Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News How to Use Metrics to Influence an Agile Environment

How to Use Metrics to Influence an Agile Environment

This item in japanese

Larry Maccherone, a Data Scientist at Tasktop Technologies, gave a talk at QCon London 2015 regarding the importance of metrics usage and how they should influence important decisions in the organizations.

During the talk “What? So What? Now What?”, he brought to the room real-life’s examples that showed how important is to not ignoring what the data is saying at any given moment. He suggested that “every decision is a forecast” so, the odds of making a good one depends on the amount and on the quality of information that is used for it.

Maccherone remembered the audience that feeding teams and processes with inputs from past experiences is what we’ve been doing in Agile for the last few years. And that is because Agile is rooted on the principles of an iterative approach, “Agile is all about feedback”. He pushes the discussion to the quantitative feedback by introducing the Rally’s initiative on quantification of agile practices’ impact, the Software Development Performance Index (SDPI).

In order to provide high-quality information, Maccherone presented the audience with the following set of criteria:

  • Answers the question “Compared with what? (So What?)”
  • Shows causality
  • Tells a story with whatever it takes
  • Is credible
  • Is impactful in the social context and has business value
  • Shows comparisons easily
  • Allows you to see the forest and the trees
  • Informs along multiple dimensions. Is multivariate
  • Leaves in the numbers where possible
  • Leaves out glitter

Maccherone continued his journey through quantitative data by showing some trues about cognitive bias and how one can deal with them:

  • Few people are immune to it
  • We all think that we are part of that small group
  • You can be trained to get much better. Douglass Hubbard – How to Measure Anything
  • We do a first fit pattern match. Not a best fit pattern match. And we only use about 5% of the information to do the matching

He finalizes by showing how people lie with data and listing the key tips to get others influenced by your data:

  • Tell a good story
  • Become known for being right
  • Avoid wars about semantics
  • Imperfect evidence is better than no evidence
  • Change the nature of the conversation

Rate this Article