Agile engineering practices
Part 1: Do the most valuable thing, unless you can't
This is the third and final part of the 'Do the most valuable thing, unless you can't' article (see Part 1 and Part 2). In this part we look at the process of sequencing prioritized work within capacity constraints and discuss the challenges facing many organizations today.
A prioritized list of work represents the order of importance/value of work to the organization. In a perfect world, with infinite people and resources managing loosely coupled systems, work could be completed in the exact order of priority.
However, there are factors that influence the order in which work can be done, for example, inter-dependence on other pieces of work, capacity constraints, cost of delay, etc. These factors bring constraints, feasibility and logical sequence into consideration when deciding what work to do next. So where prioritization defines the work we want to do, sequencing determines the order in which we can do the work.
Most people are aware of the difference between ‘urgent’ and ‘important’. Some organizations make the mistake of scoring work with a looming deadline more highly to ensure it gets completed on time. A better approach is to give the work the correct business priority and trust the sequencing process to use genuine deadlines as one of the constraints when sequencing work. Note that deadlines can be used to game the process so it's incumbent on the management team to challenge what appear to be unrealistic or unnecessary deadlines that can 'poison' the sequencing process.
In the sequencing of work, the following steps are cyclical. Sequenced work requires constant review and fine-tuning when new information is uncovered or new insights are gained.
Organizations are often frustrated at the pace at which value is delivered, and mistakenly diagnose the problem as one of capacity. The reality is the problem is more complex and often what is missing is that they don't have a good mechanism to determine what is valuable, they don't have a clear view of the capacity they have to do work, and they don't manage the sequencing of work well.
In our view, far greater improvements and longer term gains can be made by simply working on what matters first, not working on what doesn’t matter and sticking to the old cliche “stop starting and start finishing.” A system that is principles led, which enables the business to articulate what is important and to organise itself around delivering what is important sooner and more reliably, is a more meaningful approach.
Senior management and general management must actively work together to determine the most valuable work to meet the desired objectives. Each has a defined role and must hold the other accountable; senior management own and must be held accountable for the objectives and the value criteria, general management own and must be held accountable for the selection of work and allocation of capacity. There are no shortcuts. This will not work if you are not all-in.
Managing work this way provides a far more realistic and reliable view of the organization's capacity to implement change. And it has the advantage of surfacing constraints and organizational blockages that can be methodically addressed through continuous improvement. Pick a small area of the business, test and learn what works. Start tomorrow.
A prioritized list of work represents the order of importance/value of work to the organization. In a perfect world, with infinite people and resources managing loosely coupled systems, work could be completed in the exact order of priority.
However, there are factors that influence the order in which work can be done, for example, inter-dependence on other pieces of work, capacity constraints, cost of delay, etc. These factors bring constraints, feasibility and logical sequence into consideration when deciding what work to do next. So where prioritization defines the work we want to do, sequencing determines the order in which we can do the work.
Most people are aware of the difference between ‘urgent’ and ‘important’. Some organizations make the mistake of scoring work with a looming deadline more highly to ensure it gets completed on time. A better approach is to give the work the correct business priority and trust the sequencing process to use genuine deadlines as one of the constraints when sequencing work. Note that deadlines can be used to game the process so it's incumbent on the management team to challenge what appear to be unrealistic or unnecessary deadlines that can 'poison' the sequencing process.
In the sequencing of work, the following steps are cyclical. Sequenced work requires constant review and fine-tuning when new information is uncovered or new insights are gained.
- Determine relevant factors that affect the sequencing of work. These almost always include available capacity, internal and external dependencies, and next milestone delivery date but can also include things like bundling efficiency (where it makes sense to do several pieces of work in sequence), cost of delay (where the cost of delaying one piece of work is higher than another), and quick wins (where we can pick up small pieces of work to deliver quick value).
- Determine the order of execution of work according to the agreed factors but guided by the following good principles:
a. Select the most important pieces of work first (unless there are constraints that prevent the work being done). NOTE: Make sure there is a definition of 'done' before pulling any work from the prioritized backlog.
b. Select the next most important pieces of work, keeping in mind that the first pieces of work should be completed or almost complete before the next pieces are pulled into work-in-progress. This avoids diluting the focus on a large number of pieces of work.
c. As much as possible, avoid pausing work to start new work (this often happens when there are many external dependencies). Focus first on addressing the dependencies or constraints, internalising them as much as possible to increase the level of control. This might mean things slow down for a period of time but with process maturity and experience with the new ways of working, the rate of delivery of completed pieces of work will increase.
d. Make sure the people doing the work are the ones who decide what they can and can’t do. However, challenging the work plan and the estimates is absolutely encouraged to ensure there is a balance between ambitious and realistic.
- Keep reviewing the sequence for opportunities to increase flow and minimise waste. Patterns will emerge as the process develops a rhythm and that, in turn, may influence the sequence of work of different shapes or natures.
Sequencing is as much an art as a learned skill. A person with skills and experience sequencing work is able to move between the bigger picture - what creates real value for customers and the business - and the detail - what work should go first. They are able to hold the line against strong stakeholders who may be used to pushing pet projects. They get a certain amount of joy, yes joy, from the smooth flow and execution of work. And they develop instincts about how different types of work unfold as the environment changes. Rather like a chef or a gardener or an orchestra conductor, the person with the responsibility for sequencing must develop their unconscious competence and seek to create order through a smooth flow.
Work on Your Constraints
The two most significant factors that influence what work we start are capacity and dependency. While capacity is primarily an internal constraint (i.e. we collectively have capacity to do work and we have the ability to allocate that capacity to the most important things), dependency can be both an internal and external constraint, where we require the services of someone outside the team.
There are a few good guiding principles to follow to address these two common constraints:
- Capacity - Effective Capacity should be maximised to provide the greatest flexibility to do the most valuable work. Waste work and incidental work should be eliminated, automated, minimised, rationalised, and/or optimised, where possible. A Contingency Reserve should be established in work planning and protected for use to absorb genuine overruns, accommodate urgent and unexpected work, and provide time for continuous improvement activities.
- Dependency - There are almost always dependencies where one piece of work is dependent on another being finished before it can start. However, external dependencies, where work is handed off to a business function or external team, are particularly troublesome.
We (theoretically) have better control over internal dependencies and less control over external dependencies. There is significant waste in handing off work - defects, rework, delays, excess work - that can be minimised or eliminated by internalising the dependency.
As a result, it's critically important to work out why work is being handed off and determine whether creating true cross-functional teams with all the skills and access to resources, even for a temporary period, can reduce the need for work to be handed off outside the working team. Co-location can also be a big help to reduce the distance, both physical and mental, between internal and external functions.
Ultimately, a team that can benefit from being autonomous should try to internalise dependencies as much as possible to allow them to control what work they do and adjust capacity to minimise the impact of dependencies on the delivery of work.
Summary
Very few organizations are satisfied with the responsiveness and cost-effectiveness of their delivery capability. Attempts to address this often erroneously focus on cost-cutting, off-shoring, or simple brute force to try to get more for less.Organizations are often frustrated at the pace at which value is delivered, and mistakenly diagnose the problem as one of capacity. The reality is the problem is more complex and often what is missing is that they don't have a good mechanism to determine what is valuable, they don't have a clear view of the capacity they have to do work, and they don't manage the sequencing of work well.
In our view, far greater improvements and longer term gains can be made by simply working on what matters first, not working on what doesn’t matter and sticking to the old cliche “stop starting and start finishing.” A system that is principles led, which enables the business to articulate what is important and to organise itself around delivering what is important sooner and more reliably, is a more meaningful approach.
Senior management and general management must actively work together to determine the most valuable work to meet the desired objectives. Each has a defined role and must hold the other accountable; senior management own and must be held accountable for the objectives and the value criteria, general management own and must be held accountable for the selection of work and allocation of capacity. There are no shortcuts. This will not work if you are not all-in.
Managing work this way provides a far more realistic and reliable view of the organization's capacity to implement change. And it has the advantage of surfacing constraints and organizational blockages that can be methodically addressed through continuous improvement. Pick a small area of the business, test and learn what works. Start tomorrow.
Disclaimer: The statements and opinions expressed in this article are those of the author(s) and do not necessarily reflect the positions of Thoughtworks.
