The False Efficiency of Task Splitting

Organizations often cite efficiency when splitting people’s time across multiple roles and tasks. Rather than hiring additional people, the reasoning goes, we can simply ask our existing people to pick up the slack and do several different things.

By way of example, one organization I work with using Agile/Scrum on multiple projects has elected to have a single individual from each team play the roles of “team lead” and “ScrumMaster“. Although they aren’t convinced it’s a bad idea yet, my feeling is that they are on a dangerous path.

As I framed my argument against this specific policy, I realized that, as is often the case, the Agile principles seem to coincide with the economics of the situation.

Argument 1) Knowledge workers aren’t really fungible. If individuals are “split” between tasks they become ineffective and unproductive (see “Slack” by Tom DeMarco).

DeMarco argues that task switching is a leading cause of productivity loss due to 1) the mechanics of switching tasks, 2) the need for knowledge workers to immerse themselves uninterrupted in tasks, and 3) the frustration that accompanies being interrupted too often (see Slack, pg. 16-21). DeMarco refers to evidence generated in multiple empirical studies indicating that on average most workers who are multitasking lose at least one hour in every eight hour day to task-switching. Over the period of a project, these lost hours add up to significant amounts of money wasted on non-productive work.

This “task-switching” problem also directly affects the productivity of the other team members on the team (ripple effect) who now lack a fully productive team member providing focused direction and technical decision making.

Argument 2) There is also a logical conflict of interest when team members act as SM simultaneously. Scrum requires constant communication and relationship building between the product owner and the development
team. Someone needs to act as a mediator between the product owner and development team when conflicts or discrepancies arise, and traditionally this is the SM. However, if the SM is on the team the SM cannot be impartial arbiters of conflicts since they are at the same time team advocates.

Argument 3) A major SM role is to take care of the administrivia overhead and impediments that affect the team. If the SM is also a team member then there is no real net benefit to the team. Consider why professionals often have secretaries; it would be ineffective for the professional to do administrative tasks when someone more capable in that area could handle it which would free up the pro to focus talent more specifically on their area of expertise.

Practically, however, I know that often team members act as SM when there are a limited number of people available and lots of work to do. It’s not a disaster but it is inefficient use of people’s time. At one point, the economic inefficiency of task splitting outweighs bringing on a dedicated SM. I’m curious to hear what others think on this subject.

Victor Szalvay

Victor Szalvay currently leads product development for CollabNet’s ScrumWorks® product suite. In that capacity, he works closely with customers, stakeholders, and the development teams to deliver high business value each release cycle. With more than 150,000 active users worldwide, ScrumWorks is used by more than half of the Fortune 100 and boasts the largest market share of any Agile management tool.

Posted in Agile
5 comments on “The False Efficiency of Task Splitting
  1. Scott Hampton says:

    In medical product development (my world) we have loads of Really Smart People working our projects. The great thing about RSP is that they are smart; this has a downside, however, for they are also easily distracted.
    Since we have multiple projects, and sometimes the biochemist is only needed for a few weeks every month, everybody is part of several teams. Invariably each person works the project they find most interesting, so as the program manager I spend a lot of my time coaxing RSP to work project critical tasks (the boring stuff).
    Even computers don’t really multitask. RSP pretend they can, but it is an illusion. We have arrived at the method of Bite Sized Pieces — if the biocompatability expert’s role is to handle material validation, then the work is to be in chunks that can be completed in one day. That BSP may have a week on the calendar, but the RSP is to spend the day on the task.
    We have fixed times each day of the week for all meetings, and Fridays are blocked out entirely for work, with the result that people have nice chunks of time to work on the ONE thing that needs to be done.
    YMMV.
    Good stuff here, BTW, I’ve linked your site.

  2. Pete Behrens says:

    Victor,
    You make a good case for the distinction of roles on the team. However, I believe it is idealistic in most cases. To avoid a long comment response and to propose a separate question, I blogged my thoughts separately at http://trailridgeconsulting.typepad.com/ pete_behrens_blog/2005/10/process_facilit.html

  3. Victor Szalvay says:

    Scott: Excellent ideas. I agree that blocking out uninterrupted periods for work is a fantastic idea. I haven’t been able to achieve that level of discipline for myself, but I’ll work on it 🙂
    Pete: I enjoyed your response very much. It appears we are reading the same books, so we are familiar with each other’s references. Your point about Organizational Patterns and cross-functionality is well taken and I’d like to clarify it. I also encourage cross-functionality at both the team, and if possible, the team member level. I believe the team should do whatever it takes to complete committed work inside the iteration, and most times that is best done by overlapping roles to increase communication bandwidth.
    DeMarco calls it “task splitting”. “Task” for agilists has a specialized meaning. The way I read DeMarco’s meaning was splitting someone’s time between totally unrelated tasks, like separate projects or in this case sharing two distinct and independent roles.
    Your inquiry about the part-time facilitator: some facilitator is better than none, and I would take a dedicated part-time facilitator over a team member acting as facilitator some of the time. For one thing, the team reclaims a full-time team member (who is no longer split between roles). Second, a dedicated facilitator, even part-time, can focus all attention and energies on the task at hand, rather than being sucked away for extended periods of time (as I often witness) by a vortex created by another equally important responsibility.
    Thanks for the fantastic comments guys!
    — Victor

  4. Mark Yarbro says:

    In response to Pete’s commentary above I’ve posted my equally lengthy observations on his blog, but the short version is that I think Victor was making an important but subtle point that multi-tasking of diverse activities, while often necessary, increases project risk in a manner that is not immediately obvious.

    The shorthand is that the amount of time and effort wasted on task switching is NOT constant, but increases substantially when the roles differ greatly (for example, construction roles verses support roles) and can cause fatal blind spots that end-up costing much more in time and effort than was saved by the staff reduction that necessitated the task switching in the first place.

  5. Denis Haskin says:

    This is discussed at length in Tom DeMarco’s excellent book Slack

Leave a Reply

Your email address will not be published. Required fields are marked *

*