September 4, 2016 at 3:02 pm #28500Max GuernseyKeymaster
Guy Beaver (one of our senior consultants) pointed me to an article on SAFe’s Weighted Shorted Job First (WSJF) that discuss the challenge of having the cost of delay being calculated as equal to Value + Time Criticality + Risk Reduction and/or Opportunity Enablement (RROE). In a nutshell point is that if there is no value or RROE then the time criticality doesn’t matter. Something critically urgent with no value doesn’t need to be done (quadrant 4 of Covey?). There is actually more going wrong with the SAFe WSJF than just this.
Our SAFe webinar page has a 13 minute video on the SAFe WSJF. At the end we discuss how assuming Value, Time Criticality and RROE are equal doesn’t make sense. Due to the normalization of these three factors (remember, when calculating them, each of these is “normalized” by having the smallest value be changed to one with the others being relative to this. However, this means that if overall RROE doesn’t really amount to much but there is a large variation, then that number can skew the overall ranking. For example, true Value can range between $100,000 and $300,000 while true RROE ranges between $10,000 and $80,000. Let’s say we have the following situation:
Feature 1: Value= $100,000 RROE = $80,000
Feature 2: Value- $300,000 RROE = $10,000
The reality, of course, is that we don’t really know the true dollar amounts which is why SAFe WSJF uses relative estimation. But that ay result in:
Feature 1: Value= 1, RROE 8
Feature 2: Value= 3, RROE 1
This is clearly wrong.
One can fix this by weighting the different columns.
The point is that taking shortcuts to make something easier does not make it better
TSeptember 4, 2016 at 3:15 pm #28505Max GuernseyKeymaster
Another challenge with SAFe WSJF is that it substitutes effort for duration in the denominator. Effort and duration are correlated, but not the same. When working at a client Israel Gat (the Agile Executive whom I’m blessed to have on our team) took me to task after he saw me describe WSJF to a client using SAFe’s method instead of Don Reinertsen’s approach. We, of course, had a conversation about it. I should have known how it was going to turn out because anytime I’ve disagreed with Don it’s always just been an indication that I was going to learn something later. 🙂
The thing is, effort and duration are not always correlated. A project that will take 30 person months will have different durations (and therefore CoD) based on the number of people, and their effectiveness in working together). This was not news to me, I was just ignoring it 🙂 Jason Yip wrote a blog almost 4 years ago Problems I have with SAFe-style WSJF Don’s reaction to this was somewhat pragmatic as he didn’t get into what was right or wrong but appreciated that at least Dean was getting the idea of WSJF out in the world – which is definitely a good thing.
- You must be logged in to reply to this topic.