimes to the point where we believe we’re experts. Then, as we dig deeper, we realize the “iceberg” is much bigger than what we saw at the surface, and our confidence dips. Eventually, we know enough to properly judge what we actually know: our confidence keeps growing from there, but this time in step with our real competence and never again reaching that early, inflated peak.
The bias behind this effect is illusory superiority. Unlike overconfidence bias, which makes us too confident in our own abilities, illusory superiority makes us believe our abilities are above average. As the researchers who coined the term put it: “the miscalibration of the incompetent stems from an error about the self, whereas the miscalibration of the highly competent stems from an error about others” (Kruger & Dunning, 1999).
So, depending on our level of expertise in a given area, we might overestimate or underestimate our abilities, and therefore how long a task will take.
Back to our exercise: your confidence in organizing this party by yourself will depend on your knowledge and past experience. If you’ve never done anything like it before, or simply believe you’re not an organizer, you’ll probably ask for help. If you’ve successfully hosted a family dinner or attended a surprise party before, you might feel overconfident and estimate too optimistically: a month can feel like plenty of time, and you might even think you could pull it off in two or three weeks (which, in the end, could prove risky). On the other hand, if you have a bit more experience and are aware of the risks involved, you might swing the other way and over-plan. It’s usually only once you’ve become a genuinely seasoned organizer that you can balance past experience with the needs of the current project and land on a realistic estimate.
Some of us have a natural talent for organizing; others learn it along the way. Through feedback, observation, and practice, we get better at judging our abilities against the average and our confidence grows accordingly. The real danger of this effect shows up when we have very little information about something and fall straight into overconfidence about our ability to handle it.
What can we do about it?
Click or tap on each point for more details
Start with a self-check.
What do you believe you’re genuinely good at, and where do you tend to struggle? Then look around you: why is this task taking my colleague so long? Am I missing something, or is this simply the kind of task I’d normally be good at?
Ask your colleagues for feedback.
This effect can trick us into assuming that whatever comes easily to us also comes easily to everyone else, when in fact, it might be exactly what makes us valuable.
Track your estimates against actuals.
For every task, write down what you estimated and what it actually took. Notice the pattern emerging over time: maybe you consistently underestimate by 30%, or you overestimate simple tasks but underestimate complex ones. This is the standard technique used in forecasting to gradually calibrate a biased estimator, and it works whether you tend to over- or underestimate.
The Dunning-Kruger effect might hit harder when we work alone and have no reference point, which is often the case for solopreneurs and freelancers delivering a project start to finish on their own. Estimating as part of a team helps mitigate this: it gives newcomers a chance to see what they still need to learn, and gives experts a chance to recognize their own talents.
Trap #3: The Anchoring Effect
Do you think a concrete bridge is built in 18 months? How long would you estimate the complete construction of a medium-sized bridge to take?
Be honest: did your estimate landed close to 18 months? That’s the anchoring effect at work: our tendency to rely too heavily on the first piece of information we receive when making a decision.
What does this mean for estimation?
If, during team planning, one person shares an estimate out loud first, the rest of the team tends to get anchored to it, asking fewer clarifying questions and simply refining that initial number instead of estimating independently.
The fix: let each team member estimate individually first, and only then bring the estimates together for discussion.
This is exactly the principle behind the Delphi method (also known as “estimate-talk-estimate”): gather anonymous estimates from each team member, discuss the results as a group, and repeat the process until you reach consensus.
Planning Poker, widely used in software development to estimate the complexity of user stories, is essentially a lightweight, gamified version of the same idea: each team member privately selects a card representing their estimate; only once everyone has chosen does the team discuss openly and work toward a shared number.
Anchoring is not only a team problem, either. It shows up just as often on an individual level, for instance, when a client hands you a deadline first (“I need this in three weeks”) before you’ve had a chance to estimate anything.
Going back to our exercise, before you estimated a single task, you already had one number in your head: the one-month deadline. That number influenced how you split the days across the four activities, even though 30 days has nothing to do with how long it actually takes to send invitations or clean the house for a party. That’s the anchoring effect at work, before you had even started this course.
Even when a deadline arrives from someone else, a client, a stakeholder, or simply the calendar, it’s worth estimating each task on its own merits first, and negotiating the gap, rather than reverse-engineering your numbers to fit it. Chances are the other party is just as unaware of the risks as you were.
Before We Go: A Quick Recap Exercise
Read the six project scenarios below and see if you can spot which of the three mental traps discussed in this course is most likely at play.
Summary
We’ve covered our three most common mental traps: heuristics, which help us act quickly in complex situations; cognitive biases, which push us toward certain default patterns of thinking, even when they’re not the best ones; and logical fallacies, which help us justify beliefs we’ve already settled on. We also looked at how the planning fallacy, the Dunning-Kruger effect, and the anchoring effect specifically distort our estimates, and what we can do about each one.
If you want to find out what other cognitive biases might be at play in other areas related to your projects (e.g. decision-making and risk analysis, social and group influence), visit BiasCodex. The team has mapped 200 mental shortcuts, with definitions, examples and practical actions to counter them.
Now it’s your turn: go back to the exercise at the start of this short lesson, revisit your original estimates, and refine them if needed, considering what you’ve just learned.
References:
- McRaney, D. (2013). You Can Beat Your Brain. Oneworld Publications.
- Kahneman, D. (2013). Thinking, fast and slow (1st pbk. ed.). Farrar, Straus and Giroux.
- Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. Available at: Wayback Machine.
- Bias Codex