I've been called to task - asked why I teach teams to distinguish between Task Hours and some non-denominational non-specific non-absolute unit. The latest team I guided used PEZ! In the daily stand up a person might be heard to say:
Benefits of "Task Points"
Because it works and is more fun. Not a sufficient and rational reason.
Because it's a fractal of the concept of Story Points. Not a practical enough reason.
Because it encourages transparency and acknowledges that we are really bad at estimates. Not sure we wish to propagate that notion.
Because Scrum has a value of Openness and in that spirit can we admit that a team of 7 people doing 50 "hours" of tasks a week is really a poor metric. Is the metric WRONG - or is it being perceived poorly - or do we just not know what we have measured?
Have we truly measured the sprint progress in the units of HOURS? You know that aggregate of 60 minutes. That man-made concept of time. That universal unit that all societies agree upon, even the Imperial Unit based USA.
Let's take a look at a team's burndown, using traditional hours. It shows that in the first week of the sprint they accomplished 50 hours of work. But there are 7 people working for the week. Generally speaking, that's 7 X 40 hr/week = 280 hrs. Now many people like to point out that there are not 8 hrs of work in a typical workday. Many people believe that every day contains a few hours of non-productive hours. Many believe there are 6 hrs of productive time in a typical day (7 X (6hr X 5days) = 210 hrs). What can we deduce from our burndown? Maybe what we've been estimating and burning down is not this well know unit of time called an hour.
We have been estimating in hours - so it must be hours. This is a myth. The analysis at RescueTime has shown this myth to be busted by looking at the habits of use of their tracking tools. Their tools calculate a productivity rating in addition to hours on tasks. To spoil the read... "And with an average productivity pulse of 53% for the year, that means we only have 12.5 hours a week to do productive work."
So maybe that study puts an end to the concept of Ideal Engineering Hours. I've been involved in these dialogues and the general consensus is 5 - 6 ideal engineering hours per day. It does not quite jive with what a real study finds (12.5 / 5 approx. 2 - 3 hr/day).
If a team of 7 should be doing something like 200 hrs of productive work a week - then what do we think of a team of 7 only completing 50 hrs in a week? Well is that rational? No. So what's wrong with this irrational logic? One could blame the individuals on the team and try to improve their ability to estimate tasks. Yet if they are off by a factor of 4 - how much work will go into improving their poor estimation technique? And what about this empirical measurement technique? If we believe in the empirical technique then does the accuracy of our estimates matter?
My belief in transparency (openness) has me concerned that if we hang a chart on the wall that professes that the team completes 50 hours in a week - that someone is not telling the truth. It is obvious if one observes the team, as a Team Agility Guide, might do and see that the team members are working on the sprint stories for the greatest part of every day and working diligently then these units might just be the problem. When the estimate is made in good faith but doesn't equate to what the wall clock tells. Something is telling a fib, a lie. Let's quit lying and make our burndown truthful.
One technique is to realize the empirical measurement does not require hours to work, choose any unit you wish - try PEZ. And then one can correlate the PEZ to work hours. And stop talking about hours that require half a day of work to complete.
Mike Cohn's critique of "Task Points"
I've read and reread Mike's critique and advice, but I don't get his point. He finds tremendous value in Story Points, and notes that Task Points have similar disadvantages:
Perceived Benefits of Task points
How does having a team track remaining effort on work (story or task) benefit the delivery of value? Great question. Will a list of reasons do?
Is it useful to track remaining effort at a task level? Yes and No; it's situational.
Let's first define the No situation, when the team is performant, gets stories done easily, has a known velocity, no problem planning and achieving a sprint commitment.
Now the alternative situation... is it useful... I think so.
See Also:
Don’t Estimate the Sprint Backlog Using Task Points by Mike Cohn
"Yesterday I was working on the code to DisCombobulate the GUID so that we could distinguish customers from guest and the task to peek into the hashed session state was harder than I anticipated - I will need 2 more PEZ to get that done today."
Benefits of "Task Points"
Because it works and is more fun. Not a sufficient and rational reason.
Because it's a fractal of the concept of Story Points. Not a practical enough reason.
Because it encourages transparency and acknowledges that we are really bad at estimates. Not sure we wish to propagate that notion.
Because Scrum has a value of Openness and in that spirit can we admit that a team of 7 people doing 50 "hours" of tasks a week is really a poor metric. Is the metric WRONG - or is it being perceived poorly - or do we just not know what we have measured?
Have we truly measured the sprint progress in the units of HOURS? You know that aggregate of 60 minutes. That man-made concept of time. That universal unit that all societies agree upon, even the Imperial Unit based USA.
 |
| Burndown in Hours |
 |
| Burndown in PEZ |
We have been estimating in hours - so it must be hours. This is a myth. The analysis at RescueTime has shown this myth to be busted by looking at the habits of use of their tracking tools. Their tools calculate a productivity rating in addition to hours on tasks. To spoil the read... "And with an average productivity pulse of 53% for the year, that means we only have 12.5 hours a week to do productive work."
So maybe that study puts an end to the concept of Ideal Engineering Hours. I've been involved in these dialogues and the general consensus is 5 - 6 ideal engineering hours per day. It does not quite jive with what a real study finds (12.5 / 5 approx. 2 - 3 hr/day).
If a team of 7 should be doing something like 200 hrs of productive work a week - then what do we think of a team of 7 only completing 50 hrs in a week? Well is that rational? No. So what's wrong with this irrational logic? One could blame the individuals on the team and try to improve their ability to estimate tasks. Yet if they are off by a factor of 4 - how much work will go into improving their poor estimation technique? And what about this empirical measurement technique? If we believe in the empirical technique then does the accuracy of our estimates matter?
My belief in transparency (openness) has me concerned that if we hang a chart on the wall that professes that the team completes 50 hours in a week - that someone is not telling the truth. It is obvious if one observes the team, as a Team Agility Guide, might do and see that the team members are working on the sprint stories for the greatest part of every day and working diligently then these units might just be the problem. When the estimate is made in good faith but doesn't equate to what the wall clock tells. Something is telling a fib, a lie. Let's quit lying and make our burndown truthful.
One technique is to realize the empirical measurement does not require hours to work, choose any unit you wish - try PEZ. And then one can correlate the PEZ to work hours. And stop talking about hours that require half a day of work to complete.
Mike Cohn's critique of "Task Points"
I've read and reread Mike's critique and advice, but I don't get his point. He finds tremendous value in Story Points, and notes that Task Points have similar disadvantages:
- A foreign concept to many team members
- The need to establish baseline values against which relative estimating can begin
- A concern that estimates drift over time in comparison to the original relative values
Perceived Benefits of Task points
How does having a team track remaining effort on work (story or task) benefit the delivery of value? Great question. Will a list of reasons do?
- It moves the focus from how much effort - to how much effort is remaining - imagine the sunk cost fallacy;
- when an item takes longer that your typical inspection cadence (sprint for a story; a day for a task) it requires an explicit reestimation and recommitment to the item getting to done;
- when the list of estimates on an item grows long (say past 2 or 3) it is a wonderful indicator of an impediment / when a task (should be much less than a day) last in-process for a week it is a great indicator that help is needed - training a team to hold members accountable is easier with a visual signal
- when the item is done-but... and a person wants to mark a zero, but ... it's a great indication of a newly discovered task
- I suppose the metric (burndown) is more accurate to reality than an ALL or Nothing assessment of tasks/stories effort.
- the practice reinforces the inspect/adapt micro cycle, and someone once told me - if it's hard to do; do it more frequently.
- ... there may be others ...
Is it useful to track remaining effort at a task level? Yes and No; it's situational.
Let's first define the No situation, when the team is performant, gets stories done easily, has a known velocity, no problem planning and achieving a sprint commitment.
Now the alternative situation... is it useful... I think so.
See Also:
Don’t Estimate the Sprint Backlog Using Task Points by Mike Cohn
Comments