Estimating September
Milestones, timeboxes, and the pursuit of good enough
Happy 1st birthday to this little newsletter! My professional world has unfolded in entirely unexpected ways since the first edition. Getting 10 newsletters out the door by my self-imposed end-of-month deadline has been a miracle on the order of loaves-and-fishes. Between work and parenting, most of my existence feels like logistics and plate-spinning, so it’s hard to carve out the uninterrupted blocks of time that I need to write longform. This week, I’ve been laid low with a cold, which has left me even more down-to-the-wire than usual. But somehow I’ve managed, and I’m so glad I’ve found time to write and reflect. It nourishes me in a way that the sturm und drang of project work cannot. Thanks for reading and subscribing — it plays no small part in my ability to prioritize making the time :-).
‘Good Enough’ and the four-day workweek
One of my friends leads an organization that is piloting a four-day workweek. They are doing what I think of as a “proper” four-day workweek1: reducing the number of work hours by 20%, but keeping everyone’s salary the same. When my friend announced this new pilot, her staff didn’t greet her with unbridled joy, but rather skepticism and anxiety. It makes sense, when you think about it — when you feel like you’re going all-out, it’s unfathomable you could ever get your work done in fewer days.
As we puzzled together over the lukewarm reception to a four-day workweek, we wondered whether her team (disproportionately young, disproportionately high achievers) might need some help with discerning when their work could be just ‘good enough,’ so they could save their time and energy to invest in excellence where it really matters. It’s not a natural instinct for so many of us: both schooling and society at large condition us to strive for straight As across the board. But in modern knowledge work, that is a recipe for burn-out and exhaustion, whether your workweek is four days or five.
In the weeks after our discussion, my friend texted me a question:
I’d love to hear if you have any good resources on how to train folks to focus on ‘good enough’ outcomes and how to estimate the time it will take?
Alas, I didn’t know any resources, and google left me wanting. There are plenty of excellent materials from the worlds of project management and agile, of course, but nothing that seemed quite right for this particular case. So, this newsletter is my attempt to answer that question, with a four-day workweek in mind.2
Pareto vs. Parkinson
The Pareto principle, or the 80/20 rule, says that 80% of the outcomes come from 20% of the efforts.
Parkinson’s law says that work expands to fill the time available for its completion.
If you think these adages are basically true (as I do), then the implication is clear: you want to be spending most of your time doing the 20% of the work that delivers 80% of the value (to take advantage of the Pareto principle). But if you leave yourself unchecked, Parkinson’s law will take over, and you’ll find yourself often putting in a lot of effort for little marginal benefit.
To me, this is why a four-day workweek usually doesn’t decrease productivity. A lot of the effort that gets eliminated is pink-box effort, which wasn’t delivering much extra value for all those extra hours. With one fewer day to work with, a greater proportion of time is spent in the orange box, creating value at a much higher rate.
That is not to say you should never spend time in the pink box. In any given organization, the orange box is going to be appropriate for most things. The essence of strategy is making choices about the very few areas where it will really make a difference to work in the pink box, doing the hard, hard work to get from 80% to 100%. But it’s important to make those choices carefully and deliberately.
The reality of modern knowledge work is that important work is going to go undone each week, because the work is endless and time is not. This pink box/orange box framing helps us spend our finite time striving for excellence where it matters most. Then, when we’re in the orange box (which is often), we can focus on the different goal of getting to ‘good enough’ as efficiently as possible.
Timeboxes to translate good enough
Good enough, alas, is a slippery, shape-shifting beast — all the more so when the work is ambiguous, creative, emergent.3 Good enough is not shoddy or mediocre work, but rather work that has made key trade-offs to deliver similar value with less effort. It is difficult to define in advance, and it’s subject to change as the work unfolds, as time passes, as priorities wax and wane. And often good enough for the people doing the work isn’t the same as good enough for the person asking for the work. It’s very easy to ask for something without understanding just how much work is required; it’s likewise easy to work to a level of quality and attention to detail far beyond what others will value or even notice. Even if you make a deliberate practice of discussing what good enough looks like for each and every assignment, a lot will get lost in translation.
Enter timeboxes. A timebox is a predetermined maximum number of hours that you will spend on an activity. You don’t just define what you want done, but how long it should take.
“Please can you do [x] — don’t spend more than four hours on it.”
This works because the timebox functions as a unit of exchange. Where good enough is amorphous, time is concrete — each individual has the same number of hours each year, and nobody gets extra. With a four-day work week, you have about 360 productive hours each quarter,4 so when someone asks you to spend four hours on something, they're saying it's important enough to warrant ~1% of your efforts this quarter. If you end up spending 5 hours, that's probably fine. But if you spend 18, you've just given that activity five times more weight than it needed.
Timeboxes get even more effective when they become the basis for two-way communication:
“Please can you do [x] — don’t spend more than four hours on it.”
“I’ve spent two hours on [x], and I don't think I’m anywhere close to half finished. Can we meet to discuss?”
Perhaps the timebox here was too small, and when the two people meet, they’ll agree that it should be bigger. Perhaps the person has been working in too much detail, and they need to take a different approach. It doesn’t really matter. The important thing is the timebox helped identify a lack of alignment between the asker and the doer. Moreover, the doer was able to flag the disconnect after only a couple of hours of work, reducing churn and wasted effort.
Estimating timeboxes to arrive at good enough
Ad hoc timeboxes work really well to define good enough for smaller tasks. But when it gets to larger efforts, timeboxes lose a lot of their effectiveness. Let’s suppose you have an initiative that’s supposed to take up 50% of your time this quarter. That gives you a timebox (~180 hours5), but it doesn’t really define good enough, because human beings—as a species—systematically overestimate what can be accomplished in a given block of time.
For timeboxes to be effective over larger blocks of work, you need to estimate effort to achieve specific outcomes. The most accurate way to estimate is by breaking work down into smaller and smaller chunks, until the chunks are more concrete and easier to timebox. For example: “Do a literature review on x” is a very open-ended activity; no matter how long you spend on it, you can claim to have reviewed literature. If you are working on your PhD, it’s probably necessary to spend hundreds of hours on this activity. But if it’s just a stepping stone towards an objective you need to deliver this quarter, you’ll need to spend a small fraction of that. Estimating the literature review both defines your timeboxes and articulates some of your assumptions:
The payoff from building these estimates is twofold:
Confidence you have a path to deliver the objectives
Concrete assumptions you can adjust as you refine your definition of good enough.
Number 2 is especially important. The true power of estimating emerges when it is a give-and-take, so that the estimates shape the objectives and vice versa. If 18 hours is too long to spend on the literature review in view of everything else that has to get done, then the estimate helps us navigate the trade-offs (read fewer articles, skip the summary). Conversely, if we think 15 articles isn’t sufficient, we can quickly translate that into an extra time investment.
Now, I want to be clear that this kind of estimating absolutely takes more time and energy than arbitrarily assigning a timebox and calling it a day. It isn’t always worthwhile to spend that extra effort. Sometimes you should just pick an arbitrary timebox and get to work. But when objectives are fuzzy and timeboxes are large, investing in upfront estimating is the best way I know to have everyone aligned on what good looks like, to deliver the objectives on time, and to have everyone (mostly) stay within their regular work hours.
Timeboxing tips and tricks
Before I send you off to estimate and timebox to your heart’s content, I have some practical advice to share:
Be realistic about the size of your timeboxes: Speaking for myself, the stress so often comes from not managing my own expectations about what’s possible in a given window of time. Working eight hours a day is not the same has having eight hours of timeboxes, unless you never send an email, give advice to a colleague, file an expense report, or attend a meeting. I find eight hours of work generally translates into six hours of productive timeboxes6. Using that mental conversion also builds in contingency7, so my timelines have some flexibility when the unexpected happens (as it often does).
Work breadth-first, not depth-first: Working depth-first is writing the first paragraph of the newsletter and futzing with it until I’m quite happy with it, and then moving on to the second (with futzing), then the third (futz), and so on. Working breadth-first is sketching out an outline of the key ideas for the whole newsletter, then going back and adding bullets under each idea, then crafting paragraphs from each set of bullets, then futzing the paragraphs. It’s not always easy or intuitive to work breadth-first, but it makes such a difference. If you’ve worked depth-first and your timebox is up, you have half-finished work. If you’ve worked breadth-first and your timebox is up, you have a complete thing, just at a lower level of fidelity8. Working breadth-first also makes it much easier for you to…
Share early and often: Timeboxes are not the amount of time you spend working on something before you show it to others and get feedback. Especially for ambiguous work and longer timeboxes, you should be checking in frequently so that you can adjust the direction and approach to make best use of the remaining hours in the timebox. Try to put pencil to paper (pixels to screen) in your first two hours, and get comfortable letting others react to your shitty first drafts. As the work takes shape and ambiguities resolve, the check-ins can be less frequent, while you put your head down and crank through the work. On the flipside, if you’re the one giving the feedback, you should be tailoring your comments so they can be reasonably executed in the remaining time. Reworking the whole approach is great when you’re 10% of the way through a timebox. It’s a disaster when there’s only 10% left.
It gets easier with practice: Thinking in timeboxes and estimates is not the most natural instinct for most people. It can feel pedantic and painstaking. It definitely gets easier and quicker over time. Your estimates don’t need to be as detailed because you can rely on past experience. You learn the activities you always, always underestimate. You’re just faster at it. But the thing many people are surprised to find is that once you get going, it is a surprisingly creative task. You’re not just breaking down tasks into sub-tasks: you’re imagining how the next few weeks or months of your life will unfold, and designing how your goals will look and feel, once you achieve them.
What I’m Working On
All year, I’ve been working on a rather big thing that finally launched this month. As is often the case, my role was to connect dots, spin plates, and offer opinions on em-dashes, while others did the hands-on work of making. It was a pretty impossible timeline, so there were plenty of good enough trade-offs made along the way. But in the end, we delivered not just on-time but a week ahead of schedule, and I’m filled with gratitude to have worked with such a talented collective of individuals. It’s not often you get to collaborate with people who commit to shared goals with tenacity, pragmatism, and a healthy dose of snark. I’m so proud of them and their wonderful work, and I feel incredibly fortunate to have worked with such a wonderful team. And since they mostly subscribe to this newsletter, I thought I’d make a point of saying so.❤️
Not that we will have much of a chance to catch our breath. The thing we worked so hard on for nine months has now become phase 1, and it’s time to roll up our sleeves on phase 2. No rest for the wicked, but I suppose that’s what the timeboxes are for.
In comradeship,
S.
Often, you’ll read about a company doing four-day workweeks, and discover that they are asking employees to compress the same number of working hours into four extra-long days. That doesn’t strike me as a change for the better. Moreover, ongoing pilots suggest that most organizations who move to a 32-hour workweek see their productivity remain constant (46%) or improve (49%).
Did you know you can send me questions and I might write whole newsletters about them? It’s true.
Which is to say, most of the work humans do, as the robots and AIs take over the predictable and repeatable tasks.
32 hours a week x 45 weeks=1440/4 quarters = 360. This assumes four weeks of vacation, 12 statutory holidays, and 3 days lost for sickness, dentist appointments, etc. If you work a 40 hours a week, you have 450 hours a quarter.
Actually, probably less, but I’ll get to that.
For individual contributors; it’s usually less than that for managers.
The trick is to have just a bit of contingency. Too little, and you feel like you’re in a constant state of scramble. Too much, and we’re back to Parkinson’s Law, with the work expanding because there was too much time available in the schedule.
For instance, I think the transitions in this newsletter need some work, but it’s after 6pm on the last day of the month, so I’m gonna send it anyway.
This is gold! Thank you for the thoughtful and practical approach to this challenge. I’ll continue to question aloud around you!