Continuous Project Planning

Continuous Project Planning is a set of ideas and principles designed to address some of the inherent difficulties associated with planning and managing projects.

Core ideas

Continuity

The core of continuous project planning is to always keep the plan up to date, continuously monitoring that it leads to the fulfillment of the goals, and to continuously adapt it to reflect changing conditions, evolving circumstances, and new knowledge.

The plan is a map of the future showing what you will do to reach your goal. With an inaccurate and out-of-date map you risk running aground.

Continuous means that the plan is updated every day with what work actually has been done. It is updated as soon as you gain new insights into how much work is required, and whenever resource availability, scope, or priority changes.

Good information is key to making good decisions. By continuously updating, you are always ready when decisions need to be made, and you will discover if the success of the project is at risk earlier, giving you more time for corrective actions.

Manage uncertainty

Uncertainty and risk are inherent parts of projects. While reducing risks is good, they cannot be eliminated. You need to manage the uncertainty and risk that remains.

Make the uncertainty in estimates clear by quantifying them. Estimate time for tasks or activities with one number for the best-case scenario, and one that is reasonably sure will be enough. Define reasonably sure as a 70% probability, and assume an exponential probability distribution. While certainly not completely accurate, this method is simple and will give you much more information than a single number will.

Use these estimates to calculate uncertainty and probability distributions for both total amount of work and for when milestones will be reached.

While a risk register for more specific risks is good, having the risks in the plan and in the schedule makes them more visible, and easier to prepare for. Discrete risks, if realized, will cause more work and delays. Even when you do not know how much, you can put estimates with wide uncertainty ranges in your plan.

Agile with predictive planning

For an organization, being agile means being able to quickly adapt to change. Predictive planning outlines what is expected to happen and what you will do in the future, so that you are prepared.

Continuous project planning is about both having a plan so that you are prepared for the future, and continuously updating it and changing it to reflect the most current information you have.

The predictive specifier in predictive planning seems superfluous. Planning is a predictive activity. It is there as a distinction from Agile methodologies, where agile is a reference to the “Manifesto for Agile Software Development”. The fundamental idea of the manifesto is essentially to make sure the project delivers value, rather than to efficiently deliver on time.

Continuous project planning is not a framework giving you exact recommendations on how to make sure your project goals and priorities remain relevant. You should, but how depends on your organization and the nature of your project. Regular recurring meetings are good, but if it is possible to address problems and changes as soon as they are known, it is even better.

Another principle from the Agile Manifesto is not to plan more than you need. With continuous project planning this is reflected at two levels. You can start off a plan with large activities with lots of uncertainty. Put in the work to reduce the uncertainty and refine the plan only when you need to. The other is in the short-term planning, where you can rely on the feedback from time reports instead of adjusting detailed planning in advance.

Continuous project planning is a hybrid approach that promotes agile response to change and addresses several of the non-software-specific points of the Agile Manifesto. How to implement practices to ensure new external information is gathered and used to update goals and priorities is left out of scope.

Tools and automation

In software development, there is continuous integration and continuous delivery. Both require tools and automation. So does continuous project planning.

The automation required to make continuous project planning possible will not only give you better plans, it will also save you time on project administration.

Automatic scheduling

Automatic scheduling based on priority, dependencies, resource availability, and estimates is key to automation in project planning. It separates your plan of what to do from when it will be done. It makes changing plans easy, and it lets you see the effects of changes immediately.

With manual scheduling you risk overbooking resources, assigning too much work to individual team members. With automatic scheduling that does not happen.

Automatic feedback

The work actually done will not always match what you have planned. By using time reports from the team to update the plan and the schedule, you will always have a plan that reflects what has already happened. Work that did not happen today will be scheduled for tomorrow instead. Any impact on milestones or deadlines will be visible in the automatically updated schedule, no status meeting required.

Performing tasks out of order, or helping a colleague, will not break the plan; it will be updated.

Probability calculations

Estimates with uncertainty and probabilistic risks can be handled using Monte Carlo simulations when applied to a timeline schedule. Automatic scheduling is a prerequisite for this to work.

The result is probability distributions for when tasks and milestones will be reached, and probability numbers for whether milestones will be reached on time. It also results in a probability distribution for the total amount of work, and thus for the associated cost.

This eliminates the problem of merge bias, and it enables risk-aware decision-making.

The tool will not only look at a single project, but will take into account the accumulated uncertainty from all projects sharing resources. A delay in a different project may block a resource and create a bottleneck in your project. Without a tool, this can be very difficult to discover.

Scenarios and contingency plans

When it is easy to change plans and see the effects, you should expect your tool to let you test and experiment with different scenarios. Explore what happens if a risk becomes reality and prepare a contingency plan. Explore what impact a change in priority or scope has before making a decision. Compare strategies to select the most effective, or least harmful, ensuring all stakeholders understand the implications.

Optimization

A project planning tool should not only detect problems but also help identify the best solutions. When you accept that there are uncertainties everywhere, there is also a probability associated with where bottlenecks will occur, and whether a mitigation strategy will work.

Optimization is not just about solving problems. It is also about seizing opportunities. Use the tool to discover underutilized resources, and help you assign work to get the greatest benefits. Let the tool help you differentiate between delays that can cause the project to fail, and those that will not have any adverse effect at all.

It is common to organize project work in a way that makes project planning manageable, trading efficiency for predictability. With a tool that handles the scheduling complexity, you can organize work to be as efficient as possible, and reduce the overall risk. When resource sharing across projects becomes manageable, experts can be utilized more effectively and the team can be mobilized to address problems as they arise.

Another common strategy to make planning manageable is to create intermediate deadlines and plan for arbitrarily sized buffers. With a tool that handles uncertainties on the timeline you do not need to waste energy on meeting deadlines that are meaningless, and buffers are built into probability margins you have for the truly important milestones.

Work with human nature

Some aspects of human nature tend to make project planning more difficult. We can’t change human nature, but by being aware of the problems we can adapt how we work to reduce them.

Over-optimism is one aspect. Evolution has made us optimists, because in evolutionary times it has been better to try and fail, than to certainly fail by not trying. This tendency is especially strong in group settings, and causes us to overlook and neglect risks and problems.

Another is the will to please and to present ourselves in a good light. This can cause us to be overly optimistic even when we know better. Sometimes caused by peer pressure, sometimes to show off a brave can-do attitude.

Humans do not like to be wrong. Once we have made a time estimate or created a plan, we do not really want to change it, as this would mean admitting we were wrong. This leads us to stick with our initial statement much too long, delaying the discovery of problems and reducing the time available for effective mitigation.

Two number estimates

Estimating with uncertainty is valuable not only for the extra information it provides for the plan. It also helps us achieve better estimates. All the bravery and optimism can go into the first number for the best-case scenario. Then you consider a number you are reasonably sure will be enough. This number helps you consider all the things that could go wrong, and shifts the perspective toward cautious realism.

Defining ‘reasonably sure’ as enough time in seven times out of ten cases gives it a precise definition. This helps everyone involved make estimates in the same way and speak the same language.

Humans are generally bad at probability, especially regarding things they have not done before. While 70% may be an arbitrary number, it aligns well with the description ‘reasonably sure’, and is a probability level we are able to reason about.

Focus on the future

Do not ask if the plan holds or if a task will be completed in time. Ask how much time is remaining. This puts the focus on making a new accurate estimate, with uncertainty. Not on whether your first estimate was good or not. Even if the total estimate has increased, it does not mean that the initial estimate was wrong. You are supposed to exceed the reasonably sure number three times out of ten.

By not forcing people to admit they were wrong, you are more likely to receive accurate information about the current state and expected future.

Avoid measures such as ‘percent done’ and ‘green, amber, red’ signaling. They do not work well with human nature. Focus on the amount of work left and the forecast on when it will be completed.

You cannot do anything about the work and time already spent in a project. You can do something about the future.

Empower the team

Micromanagement and rigid planning benefit no one. Not only does it create more work for the micro-manager, it takes away agency and joy from those doing the work.

A shared plan that visualizes priorities and dependencies enables team members to independently make well-informed decisions about priority. Automatic feedback makes sure all work counts, and small deviations from the schedule do not break the plan. This promotes spontaneous collaboration to solve problems, and to focus on the overall best for the organization over personal goals.

Do not break up your plan in too small tasks. Use logically scoped, well-defined activities of reasonable size. This will not only reduce the number of items to manage in the plan and make it more readable and easy to understand, it also allows the expert performing the work the freedom to do it in the most efficient way. Reasonably sized means that an activity can cover several to-do items, and take several days to complete. Logically scoped is the important keyword.

Stress is one of the greatest health risks for office workers. Being empowered and having influence over your situation reduces stress and makes work more enjoyable.

Applicability

The concepts and tools described here have been developed and tested for product development teams, characterized by prevalent risk and uncertainty, the need to coordinate availability of key personnel, the need to coordinate between projects, and the need to predict delivery time.

Even though this describes the primary target, it has proven useful whenever there is a need to plan for an uncertain future, regardless of team size.