Managing projects with Wardley Maps — part II — cutting costs

This is a continuation of a series of articles about managing projects with Wardley Maps. Previous part can be found here.

At this point, you should have a map of your project (like one below), and understand which components bring uncertainty.

A map created for a project.

Uncertainty revisited

Uncertainty itself is usually undesired. It is responsible for contract amendments, missed deadlines and ever increasing costs. The map above has only 20% of moderately uncertain components, so you can plan the work for other, more stable components, and know nothing about how much the project will take.

What’s more — this work for 80% of components does not constitute 80% of work. It’s an extrapolation trap.

Ignoring uncertainty while crossing the river by feeling the stones.

The 20% of unstable components can require much bigger effort than the rest of the project (especially if that rest is executed efficiently). Unexpected obstacles can also completely derail the project or even entire company.

That said, at this point of time, you should focus on removing uncertainties from the project. This process is itself unpredictable, as you never know what you will find out. Of course, if someone had done something similar, you may assume the task can be completed. It may turn out it is more complicated than it looks, though. Anyway, the best way to deal with uncertainty is to remove it by experimenting. You have to verify that:

• you know how to build and operate uncertain components. In the case of the map above, it means f.e. discovering a way to identify the best time to post anything on social media. This is a vital, uncertain part of the system. You do not want to discover you have no idea how to implement it at the end of development process.
• you know how to integrate mature components. Two mature components may not work together out of the box, and connecting them may not be trivial.

That’s plenty of proves of concepts to create. Since experimenting is unpredictable, you have to some fixed time for research, and evaluate results.

Once you have dealt with uncertainty, it’s time to think about… time.

The flow of time

Time is very important for project managers, as it usually constraints the project. But what is often forgotten, is the fact that the simple pass of time changes the value of the work being done.

What is a competitive advantage today, will be a boring cost of doing business tomorrow. High margins will be replaced with volume operations, which means that whenever you are comparing costs of buying vs implementing, you need to take into account expected price drops.

Secondly, even if the cost of outsourcing stands still, a $1000 today is worth more than a $1000 in 5 years, due to the inflation.

Both those aspects of time have tremendous impact on your project, as you have to consider not only the cost of doing something today, but also possible future changes.

Look, for example, at the user management activity on the map above. Such a service delivered by an external provider may cost $5 per user per year. Building the service by yourself costs $25 000. If you design for a thousand of users in 5 years, those numbers seem to be identical.

But they are not, and not only because users register gradually, which means that on average you will pay for 500 users per year when you use outsourced approach, but also because $25 000 and a month spent today is worth more than $5 000 spent for 5 consecutive years.

Buying the service from an external provider gives you a couple of very nice advantages:

• smaller initial investment, in terms of both, time and money. You can reduce the impact of potential failure and simplify your project by making it shorter, and letting your customer (or boss) use saved resources more efficiently.
• less capacity issues and maintenance. Think of what is going to happen if you spend $25 000 to build user management system for a thousand of users, but only a hundred will register? What will happen if 1100 user will appear? What will happen if user management practices change to f.e. biometric authentication? Somebody will not be happy about that, and it is good to not be a scapegoat here.
• financing from profits. If you pay as you go for external services, you are in a very favourable position, where you can have your customers pay you first, and then use that money to cover costs. If your project fails in the business sense, this means it will be unused, which in turn means low maintenance bills. One more factor that reduces the risk associated with the project.

Said all of that, there is one caveat. None of the above will happen if the component has a lot of associated uncertainty, because frequent, unpredictable changes will increase the cost of maintenance. A good rule of thumb is to look for candidates for outsourcing in the right part of the map.

This part of the map usually contains great candidates for outsourcing.

Majority of customers does not care about system architecture. It is supposed to ‘just’ work, so by proper use of outsourcing, you earn precious time and resources to focus on what is really important — on user needs.

Project value vs Costs

Summary

Previous post has stated that invalid assumptions about project goals doom it from the very beginning. This one, a little bit more down to earth, ends with the following conclusions:

1. Uncertainty is your enemy, and you can’t remove it without getting your hands dirty. You have to experiment before you are able to evaluate how much will it take to deliver the project. The tool let’s you represent research effort in this way:

A component with a linked research effort.

You may want to link the research to a wiki/intranet webpage, where progress is tracked.

2. Planning costs. So does researching. Ensure your customers understand that. If necessary, teach them how to use mapping.

3. Consider time as a very important factor that constantly changes the value of your work. Focus on things that you can do better than the market, even when considering the pass of time (future proof, not so difficult if you know the concept of Evolution).

Atlas allows you to mark a component as outsourced.

If you are afraid of using Atlas and giving up control over your projects, you can run it behind your company’s firewall, as described on the tool wiki. LDAP is supported, as well as Google auth. Details on user management are described on this page.

In the next article, I will describe how to share effort and costs between different projects.