Chapter "API Governance" from API Management - Part III (Category Management)
To finish this series of posts about governance, I will tell you about three standard patterns. (The previous parts are here: 1 and 2) 1) Design authority 2) Embedded Centralized Experts 3) Influenced Self-Governance
I've seen all the options come to fruition, but on scale and in a dynamic company, I think only the third option will fly.
Design authority In this pattern, authority acts as a gatekeeper, which controls that the results of the teams meet the minimum level of quality.
- Enforcement and incentivization. The scheme is most effective when they have the power to prevent poor-quality and high-risk decisions from being made.
- Talent distribution. In this scheme, a small number of decision-makers are concentrated in the team of an authoritarian designer.
- Costs and benefits. The main advantage of this scheme is that all APIs go through the same command for quality control. This guarantees the right decisions, but at the same time leads to a bottleneck.
Embedded Centralized Experts In this pattern, instead of checking the results of the API team, experts are embedded in the team to help make decisions.
- Enforcement and incentivization. The introduction of experts into the project team is the highest form of coercion. If your experts make decisions that align with your core goals, so will the teams in which they are embedded.
- Talent distribution. The big challenge of managing a consulting team is finding and maintaining a team of experts. For this model to work, you’ll need a pool of API subject area experts who can be distributed across project and product groups.
- Costs and benefits. The best decisions are made early, experts shuffle their expertise with a central team, but it is difficult to maintain a common understanding of the right standards among these experts, and a large pool of experts is needed. Well, the teams themselves can be demotivated by the presence of such a political instructor in each team:)
Influenced Self-Governance In modern organizations, there is a tendency to reduce centralized control and increase team autonomy within reasonable limits for the sake of speed and innovation. This leads to a third pattern that relies heavily on influence rather than control.
- Enforcement and incentivization. This model is entirely based on incentives to influence decision-making. The central team provides golden paths to solve typical issues, but local teams have freedom of choice. They are responsible for the success of their products. Ideally, this balance encourages teams to make decisions that match the central team’s proposal.
- Talent distribution. For this pattern to work, teams must be able to make good decisions independently of each other, meaning that each team must have its own experts. (What happens in every company)
- Costs and benefits. The main advantage of this model is speed. Teams can move very quickly when they have autonomy in making decisions. However, speed is combined with the risk that decisions will be inconsistent and/or inadequate, or strongly rely on local optimization. Central governance is usually added to fix these problems.
The specific governance model may change as the structure of the company changes, for example, as the company grows. The main thing is to focus on how your system works and adjust it to changes in time:)
#Architecture #Software #Governance #Management #Leadership #Processes #Leadership