Development Reviews Still Pay Off for Small Teams
First, let’s look at why development reviews don’t get adopted.
- Scheduling alone is a hassle
- There’s no one suitable to act as reviewer
- The development team is very small
Conducting development reviews is a basic premise. If there’s a single legitimate reason to skip one, it’s only item 3, the size of the development team. When development is done by an extremely small team, detailed design information and status can already be shared between members without a formal review session. This is especially true when software development properly uses a system like git, or when hardware development separates the person who draws the schematic from the person who places the order. In those cases, daily work already functions as review. This is a very convenient situation, and in my experience, a team can naturally sustain this kind of close coordination up to about 6 people in size.
Once a development team grows past a certain size, though, things stop working that way. The only time the two ends of the team learn what the other is doing is the weekly regular meeting. Task management sheets and progress status sit on a shared system, but nobody actually looks at them. When a team like this finally holds a review session, it’s often the first time anyone has seen anyone else’s schematics. Comments from someone seeing something for the first time tend to hit the essential point. Throwing naive questions at each other is a sure way to make a review session productive.
Now let’s turn to the challenges. Scheduling is a genuinely troublesome problem, and it’s often the very reason reviews end up as empty formalities. The person in charge of a given department can never make the schedule work and keeps missing sessions, or a review gets rescheduled over and over until months pass and the rescheduling itself is forgotten. This is an organizational problem, so each organization has to solve it on its own, but let me offer a few ideas.
For example, something like “9:30-10:00 on the second Wednesday of every month.” A time slot when no regular meetings are normally held is preferable. When there’s nothing that particularly needs review, release that slot. Since it’s a reserved slot, scheduling anything else into it is forbidden. This method works well for organizations with a fair number of products or projects.
This relates to item 2 as well: a reviewer needs a certain amount of experience. Not just to make sharp observations, but to look at things with a consistently skeptical eye, informed by an awareness of certain patterns of mistakes. People suited to this role turn out to be surprisingly scarce. By designating that person in advance and guarding their schedule carefully, you cut down on the effort of coordinating everyone else’s. At large companies, this role often falls to a veteran engineer, and depending on circumstances, it can also be worth looking for a suitable person externally, such as a freelancer.
Of course, one could also try to solve this through a somewhat punitive system for enforcing reviews: have a manager strictly oversee it, or build a system that prevents shipping without a review. That approach carries stronger enforcement power and would likely work faster. But it places a heavy burden on small and mid-sized companies and doesn’t suit them well. Finding an operating method that imposes as little strain as possible is a key point whenever introducing any new process into an organization, not just development reviews.
The content of this post is part of (the original text of) the following book. If you’re interested, please get a copy.
The Shape of a Happy IoT Startup
The Shape of a Happy IoT Startup
Originally published in Japanese at https://clazytech.com/2022/09/1149/. Translated with LLM assistance and reviewed before publication.