The Classification-aware Defect Ranking (CDR) Model uses a Defect Ranking Score to provide a suggested sequence for defect resolution that considers a compliance/regulatory need to resolve and then satisfying the broadest population with the least amount of capacity consumption.
The Defect Ranking Score is calculated by adding the ‘need to resolve’ divided by the ‘effort to resolve’ to a classification bias.
The need to resolve considers the Reach, Impact, and Pressure. The more people affected by the defect, the greater the consequence of the defect, and the urgency of resolving the defect are used to determine the ‘need to resolve’.
- The need to resolve is captured as ((R + I) x C + P), where C is your confidence in knowing your customers and how they use the system when you select the R and I options.
The effort to resolve considers the knowledge that those who will be working to resolve the defect have about the possible causes of the defect. If the team has a low level of knowledge about the problem domain or the code base then the effort to resolve will typically be larger than when a team has a high level of knowledge about the problem domain and the code base.
How the need for the CDR Model was identified
Starting with WSJF, which is the suggested prioritization for enterprises adopting SAFe, people find that the variables used to estimate the cost of delay are not well aligned with the typical language used when prioritizing defects. We tend not to talk about the value of fixing a defect, the reduction of risk, or the business opportunity created when fixing defects.
When discussing defects the conversation often considers how many people are affected, the impact on these people, and the severity of that impact. For this conversation, the RICE prioritization model is a better fit for the language used. Looking at the first 2 parameters we see that this model provides a better alignment with the conversation:
- Reach represents the number of people affected, and
- Impact represents the consequence of the defect’s effect.
The shortcoming with RICE, when used with defects, is that it, like most prioritization models, is intended to be used on a single backlog of similar items. But, not all of the work in a backlog is the same type of work. A backlog typically has a number of work types, such as New Features, Work on Technical Debt, or Maintenance items. SAFe handles this using Capacity Allocations. Defects are another type of work and should have a capacity allocation, but then there are sub-categories for defects that we need to consider.
Defects have sub-categories => The need for the CDR Model
Defects have a ‘natural’ separation into at least 2 work types or classifications.
- First, there are the defects that are known internally, but not by the users
- And, then, there are the defects that are known by the users and have been reported
- They may, or may not, have been known internally before being reported
In some industries, healthcare, for example, some defects also relate to compliance or regulatory requirements.
Beyond capacity allocation to work on defects, we need a prioritization model that considers defect sub-categories.
All things being equal, most enterprises would benefit from a prioritization model that ranks compliance or regulatory defects higher than known customer-reported defects. And would rank known issues higher than defects known only by the solution provider.
The Classification-aware Defect Ranking (CDR) Model provides an objective ranking of defect resolution efforts considering the natural order of resolution efforts
Why use a model like CDR when I can prioritize in other, simpler, ways?
“If everything is a priority, then nothing is a priority”
High, Medium, and Low ranking is insufficient when you have more than 3 defects!
How to Conquer the Tyranny of the Urgent | Psychology Today: In a world where everything feels important, and we constantly feel overwhelmed trying to stay on top of it all, it’s critical to ruthlessly prioritize tasks and obligations based on their true level of importance. This struggle is called the “Tyranny of the Urgent”
Being ruthless
Ruthless has a negative connotation, but you have to be able to say no to people, even yourself. That is good, because you can stop feeling overwhelmed and focus on your day knowing that you are making the most significant and appropriate contribution you can make during the day.
Using a defect prioritization model that has sufficient fidelity, like CDR, that considers typical prioritization factors enables you to objectively say no, or at least not yet, to those defects with a lower calculated rank. The calculation takes the subjectivity out of the conversation and the fidelity of the ranking values helps you avoid everything being ‘Priority 1’.
The Classification-aware Defect Ranking (CDR) Model
For enterprises that have a significant backlog of defects,
Who are dissatisfied with their current prioritization approaches,
The Classification-aware Defect Ranking (CDR) Model,
Provides clarity on those parameters to be considered for an objective approach to determining a ranking score that is the basis for prioritization,
Unlike other ranking approaches that tend to be subjective and inadequate in capturing the various considerations.
For the CDR Model, the defect classification is based on a mandate to repair and the breadth of exposure (who knows about the defect). The CDR Model uses this classification and consideration of Reach, Impact, Confidence, Pressure, and Understanding to calculate a Defect Ranking Score.