Software Engineering at Google Chapter #7 - Measuring Engineering Productivity (1 of 3)

  • Be a data driven company
  • As organizations grow linearly, communications costs grow quadratically
  • By increasing the productivity of individuals, you can increase your business’ scope without the increased communication overhead
  • Linters help code reviews scale
  • Create a team dedicated to engineering productivity. Have it’s members be from a wide variety of fields (psychologists, software engineers, behavioral economists, and more).
  • This team is essentially “Engineering productivity specialists”
  • Is a metric worth measuring?
  • Does measuring the metric slow down the organization?
  • Tracking progress may change engineers behavior (“observer effect”) and the behavior change may serve to mask the underlying issues
  • To know if a metric is worth measuring, ask people to describe what they want to measure in the form of a solid, concrete reason or answer. The more concrete the answer is, the more likely the metric is worth measuring
  • Beware “anticdata” (single anecdotes that masquerade as data), it is powerful
  • In addition to the question above, ask the following questions:
    • If the data supports your expected result, what action will need to be taken due to that result?
    • If no action will be taken then there is no need to measure
    • What result are you expecting, and why? (This acknowledges biases at the beginning)
    • If we get a negative result, will action be taken?
    • Does a negative result change your decision? If not, then why measure? You are not using the results of the measurement to make a decision, but rather to justify an existing decision.
    • Who will decide to take action on the result and when will they take action? This ensures the person asking for the measurement is empowered to make decisions based on the results or that the person is representative of someone who can make the decision. The other purpose is to determine who the decision maker is so we can get them the data in a form that they want (quantitative vs qualitative)
< BACK NEXT >
Tweet


   


   

Thank you for your time and attention.
Apply what you've learned here.
Enjoy it all.