Discover more from The Analytics Engineering Roundup
Becoming an impact-oriented data team
Also: metrics playbooks, user journeys, the state of data testing, what new year resolutions have to do with OKRs and some cheerful data!
🎆🧧🐇 Happy Lunar New Year to everyone celebrating this weekend! 🐇🧧🎆
Before we get into some hard data stories this week, a little whimsy — here’s a delightful interactive visual of things you can find at a Lunar New Year dinner table:
I’ll have nearly all of these on my table tomorrow as I celebrate with family and friends. 🀄anyone?
In this issue:
Introducing the metrics playbook, Ergest Xheblati
The state of data testing, Pedram Navid
The best new years resolutions you can make, Cassie Kozyrkov
Most data professionals are process people. We get so caught up in developing clean systems and picking the right technology that we lose sight of what we’re building all of this for.
Blake, I wasn’t ready for this real talk. 😭 Where I think Blake is spot on:
Data teams are cost centers in many organizations today. This is a real thing, and a hard place to occupy in 2023.
Impact driven data work isn’t only a good way to make sure you stay on the right side of that balance sheet, it feels good to do. Teams that drive business value also have better team member retention — it’s highly motivating to see exactly how you’re making a difference.
Qualify, test and iterate before you commit. Yes, it’s easier than ever to model data today. That doesn’t mean we should aim to be complete from the get go — sometimes a quick and dirty notebook (or yes, even a spreadsheet) that is timely can get 90% of the way to unlocking business value because the data is available at the time the decision needs to be made.
I will add one more to this list:
Becoming more impact oriented as a data team must come from a place of high agency. It’s difficult to do well, and often requires a big shift in team process, relationship with stakeholders, and time allocation. Believing that this is possible to achieve is actually a big component of being able to get there!
These lessons can be hard (and expensive) to learn for a data team. Making impact-orientation a new years resolution this year for your team can go a long way towards ending 2023 by being absolutely indispensable to your organization.
Thanks for reading The Analytics Engineering Roundup! Subscribe for free to receive new posts.
Introducing the metrics playbook
I get really excited when folks talk about generalizing common business data needs:
… at their core organizations are not that unique. They generally fall within a limited number of business models (SaaS, eCom, marketplace, professional services, hardware, etc.)
As such, many of the core business processes are the same across the orgs as long as the business model is the same.
We can easily instrument and measure these activities and define a standard set of metrics.
YES!!! 🙌 🙌
If you’ve been reading this Roundup for a while (and/or listening to the podcast) you might have seen several different takes on this idea before. It’s certainly not new. But we have yet to do this well, or even systematically, as a field.
Ergest offers one reasonably well established example in this article: growth accounting. It appears we can look forward to more detail in future posts, and I for one will be watching this conversation closely!
The state of data testing
I agree with Pedram that we really don’t talk about testing enough. And that probably means we don’t do enough testing.
In this piece, Pedram does some light user research via Twitter and breaks down the four maturity tiers he is observing in the state of testing today. I think they’re probably the right ones. I also think that most organizations likely don’t get past #2 or #3 today. #4, when done well, is pretty life changing.
This leaves me wondering: will there be a TDD movement in the data space the way we saw this permeate software engineering in the early 2010s? Or is ChatGPT going to write all of our tests (and SQL) for us? ;)
Randall wrote this piece in 2016 and only needed to update it slightly to repost in 2023. It’s still relevant today as it was then — defining a user journey in data remains just about as elusive as it was 7(!) years ago. We have many more tools now, but maybe that’s part of the problem?
This article is a really good reminder that (just like business metrics!) there are certain core analyses and frameworks for data work that are foundational, timeless, and likely.. in need of some really good templates because you’ll be doing them again, and again, and again…
The best new years resolutions you can make
Yes, we are way past the early days of 2023, but humor me on this article for just a second. Read it in full, enjoy the bitter sweet taste of feeling seen by Cassie as you really try to turn over a new leaf this year… and then read it again but imagine applying all the 15 tips to how you design your company OKRs for the year.
Resolve to create a failure action plan
This one has a lot of depth to it, both if you think about setting personal and team/company goals.
It’s easy to feel like you’ve failed when you get too busy for your brand new exercise routine midway through February. If January feels fresh, exciting and full of possibility, February often feels like a slog as you struggle to keep up with the objectives you’ve set for yourself while your life kicks into high gear. Eventually, you fail out of your new routine because you’re trying so hard — it stops being a net positive in your life because it takes too much time/effort/etc.
I think we run into similar fallacies when we set company/team objectives for the year. If (when?) we see targets starting to slip, it feels obvious to double down and just try harder. Do more with the same amount of resources — the organizational equivalent of “powering through”.
Here’s Cassie’s advice for what to do instead:
If (when!) you encounter a wobble, don’t try to solve it with willpower. Solve it by designing a better path to your goal.
If you’re not getting to the objective you set for the year, figure out where you are experiencing friction and work on removing that friction. Simple. Elegant. Timeless. :P
And finally, if you’re looking for something positive to add to your data digest, I present to you the Cheerful Data blog. Nuggets of data insights that are designed to remind you that some things, somewhere are trending up 📈
Like the number of people not living in extreme poverty: