27 February 2017

Random Thoughts on Business Models and Cost of Delay

I cover Cost of Delay and CD3 (Cost of Delay Divided by Duration) in my Essential Agile Product Leadership course. (CD3 being similar to WSJF). This part of the course is derived from from the work of Joshua J. Arnold, with Mike Cohn's business modeling "dimensions" (may be my term) as a starting point in the thinking process.

Mike Cohn's business modeling stuff is somewhat software-oriented, so it may be a good place to start for Agile (or Agile-hopeful) organizations.

I think I may have divided Mike Cohn's business modeling stuff into "dimensions" so that I could better explain them, but I'm no longer certain who did what.  I also added a little to it. It's safe to say this is strongly based on the work of Mike Cohn.


The dimension of Purpose is the selection of why we are building what we are building:
  • New: New customers, new markets, improved cash flow.
  • Upsell: Upgrades and enhancements for existing customers.
  • Retention: Avoids losing customers or cash flow.
  • Efficiency: Saves money while enhancing growth.
  • Reputation: Improves our standing in the market, e.g., feature-parity, repairing defects. (I may have added this one...)


The dimension of Necessity is a rough model of how we see this feature improving value:
  • Mandatory: A Minimal Marketable Feature (MMF).
  • Linear: Adds nominal value to the MMFs.
  • Exciter: Will create enthusiasm and joy in the user community.


Lastly, the dimension of Process, which may have been added by me as a catch-all for remaining things from Mike's set, or so I could add topics that I think folks should consider when exploring their Cost-of-Delay and Duration estimates:

  • Cycle Time: Time it takes between receiving a request and delivery (or use) of the feature.
  • Waste: Things we do that may not provide value. E.g. re-work due to defects.
  • Risk: Things we may or may not be doing that will result in a loss of value.
  • Operational expenses: Keeping the lights on, and many others.
  • Cost of Delay: What it obviously costs to wait, i.e. lost opportunity. (See below.)
  • Complexity: Keeping it smartly simple.

Cost of Delay

The dimensions give leaders some starting points to think about as they come up with their cost-of-delay estimates, and helps them identify other folks to talk to--and what to ask them--in order to uncover other time and hidden cost estimates. "CD3 is cross-functional!" I say. At least as cross-functional as the best DevOps-embracing Scrum team.

I strongly emphasize that:

  1. All estimates must be provided by the people doing that part of the work, and that know the most about that estimate based on prior experience.
  2. Estimates must come with the mutual trust and understanding that an estimate is neither commitment nor guarantee, but is being used to choose between high-level options.
  3. CD3 is a starting point for ranking features, but that leaders cannot just sit back and let the equations do the release planning for them.
  4. If these techniques are not being done at the highest levels (e.g., portfolio planning), then they're going to be pointless at the feature level.
And I sometimes quip that, since value is an estimate, and time is an estimate, that the "estimate" units cancel out. ;-)