Engineering in Lockdown
We didn't ship any product features in Q1/2021, but we did a lot of work!
In December last year, while discussing the product roadmap for 2021 at unico, the amount of work we had ahead of us seemed endless. At the same time that we were motivated with a promising future, there was also worry about the amount of company’s objectives that would fall, obviously, onto the engineering team to be accomplished.
Looking at the team and roadmap, it was clear to me that we had a lot of technical debt and not a lot of time allocated to deal with these issues. Engineers were feeling the weight of the repetitive and uninspiring work that comes with accumulating technical debt, and consumed a considerable amount of their time. I started to think how to reverse this tendency.
We also needed to organize how the engineering team interacted with other departments to ensure any changes made in the technical side of the house were sustainable. Our ways to escalate day to day issues were veery organic, and this also made the weight of the operations fall, a lot of times, onto the same people’s shoulders. The million dollar question was: how to bring engineering quality become top of mind in the whole company in a short amount of time?
I had an idea, but in the beginning even I thought it was a hard sell, but I proposed to the management team anyway. My thesis was simple: if we focus Q1/2021 in engineering quality, eliminating technical debt and better structuring the workflows of the engineering team, in the following trimesters we would be able to deliver the product roadmap in its entirety. That’s when I proposed a Lockdown in engineering: no product features for a whole quarter. If the roadmap will be delivered deserves a follow up from me towards the end of the year. But the first quarter certainly motivated the start of the change we were looking to make in the team and in the company. We are an engineering company, with focus on sustainable long-term success.
To start, we made planning very simple, asking for short bullet points of what problems they thought had to be resolved. I asked them to please be ambitious in their planning. The list didn’t surprise anyone, and had a bit of everything: from monitoring enhancement, to better testing and re-architecture of some components in the products. We embraced it all and let the team do its magic.
Autonomy was part of the design. Used to live under day to day pressure from the operations, I felt the team was a bit lost at the start. We needed the whole company to buy in this idea for it to work, and in this regard, the whole unico team deserves a huge acknowledgement because with rare exceptions everyone respected the focus and space of the engineering team, and as I like to say, let the team do its job.
While the team worked in the technical challenges, the management team was focused in bringing new people into the team, giving extra hands to those teams that needed them, encouraging internal movement amongst teams and taking care of everyone’s onboarding. Only in the first trimester we hired 17 people into the engineering team and 100 people total at unico. Besides that, we planned and validated a big org change following the tribes and squads model of autonomous teams with very well defined roles.
Three months later, we felt the time went by really fast. But we learned a lot - and much beyond engineering. In the post-Lockdown survey, over 83% of the team rated the learning opportunity as large or very large in the period. We worked closer to each other and collaborated like never before as a team. We practiced letting go of solutions that no longer serve us. To incentivise information exchange and quality of solutions, every week we had 2 engineering reviews booked, where people could come ask for feedback on what they were doing. We started projects that involved the whole team, getting people closer to one another and enabling learning and exchange within the team.
Some teams reduced customer escalations by half. Others learned about inefficiencies on their backend configuration. We prioritised security, standardised monitoring. We improved documentation of our products internally and externally and invested on test infrastructure improvements and test coverage. We planted a lot of seeds that we’ll certainly harvest later on.
It was a huge housecleaning job, and we left the Lockdown energised with the challenges for the remainder of the year and also, with the certainty that we still have lots to do to minimize the inevitable technical debt that is so easily to accumulate over time.
As a company, we left this period committed to continue dedicating time to engineering excellence items, and committed to respect the necessary investment it requires. For that, our product and engineering planning is now done explicitly calling out how much is being invested in new features vs engineering items. In the end this is a marathon and not a sprint and we need to ensure our products are built for the future.
One step at the time, together we are building an united engineering team, strong and we delivery quality of engineering through simple, reliable and scalable products to our customers.