This week we discussed sections 3.3 through 3.5.
3.3 — Visions and Strategies
- Being out of alignment as a manager of other managers (i.e., understanding what the team is actually doing) is an opportunity to miss out on potential productivity.
- Terminology
- Vision — what we want to become
- Strategy — how we will achieve the vision
- Tactics — individual decisions that implement the strategy in the midst of the work
- Example: My vision is to get a suntan at the beach. My strategy is to drive to Fort Lauderdale. One tactic is when I reach Atlanta, do I take the bypass or drive through?
- Strategy
- Diagnosis — a very thorough problem statement
- Policies — description of the approach that you’ll take
- Actions — you need to follow through and be accountable
- “When you read bad guiding policies, you think, ‘so what?’ because it’s found a way to justify entrenching the status quo. When you read good guiding policies, you think, ‘Ah, that’s really going to annoy Anna, Bill, and Claire,’ because the approach takes a clear stance on competing goals.”
- You can have multiple strategies as long as they don’t compete.
- Jamie gave an analogy from geology: Porosity is worthless without permeability. You can’t have flow without space for things to flow through.
- Vision
- Get input because the first draft won’t be the right one.
- Jameson said policies can fail if they are dictated by those who don’t understand everyone’s perspectives and how those at the bottom would be impacted.
- Embrace and learn from failure, experiment, iterate.
- This section is a playbook for how to create a vision.
3.4 — Metrics and Baselines
- The notation of probability — e.g., “p95” — was new to most of us. This means you can be 95% sure that a random sample will be in some range.
- Two tests of an effective goal: 1) someone without much knowledge of the area can understand how difficult it is, and 2) that person can evaluate whether it was successfully achieved.
- Know your target state, where you are now, the trend in change, and the time frame to achieve it.
- Jamie has experienced resistance in suggesting that management start measuring things. Jameson continued with the notion that some people don’t appreciate making their work visible, as it might reveal how little they’re doing or how much is waste. This is especially common in environments where people going a long time with no oversight. For example, if you measure how much you’re spending on infrastructure and propose a solution that would save money, management has to admit they were wasting money.
- Sometimes you may not know a concrete dollar amount for current processes or proposed changes.
- When defining service level objectives (SLOs), only commit to the things you can control (e.g., disaster recovery) based on how much risk you want to take. This is strategy.
- Specify as few investment goals as possible — limit WIP. You can have too many good things to do, but you can’t do them all. The same goes with OKRs — don’t commit to too many.
3.5 — Broad Organizational Changes
- One team’s performance may be driven by dozens of other teams.
- The terminology in this section — context, nudge, cohort, baseline — contains words Lirio uses for its services.
- Houston liked the concept of nudging with push notifications that occur when a team’s metric changes. It’s clever, and jibes with Agile’s tight feedback loops so that people can learn quickly.
- Houston also mentioned a talk about feedback loops to help developers recognize code smells that involved learning the differences between house sparrows and song sparrows.