Use Pull-Systems to Manage Knowledge Workers

photograph of a woman thinkingManaging knowledge workers is an endlessly fascinating challenge.  How do you know when a knowledge worker is actually working? How do you know if they are working at, or near, capacity?  When I was a software developer a coworker used to joke that he was more effective because he smoked, which gave him time to stop and think.  I’ve heard other knowledge workers say when they get into a rut or hit a seemingly insurmountable issue they head straight for the gym, take a walk, or go for a run.  Their motivation isn’t to escape from the issue they’re facing rather to clear their minds so they can approach solving it from a different angle.  While they’re out running or smoking or walking are they working?

One way to better manage knowledge workers is to use a pull-system.  Most work teams are firmly entrenched in using push-systems.  Those are systems in which work is pushed onto performers as soon as it arrives, whether or not they are ready to receive it.  In a push-system there is no throttling of the work unless it is handled outside the system.  For example, a manager may sense his team is too busy to accept more work and intentionally hold back incoming requests until team members are free to accept more.  Most often this is not the case.  As more work comes into the system it is dropped directly onto the people who are expected to resolve it.  The individual team members are responsible for prioritizing and managing the work in process.

In a pull-system things get turned around.  Workers in a pull-system focus on completing a limited number of tasks and do not accept additional work until they are ready for it.  Establishing a pull-system allows knowledge workers to maintain an equilibrium where no new work enters the process until activities in progress are completed.

Pull-Systems and Little’s Law

Little’s Law tells us the average number of items in a system is equal to the product of the average number of items arriving per unit time multiplied by the average wait time.  Stated another way, as more work enters the system it expands the average time individual work items will remain in the system.  For example, if a team can complete 5 items per week, or 1 per day, and 10 items suddenly enter their workflow, the average time to complete each item will be longer than 1 day.  As the system becomes overloaded there is a corresponding decrease in efficiency.

Little’s Law seems painfully obvious and yet it is regularly ignored.  Converting work processes from push-systems to pull-systems empowers the workers to regulate their flow of work to avoid becoming overloaded.  When that happens the average time it takes to complete a work item comes down, managers have a better idea of how productive their teams are and more work actually gets done while quality improves.

Why Pull-Systems Struggle

As intuitive as pull-systems are when they are explained, they are not as common in the workplace as one might hope.  I’ve been asking myself why and have come up with the following theories.

  1. Pull-systems are counter-intuitive.

While Little’s Law may make sense, many would say if you want people to get more work done you have to give them more work to do.  In a push-system this equates to loading people’s to-do lists with everything you need them to get done.  In a pull-system everything goes into a backlog and items come out as quickly as workers are ready to receive them.  To a manager who has a long list of items to accomplish it can be a big leap of faith to put it all on a “we’ll get to it some day” list while their team sets the pace.

  1. Pull-systems may be viewed as a passing fad.

With so many management techniques and philosophies being pitched every year, it would be easy to assume pull-systems are just the latest management fad requiring a large investment in time and training.  In reality, pull-systems have been around for decades, they just haven’t been widely used outside manufacturing.  Thankfully they are fairly easy to implement and don’t saddle organizations with a large learning curve.  Teams successfully implementing pull-systems are typically able to do so in a matter of hours or days.

  1. The benefits of pull-systems are not well publicized outside of manufacturing.

Once people are exposed to the benefits of implementing a pull-system they tend to agree they have many advantages.  Unfortunately those advantages are not obvious to most of us until they are explained.  This simple fact is perhaps one of the biggest reasons more teams don’t adapt pull-systems.


Pull-systems are a simple and effective way for knowledge workers to self-regulate their volume of work.  While it can be intimidating for some managers to allow their teams to set the pace, it’s tough to argue they aren’t already doing so.  After all, how do you know whether your database administrator is really working or just surfing the web?  If you’re looking for a way to implement a pull-system in your environment, Kanban is an excellent technique.  Based on lean manufacturing principles, Kanban can be readily applied to operational activities and software development without changing existing processes or adding significant overhead.  Kanban gets teams to visualize their work, limit their work in process, and utilize a pull-system to optimize the flow of work through existing processes.