Explore a better way to work – one that promises more calm, clarity, and creativity.

Deep Habits: Using Milestones to Get Unstuck


In Search of Productive Simplicity

Last week, I described a kink in my project productivity systems. I was oscillating, somewhat haphazardly, between two different strategies, tracking hours (e.g., when the work is open-ended), and pursuing milestones (when the work is known and I need to hustle).

This felt too complicated, so I asked for your thoughts and you responded with over thirty suggestions.

A lot of your advice seemed to fall into the category of “different work requires different tools, switch as needed.”

This is probably good counsel.

But it still nagged at my preference for simplicity in such matters (which, as a theoretical computer scientist, I of course measure in terms of Kolmogorov complexity).

Then I got an e-mail from an academic in a field that also requires proof-style work (e.g., problems for which the process to completion is unknown in advance). He explained how he approaches projects:

Milestones for me. In [my field] I hit roadblocks and dead ends which force me to start over or take off in a different direction.

After asking him some clarifying questions, I filled in the details. When working on a complex problem, he would set a key milestone and attack it. If he failed to reach the milestone, he would then ask “why?”, and react accordingly:

  • If his answer is “I didn’t spend much time on it,” then he knows the next step is to put aside more time in the near future to work deeply.
  • If his answer is, instead, “I did spend enough time on it, but I didn’t get anywhere,” then he’s forced to either: (a) identify a brand new approach and tackle the same milestone again with this new approach; or (b) move on to an unrelated milestone.

To summarize, this approach focuses on milestones of the form “accomplish X by Y” (not hour tracking), but when a milestone is missed, it generates an immediate postmortem to figure out what comes next.

Once I understood these details, the full simplicity and power of this approach resonated with me. This academic was reacting to an important truth that I was overlooking: the difficulty of hard-to-predict processes — like solving proofs — is not just making sure you put in the time, but also making sure you don’t waste time stuck in a cul-de-sac (to borrow a Seth Godin term).

Stated more concretely: it’s likely way more productive to spend five hours each on three different approaches to a problem then to spend fifteen hours stuck on one approach.

This milestone-centric strategy is inspired by this reality. There are two reasons why it appeals to me:

First, I love the simplicity of using a single tool to unify how I approach the various important projects in my life.

Second, it resonates with my experience: if I fail to prove something with a given approach after, say, 5 – 6 hours, I’m likely never going to succeed with this approach. I either need to try something new, read something new, talk to someone, or move on to a new problem, keeping this one in the back of my mind in case I later come across a relevant new tool (a common trick of Richard Feynman).

I need to try this approach more extensively before declaring it successful. And it likely does not apply to many fields outside my own. But I thought it might be interesting to offer some insight into the meta-thinking that supports a lot of what I do.

More importantly, it’s a good excuse to bring up Kolmogorov.

18 thoughts on “Deep Habits: Using Milestones to Get Unstuck”

  1. I think this approach applies to more fields than you might realize. I am an artist and am also currently working on a novel. In both endeavors, this is how I work, and it is really the only way to work. With painting, I may try one sort of approach to capture what I want with paint, and if it doesn’t work after a while, I scrap it and try another approach. Same thing with writing: if one story line leads to either potential boredom or annoying complexity for a potential reader, it’s time to back up and try some other path out of the woods. Time spent is pretty much irrelevant.

    BTW, before my venture into the creative arts, I spent time writing mathematical proofs, then software development. There’s not much difference in how I approached these two seemingly different careers.

  2. This sounds like simple goal setting, which works well for a lot of people. However, the type of problem makes a lot of difference. For many people setting a goal with a date is a set up for discouragement if they do not reach the goal rather than information about time or process. If your goal has personal meaning, which it may, and/or if reaching it or not reaching it seems more about your capacity and less about the problem, then setting up a concrete goal can feel like a test of yourself.

  3. To use a directed graph search analogy, milestones are nodes and hard-work hours allow you to move via weighted edges. When you reach a dead end no amount of hard-work hours will get you closer to the destination node. Only when you identify nodes you have not yet traveled (new milestones) to can you wisely spend your capital to move again towards your goal.

  4. Don’t know if you addressed it in your prior posts, but what is your view on using the Pomodoro technique to manage time?

  5. As a professional musician and educator, it’s fascinating to me how often strategies in your posts mirror practicing strategies of musicians. In the fabulous “Practicing for Artistic Success,” Burton Kaplan describes “The Rule of Impatience” which posits that if you repeat a passage a certain number of times in a row without success that you should “stop playing! Give up!” and then other strategies such as “invent a new strategy”, “go for a walk and come back refreshed to practice further,” “call friends or colleagues and ask them about their strategies”. Sound familiar? 🙂

  6. The approach with milestones could be effective with constant procrastinators. For procrastinators, time is often the excuse used for not getting things done with things like family commitments or work hours being used for not being able to complete projects. Instead of being controlled by time factors, they can use milestones to reach goals.

  7. Cal,

    Which approach (time vs. milestones) would you recommend for studying / learning a new skill from scratch (for me – Objective-C programming to build new iPhone apps)?

  8. Procrastination certainly can be a problem, but for me it works the other way around.

    When I was studying for my finals in theoretical physics, I never knew how long it would take me to understand an important derivation. Sometimes, there would be some mathematical steps that would take me forever to figure out. And other times, it would go just swimmingly.

    When I set a milestone, i.e. “Do this and that by that deadline”, this was exactly what would make me procrastinate, because deep inside I knew there was no way of knowing whether I could keep that promise to myself.

    When I committed myself to work intensely and focused on a certain issue for a limited period of time, without committing to actually reaching the milestone during that timespan, it would be much easier for me to do it.

    I would agree that it is not a good idea to waste more hours when you sort of get stuck, spending days over a single problem, when you could use the time to produce other results. I did that, and it was not very effective. So I told myself that if I wouldn’t get ahead within 90 minutes, I would call it a day and work on something else.

    Often, the other day, with a fresh and rested mind, I would get ahead easily.

    I’m not sure what you meant by tracking milestones though. In my understanding, a milestone is a goal within a deadline. If something takes much longer than anticipated, as is the nature of this kind of work, this could very easily mess with other deadlines during the same time period.

    • As I understand, a milestone’s deadline is not so strict. Missing a deadline is supposed to cue reflection on your approach. You want to refine your approach to maximize your milestone achievements. Deadlines and milestones should be ambitious enough so that there is some probability of failure each time. In this context, “failure” is not something terrible to be avoided, but as a tool to motivate self-improvement.

      Time spent is merely a detail of your general approach. An approach defines what, when, where, and how you will achieve your milestone. It can be as vague or as precise as necessary. It can include how much sleep you get, when and what you eat, where you study, how you study, for how long you study, when to ask for help, who to ask for help, whether to listen to music while studying, whether to drink coffee while studying, etc. Time spent is only one dimension of an approach; there are infinite dimensions to play with.

  9. This reminds me of an algorithm for a search of the best move for chess players. Ideally, one would identify all the reasonable candidate moves, then calculate each, and it is very important to set the evaluation at the end of each line. If you don’t do so, after certain point it would be unmanageable to track all the conclusions from all the lines. Then compare all the calculated lines and based on comparison make the best move.

    What normally happens during the game though, one jumps into calculating one line, without considering all the candidate moves and spends time and effort, only realizing afterwards that this not the right move. Instead,spreading little time to identifying different possibilities and only then immersing yourself in calculation is a better algorithm.

    Give an amateur chess player a chess problem that requires grandmaster’s understanding of chess then even with infinite number of hours an amateur player will not be able to solve it. Even if he uses different approaches. Instead, I think one should use this time to improve a general game understanding and only then go back to the problem.

    This applies to my postdoctorate research as well. If I am stuck at a problem, then I will set few months to polish the skills and only then approach the problem again.

  10. This is great. I love how adaptable your concepts and ideas are– your methods can be applied to creative work regardless of the specific field.
    This is exactly what I needed to see. Thanks for the post!

  11. Setting certain milestones with a certain amount of time makes sense, but what about the people who take much longer to do something? The increase in the amount of time taken to complete a milestone or move on would then be much longer and so less efficient. If a person who naturally takes long to complete something changes approach after a certain amount of time, and then again if the next approach doesn’t work, wouldn’t having kept at the first approach for longer, and then it proves successful, be more efficient. At what point do you know that your approach will not work?

  12. I like this – it reminds or regiments you to occasionally not just “take a break” or “take a step back”, but to take a “meta-perspective” and ensure that your present path is actually going to take you where you want to be. While spending the last few hours on a given path, have you learned anything useful that tells you whether or not this is the best path to be on? Is your present method or position in the problem-solving process actually a useful place to be? Have you gotten thoroughly lost in the weeds (I’m very guilty of this one – I can spend a lot of time learning the intricate details of a given method while being quite sure that it’s going to lead to a dead end.)? Is the method you are trying out now actually going to get you to a solution? Are you better off just picking a method and running with it, or do you need a lay of the land before choosing something? Are you basing your present path on an assumption that is not true?

    It reminds me of the scene in The Hobbit where the company is in Mirkwood Forest and dwarves send Bilbo up a tree to see where they are and make sure they are going the right way. We’ll just ignore that they get captured by spiders shortly afterward….

  13. Cal – I scanned the notes above and hope I am not replicating anyone else’s comments. If so I apologize… Your approach works because it is essentially the scientific method. In process improvement we call it PDCA (plan do check act). The milestone target you set is your hypothesis – your best guess at what will happen as you pursue your goal. You do the work (run the experiment). When you come to your milestone, you check your hypothesis against your activity in two ways, to a) compare planned results to actual results, and b) compare planned method to actual method of running the experiment. Then you adjust accordingly based on what you learned in the process. Toyota attempts to build these learning loops into every process it puts in place, and it is one important key to their astounding performance,


Leave a Comment