Measure what matters

#goals #management #metrics

This week over lunch the topic of OKRs came up. If you are not familiar with OKRs they stand for "Objectives and Key Results", and are used at many companies to align everyone on set goals for a period of time (usually per quarter, but this can vary).

OKRs have been around since the 1970s, but really the bible for learning about them is the book "Measure What Matters" by John Doerr. They can work really well for aligning a company and making sure everyone is working towards the same objectives.

The key to a good set of OKRs are that, Objectives should be significant, concrete, action-orientated and inspirational. Key Results should be measurable, specific, time bound and aggressive, yet achievable.

OKRs are unfortunately not the holy grail when it comes to a company's success, and there is a delicate balance that needs to be walked, in order for them to be effective.

In the book it is suggested that teams should be allowed to set some of their own OKRs, in addition to aligning with the company OKRs that affect them. After all, teams need to feel some autonomy over their own work otherwise job satisfaction will suffer.

Where I have seen OKRs fall apart is when one team's OKRs negatively impacts the roadmap of another team. Big cross organisational projects need to be taking into account and factored into everybody's OKRs from the start.

It is a frustrating experience to spend weeks sorting out OKRs and planning out a quarterly roadmap, only for the whole thing to be get derailed a week later, when a priority feature needs to be developed, and you have to drop everything to get it shipped.

OKRs work well for software development when you have a single engineering team all working together. When companies get larger, the engineering team tends to get broken up into individual specialities each working independently of one another. This set up really helps autonomy but can make cooperation between teams difficult.

Key Results #

One aspect of OKRs that is always important is the Key Results. Every team should have metrics that they measure and try and improve on.

"If you can’t measure it you can’t improve it." - Peter Drucker

I am big fan of data driven management, but you do need to be careful what metrics you pay attention to.

The quote above says, "if you can't measure it you can't improve it". I don't believe that is always the case. There are plenty of things in business and life that you can't really measure, but you can improve, such as relationships, reputation or brand awareness.

Even though the above is sometimes true, it is the reverse that you need to be mindful of. Just because you can measure something doesn't mean that you can improve it.

As a part-time YouTuber I have been guilty of this many times. I can measure how many views I get or how many subscribers I get per month. However, short of making more videos or better videos I can't control it.

In many cases, it can take months for a video to start gaining traction, and it is difficult to correlate that with any changes I have made. Given that my YouTube Ad revenue is directly tied to the number of views I get, I can't control the revenue either.

Always try and make sure that your metrics and key results are things that are controllable. From an engineering perspective these can include:

Test coverage
Performance metrics
Defects found
Incident response time

Ignore the rest #

Beyond those key metrics that we are trying to improve on, what do you do with the rest?

In my opinion, you should just ignore them. If they aren't within your control, and they aren't key to driving your goals forward, then they are just a distraction.

When I first started my YouTube channel, I used to obsess over the number of views and subscriber counts, checking them multiple times a day. I can't do anything to improve them, so now I try and ignore them (although I am still guilty of checking about once a month).

For the last few years I have been using Plausible Analytics on my website. This is a privacy first analytics platform that doesn't invade people's privacy like Google Analytics and gives you basic stats on how many people are on your site, where they are based and what pages they visit.

As my website has gained popularity and now receives over 10k page views a month, I found myself having to pay £14.40/month (now £19/month).

I have no particular goals for my website or this newsletter, so measuring page views seems a little pointless.

It also has a negative impact as well. Obsessing over metrics about the content I create can also impact what I create. Creating videos or articles purely for the views rather than because I am interested in the topic, just causes me to lose motivation and sucks the enjoyment out of the creative process.

I have now removed all analytics from my website, so I have no idea who reads what or how many page views I get, and it is freeing.

This translates to other areas of my life as well.

I am currently trying to be more active and get fit by doing regular Apple Fitness+ workouts and walking more. The general metric for this is to jump on the scales once a week and see how much you weigh.

We don't own any scales in our house, so we don't obsess over our weight. I am interested in getting stronger and therefore if I am doing things correctly, my weight should go up not down.

The same goes for my finances. I track all our spending using Actual, making sure we keep within our budgets and that we have enough money in all our accounts.

For my investments however I try not to look at them on a regular basis. I only invest in index funds, and I am more interested in how they perform over a 10 - 20 year timescale rather than on daily basis.

In short:

Measure what matters and ignore the rest

❤️ Picks of the Week #

📝 Article - My own little patch - This is what I am striving for with my website. I am always shocked when viewing websites without my Pi-Hole running. Many websites now have more adverts than content. I have even known my phone to slow down and struggle to work with all the ads running. I wish everyone had their own little patch.

📝 Article - How Slack Built a Distributed Cron Execution System for Scale - Cron is my go to when I need to run something on a schedule on my server. Cron is great, but it doesn't lend itself to running hundreds of jobs and would become a maintenance nightmare. This article describes how it is done at Slack.

📝 Article - How a mortgage hedges against inflation - This is worth a read if you have large debts like a mortgage. At least in a decade your debt will be eroded due to inflation. Although this has been true in the past but with the current inflation rate back to only 2.3% and with mortgage interest rates above 4% it is better to pay off your mortgage early.

💻 Tech - Daylight Computer | The Fast 60fps E-paper and Blue-Light Free Tablet - My biggest issue with my Kindle is how slow it is. This is what puts me off getting something like a remarkable tablet for notes. This looks impressive though.

📝 Article - Trying Raycast: Part Two - I use RayCast on all of my machines but I am barely scratching the surface with what is possible. I need to read more articles like this to try and get more out of it.

📝 Article - Windows Recall sounds like a privacy nightmare – here's why I'm worried - I am so glad I don't use Windows any more. If Apple pulls this stunt then I am going to have to move to Linux. I wouldn't be surprised if Windows Recall was more about government surveillance than helping you remember stuff.

🤣 Comic - xkcd: Exponential Growth - I love XKCD.

📝 Article - slash pages - I am slowly upgrading my website to embrace the IndieWeb. If you have your own website and want to join in Robb Knight has put together this great site with all the common pages. I have added the following to my website so far: