Watch or listen to this episode:

Summary and Resources

Companies need to start to A/B software development to find what customers value. Relying on planning up front based on customer feedback and research just isn't competitive anymore!

Software development is not like manufacturing. We don't have to make sure it's perfect before we start to build. In fact, that's not a very agile approach - and we'll too often build things that waste money!

When we do lean software development, we look at a lean canvas, or a business model canvas, and figure out which aspect of the business we want to impact through a change.

We need to design an experiment to test that change, and part of lean product management is setting up these experiments so we can test whether the experiments we run are pointing us towards the impact to the business we want to make.

We need to choose a measurable outcome to impact first, and figure out how to measure it. What will constitute success?

Next we need to make sure we know how we will serve both the old (pre-change) version of the product, AND the post-change version to an audience of the same size to see which one "wins". This is known as cohort analysis, and will help us avoid vanity metrics (measurements that don't really represent an improvement). When we serve both versions of a product (with and without a change) to a cohort, we need to determine how long the experiment will run. When the experiment ends, we have a learning milestone where we look at what was gathered, and decide whether to pivot or persevere.

Traditional project accounting looks at % complete to determine how efficiently a team is using budget to complete an effort. In lean software development, we need a new approach. This new approach is called innovation accounting, and it instead measures the business on how effectively it is learning about where its assumptions about what's valuable are right and wrong.

To use innovation accounting will require getting leaders on board. Unless we change budgeting to account for this, and convince anyone who was involved in the prior software project budgeting approach to use innovation accounting - we'll continue to deal with "change requests" or asking for more budget after we learn, and it looking like a failure on our part.

Innovation accounting is a concept introduced by Eric Ries, author of "The Lean Startup".

Are you A/B testing changes to validate experiments to achieve an outcome? What challenges do you have with getting support? Leave me a comment below.

Related Posts

Additional Resources