Every time I finish a project, I try to make a balance report for myself. I look at the things that went wrong and the things that went right. The things that I tried and worked, and those that I shouldn’t repeat. Even those times when that final balance came up negative, maybe the project was worth it if in the next one I don’t make the same missteps again.
I have combed those balances and extracted five of the most important lessons I have learned. Some are beginner’s pitfalls. Others I should have solved by simple logic but matters weren’t so obvious when I was in the middle of things. Here they are:
It’s difficult to quantify, but ever since I started using Scrum in my projects, the number of hours spent per feature has gone down —and in my opinion quite considerably so. We were able to gain 80% of Scrum’s advantages with three things. First, the daily 15-minute follow-up meeting. Second, keeping the graph of the state of the project that tells us whether we’re doing well or not. And third, delivering software every the two weeks. Also, every fortnight, clients seemed to feel more comfortable when they received the items that we had agreed on two weeks before. In July and August it was not possible to meet. So every week we delivered a simple two-page document with the percentage of completion of each feature. This document included two paragraphs explaining why those percentages were that way. This helped us to regain credibility in a project that was quite difficult —especially in September when they could check that those percentages were real.
Not everything is positive in Scrum
Being a Scrum Master takes far more work than simply being a project manager. In those daily 15-minute follow-ups you will be told about lots of difficulties that the team needs to overcome before they can continue. Trying to find solutions will take the rest of your morning.
And having a deadline every 2 weeks can be exhausting. You cannot sprint for months. Six months later you start trotting (and that, with luck). You need to take this into account from the first planning meeting on.
Last, but not least, Scrum is not magic. Whatever method you use, you are going to need a team trained for the job at hand, ready to roll up their sleeves and work, and willing to contribute. Teams like this don’t grow on trees. If you have one, your job as a project manager is to get out of the way as much as possible.
Adding more workers to a project that’s behind schedule will only delay it further
This was sufficiently covered a long time ago in the book The Mythical Man-Month. Even so, it’s still necessary to highlight it, as there are hardly any exceptions to this rule. No matter how tight the deadlines are nine women cannot create a baby in one month.
Who owns all this?
No matter what we call it —designating the Product Owner, identifying the stakeholders, engaging them—the end we’re going to need to know who is going to validate the project in real life. And I do not mean who is going to green-light the invoice. You’ll also need to know who is going to use the product. The project will be a success only if, after delivering it, it works and it’s useful.
At some point during the project the area manager gave us all the documentation, validated the partial deliveries, tested the whole application, and congratulated us on our work. Regrettably, when his secretary came into the training session a week before deployment, she said: “that’s of no use to me: that’s no longer the right template, as I need to gather different information now.” This meant a week of extra hours and additional effort, on top of the risk of deploying a product that could be unstable.
Minimum Viable Product
If you already have something that could be helpful to the user, deliver it, deploy it, put it up for sale. Don’t wait until you have each and every one of the features finished. If you remember the Pareto principle, with just 20% of your planned features you’ll be able to cover 80% of the uses that your product will have.
When it’s in production, you’ll be able to get feedback from the users. They know with more precision what they need, and you’ll know where you’ve failed and how you can fix it. If you bet on just one final delivery, you only have one bullet to hit the bull’s eye. If we had done this, it would have saved us a lot of problems with the product of the previous example.
You can find texts like this and many other about how to manage agile projects in my book Agile 101: Practical Project Management (available on Amazon).