Project tracking

There is an enormous literature on project tracking. It was written for teams of thirty, with a project manager whose whole job that is, and meetings whose only purpose is finding out where everyone else has got to.

That is not your situation. You are two, sometimes one, you are building a product, and the person who ought to be running the status update is also the person writing the code. Everything below starts from there.

Why tracking stops, in a small team

In a small team, tracking does not die from a lack of seriousness. It dies because it is redundant.

When there are two of you, you already know where the other one is: you spoke this morning. So keeping a board becomes work that tells nobody anything, which is the definition of a chore. Three weeks later, nothing is up to date.

The problem is real, but the usual conclusion (“we don't need tracking”) is wrong. What you know when there are two of you is what is happening this week. What you lose is everything else: why you picked that approach two months ago, what you decided not to do, and how much is left before the product can be shown to anyone.

Three indicators, and not one more

A project dashboard that is useful to two people fits on three lines.

1. What is in progress, and who has it

Not the list of everything left to do: that one is depressing and serves no purpose. Only what is open right now.

The rule that makes all the difference: one thing at a time per person. If three tasks are marked “in progress” for the same person, none of them really is. The board shows you that at a glance, which a conversation never does.

2. What is stuck

A task that has not moved in ten days is not a late task: it is a task waiting for something. An access you do not have, a decision you have not made, an appetite you have lost.

Those three are treated differently, but none of them gets better by being left alone. The only point of tracking, here, is to make visible what has gone quiet: because what has gone quiet never announces itself.

3. Where the current phase stands

This is the figure most often missing, and the only one that answers the question you actually ask yourself: are we making progress?

Cut the project into short phases: weeks, not months. Every task belongs to a phase. The phase's progress then works itself out from the tasks that are done.

That is the important mechanism in this whole article: a dashboard must never be updated by hand. A figure someone has to type in is a figure that will be wrong in a fortnight, and a wrong dashboard is worse than none: it reassures.

What you can drop today

• The Gantt chart. It assumes you know how long tasks take and in what order they go. When you are building a product, you find out both as you go. So the chart is wrong by day two, and keeping it current costs more than it teaches.
• The weekly status meeting. With two people it is a conversation you have already had three times that week. Keep a longer monthly one instead, whose subject is not “where are we” but “are we still going somewhere sensible”.
• Estimating in hours. It gives a precision that does not exist and turns every overrun into a failure. “Small / medium / large” is enough to know what you can take on this week.
• Percent-complete on a task. A task is open or it is done. “80% there” is the polite way of saying you do not know.

Notes matter as much as tasks

This is the part tracking tools forget, and the part you will miss most.

A finished task disappears, and with it the reason you did it that way. Six months later you run into an odd-looking decision, remember nothing, and have the whole debate again from scratch: sometimes reaching the opposite conclusion, having forgotten why you ruled it out.

So write down, next to the tasks: what you tried that did not work, what you decided not to do, and why. Three lines are enough. It is the only document whose value grows with time.

Tracking while building a SaaS

Two features of this kind of project are worth adapting the board for.

The product is never finished. There is no final delivery, so there is no finish line to measure against. What usefully replaces overall progress: the date of the next thing you can show. A short phase always ends with something you can put in front of a person.

Half the work is not product. Search, support, billing, chasing invoices: that is real work, it takes real time, and it appears nowhere if the board only contains features. You then end up convinced you are not progressing, when in fact you simply never counted what you were doing.

Put those tasks on the same board as the rest. They are allowed to exist.

What it looks like after six months

Tracking done this way will not make you work faster. That is not its job, and be wary of anyone who promises it will.

What it gives you is quieter and more valuable: you will know what you have done. On the day the doubt arrives (and it always arrives around month four), the difference between “I feel like I'm going in circles” and a list of three finished phases is not a difference of organisation. It is a difference of morale.

And if you keep one rule: anything that has to be updated by hand will eventually stop being updated. Build your tracking so that doing the work is enough to keep it current.