This is a thread to organize the discussion around the productivity tracking breakout session.
Session Description
While productivity tracking is a popular pursuit, there are a fascinating variety of approaches. Many seem conflicting at their core, urging us to step back and ask:
[list]
[]What is productivity in the first place?
[]What metrics represent it best?
[*]How can we use these metrics to track what we actually care about?
[/list]
I think a lot of the conversation is predicated on the idea that you have to have a single metric for tracking your productivity.
When I think of ‘productivity’, I think of it as not any single activity I might do in a day, but rather, a sort of overall trait which is high or low and influences all the activities I do each day, and which has no single metric I can quantify it as. Or in other words, I consider productivity a latent variable(https://en.wikipedia.org/wiki/Latent_variable), which we might extract from a bunch of regular measurements using factor analysis(https://en.wikipedia.org/wiki/Factor_analysis).
You could take a bunch of measures (maybe something like: # of emails sent, self-rating of productivity, words written, todo items checked off etc etc) and try to extract a factor which looks like ‘productivity’. Has anyone done that?
Should also consider the purpose of the tracking:
If the purpose is to discourage goofing off, basic metrics about application and web site usage should be sufficient.
If the purpose is to determine the effect of certain variables (e.g. sleep or coffee), you’d need task-specific measures. This is easier for some tasks (e.g. picking apples) than for others (e.g. writing software). Artificial tasks (e.g. a game that requires concentration) might be preferable.
Gwern: great to see you here. I’ve been a huge fan of your blog for a long time!
I agree that productivity is a latent variable, although I’m not aware of anyone doing what you propose.
One problem I see with it is that it might end up measuring busyness rather than productivity. For example, you might finish a very important article and do nothing else that day. While that may be a great and productive day, your metric would probably not reflect it. I think it would be almost impossible to construct some metric beforehand that accurately reflects true productivity.
(I guess ‘feeling of productivity’ might be flexible enough to account for a wide range of work, but that’s too vague again.)
In my experience when I’m measuring things that poorly correlate to the thing I’m really interested in, I abandon those projects quickly…
One thing I like is setting the goals the day before and then tracking completion. One could then also weigh the goals. For example, finishing the article might have a weight of 80% whereas getting to Inbox Zero has 10%.
ejain: Totally agree with you. In my experience, tracking and restricting time-wasting activities is extremely powerful. For example, right now I’m checking email once a day for at most 1 hour between 4pm and 9pm. Of course, even though it is completely obvious to me that it dramatically improves my productivity it is a very poor measurement of productivity so one thing it doesn’t allow me to do is to compare changes over time.
I think the best measure of productivity is results. Results that are relevant for you. If you are say in commission based sales then it’s how many sales you are generating.
A trap I see with productivity is that people often measure output. Like a productive day is one where they were in flow state, got a lot done and had minimal distractions. But what if all that work created no tangible result?
Let’s say you are planting trees. If one person plants 100 trees in 8 hours, and another plants 50 trees in 8 hours, you’d generally say that the first person was more productive. Even if the second person was in a flow state and had minimal distractions.
If measuring just output and not results, then consider 2 sales people. Both work for 8 hours going door-to-door selling. The first person knocks on 200 doors, and the second knocks on 50. It’s common to think the first person was more productive. But what if the first person got 0 sales and the second got 5 sales?
I see this kind of thing at work a lot. There is an element of self-deception often present with productivity. A productive day often means a lot of activity was completed. But it takes analysis to figure out if that activity was truly productive or not.
Output can sometimes be a good measure for productivity. Say you are a writer it might be good to have a target of how many words you write in a day. The measure of results - whether your book sells or not - might be so far down the track, activity could be the best thing to track. But then again it could be high or low quality writing.
I personally define productivity as the shortest distance between A and B. A being where you are now and B where you want to be. Say it’s a certain financial, fitness, or other goal, the most productive person is the one who gets there the fastest.
This could make the self-tracking difficult, unless you have a specific goal in mind that you can measure progress along the way on.
“Results” may be a better measure, but can be more difficult to quantify when looking at individual days: How do I know if the pages I wrote today will help sell more copies of the book, or how many of the sales result in happy repeat customers (vs cancelled orders)?
I agree with both of you.
Of course, in the end it is about results and ideally you’d quantify that. But often what you do has a very indirect, delayed or unknown effect on the results you desire so it’s difficult to set up a system for quantifying it.
And then what’s left is often quantifying output. (Or even worse: input (e.g. time)).
If you know of a good system to track results in a systematic and quantitative manner, I’d love to know about it!
From the definition of productivity:
“the effectiveness of productive effort, especially in industry, as measured in terms of the rate of output per unit of input.”
I guess it’s about measuring the ratio between the time spent on ‘productive’ tools and how closer/how many of the desired results are we achieving.
If you spend longer on ‘productive’ tasks, you should achieve more (although it might not be a linear relationship), but I’d say what’s important is how much you achieve per unit of time.
Why is this important?
Imagine you measure only the results, you may be achieving what you expect, you’re happy with the results, but then you realise you’ve been working 16hr/day 6day/week. Maybe you can do better 
On the other hand, you may measure where your time is going (using tools like RescueTime) and have a very high % of productive time, even though you only work 6hrs/day 5days/week. Then you realise, wait, but I’m not achieving the results I expected.
Measuring the ratio, or even better, the changes in ratio, tell you how much faster you get to your goals with each time unit, % of time spent trying to get there and the optimal hrs/day to achieve it without going crazy.
Does that make any sense at all?
Just realised I may have just repeated what mattdk said
That’s an interesting observation Diogo.
You’re, of course, right that productivity technically means the ratio of output to input. Even though I had been tracking ‘productivity’ for years and even though I knew the definition of productivity I only really thought about it when I prepared a Show & Tell on it.
I’d argue though that nobody actually cares about productivity as understood technically. People care about results. I asked everyone during the breakout session to define what productivity means for them and no one said something like ratio of output to input. Instead people think of it as achieving results. (Or strangely being in a flow state.)
Yeah its a tricky one.
I often measure my productivity in 2 primary ways:
An initial task is studying all the things you do in a day and then rank them with the above color system.
and 2. The results.
If you are a finance trader you can probably judge your results daily, but not for most jobs. That’s weekly, monthly or quarterly. This is when you can see if you are truly productive, and if you need to redesign the color system above.
Would you say that, to achieve more results, the ratio could still be useful?
We rarely have only one focal point in life (I’m thinking: work, family, pet projects, studies). It is important to be able to achieve the results on each one of those areas with the least amount of time possible.
My point is that only measuring results won’t tell you much about how you’re achieving them or the sacrifices that you’re making. You may be achieving everything at the cost of a lot of energy.
Measuring the ratio of energy (or time sent) and results would give you an idea of the direction of an optimal productivity.
To conclude, measuring results is interchangeable to measuring success. Productivity should be how well you arrived at that same success.
Lets say you want to achieve goal A by tomorrow. A day has gone past and you’re happy, you’ve achieved it!
Which case sounds more productive?
1 - You spent the whole of yesterday (including night) working on your goal, hard work but it paid off
2 - You organised your day differently and spent 6 hours on your goal and still had time to review and relax before a great night of sleep
Both achieved the same result.
Now, I do agree, if you’re going to measure just one thing, results are the most important one in the end of the day, results are what matter, the rest are optimisations.