Not all tech debt is created equal.
I ran a tech debt workshop at Thrive. The formula I used to discuss this was (Blast Radius * Frequency of Pain) + (Engineering Velocity * Trend)
Blast Radius - What breaks if this fails?
- 1 = Single team feature
- 3 = Critical path for product area
- 5 = Company-wide outage, data loss, security breach
Frequency of Pain - How often does this suck?
- 1 = Quarterly annoyance
- 3 = Weekly friction
- 5 = Daily impediment
Engineering Velocity Impact - How much does this slow us down? (I was the least happy with this scale wording)
- 1 = Slight slowdown for specific tasks
- 3 = Blocks certain types of features
- 5 = Grinds multiple teams to a halt
Trend - Is it getting worse?
- 1 = Stable annoyance (flat)
- 3 = Slowly degrading (linear)
- 5 = Accelerating problem (superlinear)
Interpretation
- Score 30+: Critical - Immediate priority
- Score 20-30: Important - Next quarter
- Score 10-20: Moderate - Roadmap it
- Score <10: Monitor