July 2016 Archives

Seeing the underlying resource thinking

Creating stable teams and having teams work from one priority are two critical ideas in large-scale product development. This means when we discover some work, we put it into the overall backlog with appropriate priority; then, all teams share same priority, and any team may do the work.


However, two common counter ideas often pop up.


1. Change work priority to match team skill


Only the team with matched skill will be considered for the work, while other teams will not, even they are available. What will other teams do? Because they should not be idle, they do work that has lower priority but matches their skill.


2. Create project team to match individual skill


When the work requires skills across groups, people from those groups are pulled into a short-term team only for this work, so that individuals can do the part matching their skills. This is the traditional project team approach.


I felt a bit frustrated when I heard those ideas again during my recent consulting. It seems so deeply held, thus I decide to make a deep analysis resulting in the below CLD (Causal-Loop Diagram).


Resource thinking.jpg


The two counter ideas, which are in fact solutions on how to organize people for work, are illustrated by two B-loops (B1 and B2).


The B1-loop reads like this:


      the more delivery pressure -> 

      the more consideration for skill in prioritization ->

      the more match between work and skill ->

      the more efficiency ->

      the more value delivery ->

      the less delivery pressure


The B2-loop reads like this:


      the more delivery pressure -> 

      the more consideration for skill in forming team -> 

      the more match between work and skill -> 

      the more efficiency -> 

      the more value delivery -> 

      the less delivery pressure


They both try to create the match between work and skill, thus optimize for efficiency.


However, they also create intended/unintended consequences illustrated by other loops in the diagram.


1. Less value delivery


This is illustrated by the R1-loop, which reads like this:


      the more delivery pressure -> 

      the more consideration for skill in prioritization -> 

      the less value delivery (as it constrains prioritizing high value items) -> 

      the more delivery pressure


As the value delivery is affected by both value in items and the number of delivered items (efficiency), R1 and B1 together create an economic tradeoff. Depending on which loop is dominant, it may deliver less value, even though the efficiency goes up.


2. Lower team performance


This is illustrated by the R2-loop, which reads like this:


      the more delivery pressure -> 

      the more consideration for skill in forming team -> 

      the less stable the team is (as the work is dynamic) -> 

      the less efficiency at the team level (as it takes stability to perform at the team level) -> 

      the less value delivery -> 

      the more delivery pressure


R2 shows the fix (create project team to match individual skill) that backfires, as those teams exist only in short term while the stability is essential for team performance.


3. Less learning


This is illustrated by the R3-loop, which reads like this:


      the more delivery pressure -> 

      the more consideration for skill either in prioritization, or in forming team (actually two loops) -> 

      the more match between work and skill ->

      the less learning ->

      the less efficiency (in the long term) ->

      the less value delivery ->

      the more delivery pressure


R3 shows the fixes (change work priority and create project team) that backfire, as they both create more match between work and skill, which reduces learning!


What we are seeing in the diagram is really the underlying resource thinking.


  • Skill utilization is more important than value delivery
  • Team is no more than sum of individuals
  • People are fixed in one skill and they cannot learn

As learning is essential in any product development, the mismatch between work and skill is not something nice to have, but must have. By creating stable teams and having teams work from one priority will you create the mismatch to promote learning!


Build a Real Management Team

Real development team is already rare, let alone real management team. When talking about real team, I mean that the group has common goal and takes shared responsibility. Usually, management team is not a real team, but a working group. This is indicated by every member having clear individual responsibility. Even though it is claimed that they have common goal, it is mostly the leader who is accountable for it, and other members are mostly responsible for their own parts. This is the traditional organization hierarchy.


As the hierarchy is defined by reporting line, let's look at various settings when we create cross-functional feature team in adopting Scrum.


1. from multiple managers to one manager


As the traditional organization is often based on functions and components. When we move from single functional and component team structure to cross-functional and feature team structure, if original reporting structure remains, members in the same team will report to different managers. Those managers are still responsible for their own functions and components. Even though they are members in management team, they do not take shared responsibility. This is not desirable for team, as the more managers, the more difficult it becomes in alignment. 


Reporting line - functional.png


Therefore, the common advice is to change the reporting line and make the whole team report to same manager. Then, those managers become cross-functional managers. However, this does not make them take shared responsibility in management team, as usually they are still responsible for their own teams.


Reporting line - cross-functional.png


2. from one manager to multiple managers


Recently, in my LeSS course, a few participants from same product organization shared that the members in their teams report to multiple managers, but NOT managers for different functions and components. Instead, it was a bit like everybody randomly picking up a manager. It sounded weird at first, then, it revealed an interesting purpose of doing that. They want to build a real management team!


Reporting line - random.png

The rationale is, when any team is not owned by one manager, it would avoid the situation of any manager only responsible for their own teams. Instead, multiple mangers will collectively  ensure the success of all teams. In this setting, any misalignment among managers would become serious impediment towards well-working team. What impressed me was that they made informed choice, with the attitude of experimentation.


This reminded me of my NSN (Nokia Siemens Networks) experience. When I led a department with 100+ people, we had 4 R&D managers, together with me we formed a management team. We put deliberate efforts in building a real team - taking shared responsibility for the whole department goals.


We tried:


  • Managers acted as ScrumMaster for teams who did not directly report to them. See Manager as ScrumMaster paper.
  • Created department initiatives to foster collaboration among R&D managers, especially on removing organizational impediments, continuous improvement, and community of practices.
  • Not cascade down department objectives to next level for R&D groups, but kept it at department level.
  • Instead of me giving the feedback to every manager as part of performance appraisal, we had management team feedback session, everybody giving and receiving feedback in the open.


We did not try having members in one team report to different managers by random. What an interesting experiment, ever though its effect remains to be checked.


I applaud for any attempt in building the real management team, as it is hard, but a lofty goal to seek.

About this Archive

This page is an archive of entries from July 2016 listed from newest to oldest.

June 2016 is the previous archive.

January 2017 is the next archive.

Find recent content on the main index or look in the archives to find all content.