How deconstructing my calendar told me a fascinating story
Too many meetings, Too little time
Often individuals in management roles find themselves occupied in too many meetings. I find myself looking at my calendar at times and wonder if I have any free time to do any work. We all at some point have likely joked about meetings being counter-productive but they do serve a purpose and add value — why else would we have them? When done right meetings can get real work done. Broadly speaking they help provide support, exchange information and make decisions.
I recently read High Output Management by Andy Grove and loved the chapter Meetings — The Medium of Managerial Work. This topic is a personal favorite of mine for 2 reasons. Firstly, Time is a limited and precious resource and if I can manage to optimize my meetings to derive higher output that will be awesome. Secondly, Time is an irreversible resource. A decision or action that wasn’t taken yesterday can’t ever be reversed. Once you lose time you are always going to be late. In case of many other resources like money or man power the constraints and budgets can be changed. So any improvement in time management has a multiplier effect on my productivity.
In that chapter Andy classifies his managerial meetings at Intel into two types :
- Process-Oriented Meetings : Ones where knowledge is shared, information is exchanged, and they take place on a regularly scheduled basis. These meetings encompass One-on-Ones, Staff meetings, Operations review.
- Mission-Oriented Meetings : Ones where specific problems are solved and decisions are made. They occur on an ad-hoc basis.
He talks about the purpose, best practices and cadence for each meeting type. At the end of the chapter he concludes —
“the real sign of malorganization is when people spend more than 25 percent of their time in ad hoc mission-oriented meetings.”
As soon as I finished reading the chapter the first obvious thought that came to my mind was “How does my meetings portfolio look like? Am I doing it right?” I didn’t know the factual answer to this question. I had a feeling that my some of my days are more heavy on One-on-Ones and some on Staff meetings but I wasn’t sure how much time I was spending overall. I also didn’t know if the time spent was optimal. There were a lot of unknowns and I had a strong feeling that I could definitely do better. It didn’t take too long for the engineer in me to start prompting — “Let’s data science the shit outta this …” and thus began a weekend project to deconstruct my calendar.
And then science met management
Before I began analyzing my calendar I listed the goals of the exercise. I wanted to validate a few hypotheses based on anecdotal evidence of how my calendar looked and felt. I have a few ground rules that I have set to optimize meetings for specific purposes. At the end of the analysis I wanted to see if I actually abide by those rules. Lastly, I wanted to observe new trends and gains insights to improve the composition of my meetings and the respective schedule going forward.
Assumptions and Hypotheses
- There are tons of resources which talk about optimizing maker’s time. Paul Graham has written a great essay about Maker’s Schedule, Manager’s schedule which every manager should read. To ensure all the makers in my team don’t have a fragmented day we’ve set a ground rule of no meetings before 2pm. I definitely wanted to verify if my One-on-One meetings are indeed happening after 2pm.
- I’ve also set a rule for myself to maximize scheduling of all One-on-Ones to Mondays and Fridays so that my peers, direct reports and skip-levels aren’t interrupted in the middle of the week. My hypothesis was that most of my One-on-Ones are on Monday and Friday afternoons.
- Working for Spotify in NYC with engineering offices and distributed leadership in Sweden (CET) and US East Coast (EDT) requires most Staff or Operations Review meetings to happen during morning hours on my end. We also tend to optimize our knowledge sharing and alignment meetings to happen on Thursdays. My hypothesis was that a lot of our Staff and Operations review meetings are on Thursdays and usually in the mornings.
- My final hypothesis was that Problem Solving and Decision Making meetings (under which I include things like recruitment interviews, system escalations, strategy offsites, etc) have no recurring pattern and they will be all over the calendar.
As any data scientist will tell that any analysis is only as good as the data set. So the first step was to clean and organize my raw calendar data into analyzable data set. Some meetings tend to be seasonal like annual reviews and some tend to have a good recurring pattern — therefore I wanted to even out any such meeting spike. Ideally analyzing all my meetings from the past 12 months would have been perfect. However since the generation of the data set from raw calendar data required manual annotation, I compromised the data set size and settled for performing analysis only for my meetings from the first 6 months of the year i.e. first 26 weeks. In a way I was doing my 2016 mid-year meeting performance self-evaluation.
Since my goal was to gain actionable insights I deviated a bit from Andy’s meeting classification to suit my needs and categorized and hand annotated all my meetings as follows:
- One-on-Ones : All One-on-One meetings with direct reports, my manager, my peers, skip level reports. The reason to distinguish One-on-One from process-oriented meetings was to specifically measure it’s own performances.
- Staff meetings and Operations Review: These would be recurring meetings with senior tech management, localized tech management and my own direct reports. This also included certain other cross-functional information exchange and discussions meetings with leaders in product, design and operations.
- Decision making and Problem Solving: These are mostly ad-hoc meetings where we’ll actively be solving problems and making decisions. Things like discussions with vendors, partners, recruitment candidates, problem escalation, ad-hoc initiative kick-off meetings (before they start operating at a regular cadence), etc. were classified under this category.
After the tedious work of hand-annotating all my meetings; I started by plotting the graph of the time spent in meetings per week grouped by the category. The plot can be seen below:
The basic analysis revealed that on an average I had been spending 16.5 hours every week in meetings. This is about 42% of my time. Of those 16.5 hours about 4.5 hrs on average every week were spent in One-on-Ones. Which is about 11% of my overall time and about 27% of my meeting time. I was spending the remaining time of about average 7hrs and 5hrs on Staff/Operations review and Decision/Problem Solving meetings respectively. Although 1 out of every 4 hours being spent on One-on-Ones didn’t seem bad it did feel to me like I might have been actually spending more time than that.
I felt I needed to analyze the volume of meetings every week. To do this I plotted a histogram of number of meetings by week grouped by category. The next plot revealed another interesting picture.
This histogram revealed that on average I have about 18 meeting every week of which about 8 are One-on-Ones. So although the time I spent on One-on-One is about a quarter of all my meeting time, in terms of volume the One-on-Ones are about half of my weekly meetings. This answers why I felt over-indexed on One-on-Ones. The reason for a difference in volume and time spent is due to the fact that I tend to have more frequent 30 minute One-on-Ones where as most Staff meetings tend to be 1hr long. The average number of meetings was split evenly at 5 between Staff/Operations review and Decision/Problem Solving.
This initial analysis gave me a good overview of my meeting portfolio and the time spent. However, I still had no idea about my meeting distribution and how fragmented (In other words : disorganized) my days looked. The next step was to create a heat-map of my meeting time distribution over a week for each category. The following diagrams show how the individual distribution and combined distribution looks like.
Results and Insights
Looking at the volume of meetings, meeting distribution and the overall time spent in meetings has revealed lots of interesting insights.
- I confirmed that I’m not mal-organized since I was spending less than 25% of my time in ad-hoc meetings. I was spending exactly half of that (12.5%) which brought a sigh of relief.
- I clearly wasn’t having my One-on-Ones only on Monday and Friday afternoons. In fact Wednesdays were my most busy One-on-One days — what a bummer. However, I was good at keeping these meetings past 2pm. I’m glad I discovered this else I’d have continued to schedule future One-on-Ones re-actively based on my availability as opposed to spreading the load of these meetings over to Mondays and Fridays.
- Thursdays and most mornings were indeed most busy times for me to have Staff/Operations review meetings. In this case my feelings matched the reality.
- Decision making meetings were evenly spread out all over the Calendar except 2 slots which happened to be recurring meetings where recruitment and tech operations were reviewed. Technically we were solving problems in those meetings but they should have been ideally classified as Operations review meetings. So all good here. It’s also great that decision making was happening when needed as opposed to being reserved or postponed to certain times/days.
- My calendar seemed to revolve around my Thursday heavy Staff/Operations review and self-imposed One-on-One rule. This was good since it brought predictability to my schedule and I could plan things more efficiently.
- The red blocks on Wednesday and Thursday reflect my staff meeting with upper management. The red block on Friday reflects my staff meeting with my direct reports. The order of these meetings does help me to pass on crucial top-down information in a time-sensitive manner. It turns out that this fact though not very obvious when I scheduled the Friday meeting became more obvious after this analysis. Calendar hacking FTW!
Some of the lessons from these insights are specific to me and will definitely help me continue my good habits and get rid of the bad ones. However, some of the lessons could be generally applicable to anyone in a management role. Below are some of my takeaways -
- Plan your week around Staff and Operation meetings. Since these meetings involve many people design and schedule them predictably so that all the participants can set aside some time for those without impacting their schedule.
- Minimize disruption to maker time. Set firm rules about time and days for recurring One-on-Ones to cause minimal disruption for makers. Afternoons are better since mornings tend to be more productive for many. Also people tend to come in to work at different times so afternoons works out better for everyone. Avoid middle of the week as much as possible.
- Schedule the order of meetings for information flow. Schedule the order of Staff meetings to facilitate timely flow of top-down information within the organization.
- Have a calendar strategy. Don’t be reactive and schedule meetings on the fly. Have a strategy for how you want to spend time and optimize your calendar for that.
- Analyze your calendar and validate. Once you have a calendar strategy make sure you analyze your calendar and get a sense for how you are spending time. Facts beat feelings — validate if those two actually match. Do this specially if you are a manager and spend tons of time in meetings.
Hope when you get a chance you’ll work on deconstructing your calendar and share additional tid-bits about your calendar strategy with me in the comments ;)