Nothing compares to the stress of hearing the whooshing sound of a deadline passing by. And if you’re tired of making close calls on your projects, now would be the right time to learn why it’s such a big deal and how you can improve your time estimation skills.
But before we start…
Who Needs Deadlines Anyway?
For one, deadlines help add a little spice to your otherwise boring everyday life in the office. The word itself has a certain thrill of adventure to it, doesn’t it? But while it’s undeniably cool to make a close call on a deadline and feel like Neo dodging a bullet in the nick of time, you really don’t want to push your luck next time. Because whereas nailing the deadline is taken for granted, failing it could mean some serious trouble.
Seriously though, the ‘dead’ in ‘deadline’ just about sums it up, making it really hard to justify its existence. But any deadline is just a product of time estimation. So we’d have more luck asking why would one need to estimate time required for a certain job or project to be done. Now, that question I find a lot easier to answer – clients, among other things, need it to analyze how much money they will have to invest and vendors risk ending up with an unrealistic budget and schedule for their project unless they do the estimates. Freelancers have it even worse – making a wrong guess on how much time they need to complete a certain task can ruin their reputation. And to top that off, without proper time estimates, there’s no way for either of them to tell how much the work is going to cost when all is done. And we know only dentists can get away with that.
The Art of Estimation
Multiple surveys show that even trained estimation professionals known as project managers give the process of time prediction a whole 4.17 out of 5 on the scale of importance, while 9 out of every 10 managers admit it to be ‘problematic’. And I wager that ‘problematic’ is putting it mildly.
So you see, you’re not alone in this. Although people have been doing it for almost a century now, accurate time prediction still remains a problem unsolved. Multiple researches show that over 70% of all projects end up over their schedule or budget, although the average overrun is somewhat minor – 30-40%. How is that even close to ‘minor’? Well, taught by years of experience, most managers and clients find a -20%/+20% deviation in estimated vs. actual time acceptable if not perfect. The funny thing is – those numbers have been very consistent, with no signs of improvement over the years.
So what’s causing the time estimates and hence the deadlines to be so inaccurate? Well, you’re gonna like that, because most managers like to shift the blame to things they can’t be held responsible for, like minor and major changes in their clients’ demands (more than 1/3 of the problem, if you take their word for it). So to put it in one short summary: estimates are very important, yet they just won’t work and it’s not your fault. You could stop right here, but I promise you it will be small comfort if you try to communicate that summary as an excuse for failing all deadlines to your boss, your team or your client.
There are dozens of estimation methods practiced worldwide, but essentially they’re just one of two possible approaches: expert estimation and formal estimation. First, let’s see what either of them is about and then try to tell which one is better.
The expert approach remains the most preferred one. In it, all time estimates are predicted by human experts based on their experience, skills and intuition. It’s often called a ‘guesstimate’ – and I won’t lie, there really is a lot of guesswork behind it, but it does have its advantages.
The key advantage of the expert approach is its flexibility. A good expert has enough experience and information on their team members’ individual productivity to anticipate possible problems and make necessary adjustments on the go. People also tend to show more commitment to human-made estimations than to those generated by a formal model. Think of it as of a ‘self-fulfilled prophecy’.
The downside of this approach is the annoying trait of human nature to make mistakes. It might sound weird, but the worst problem of expert estimation is too much optimism. Optimism causes 44% of the overrun. The other main disadvantage of this approach is that humans are bad at processing big data. Overruns tend to increase with project size (>200 man months) and complexity. Even experts admit that they often overlook tasks and details for big projects.
Originally created to replace the presumably bias expert-oriented techniques, formal estimation is based on pre-calculated models expressed in consistent formulas. This is pure math and statistics, so now you know why it’s the less popular one.
Think of formal models as of math formulas executed by a computer. It doesn’t make mistakes (humans who write the formulas do) and it doesn’t care how much data it needs to process. And it’s a real downer when it comes to optimism too. That’s why formal models are great at processing large and messy chunks of data and eliminating all biases and wishful thinking from the estimate.
And yet, formal models are very rigid, whereas most projects are not. They’re based on the assumption that the same input always leads to the same result. That makes the formal approach very ineffective in unstable projects with a lot of context and turbulency going on. Also, creating an accurate estimation model and calibrating it to the needs of your company can be very complicated and it requires a lot of formal validation data.
Showdown: Man vs Machine
Now you must be wondering which approach is better. Statistically, each of them can be similarly accurate or unreliable. The trick is – how and when to choose either. So we urge you not to look for ‘the best estimation technique ever’, but think of it more in the sense of – which approach is best in certain context rather than in general:
Expert-based techniques are better suited for smaller but unstable projects and changing context. Formal models work better with more complex, but stable data with little context dependency.
Expert predictions are flexible but inconsistent, models are consistent, but hard to re-calibrate.
Experts have the natural advantage of possessing more information, but tend to be over-optimistic. Models are exceptionally good at eliminating situational biases that lead to such optimism.
That said, the key to making an accurate estimate is understanding which approach works best for which kind of project and then combine them in your work. According to multiple researches, properly mixed estimation techniques produce accurate results more frequently.
The choice is yours. Time estimation skills need time to develop(no pun intended), and if you’re new to it, the only way to find out which techniques work best for you is by running a few experiments. What’s the worst that could happen, right? Just be sure to let us know how that goes in the comments below and check back later – as we’ll have more articles covering this topic real soon.