Pick an example or paste your own project list in priority order, with a rough size for each. This compares rotating through every project at once, which is how most teams work, against a hard limit on how many are open. Same projects, same total effort. The only change is how many are in flight.
| Project | Size |
|---|
Copy two columns (name and size) from Excel or Google Sheets and paste into the first cell.
The charts above start with raw math: switching between projects is free. In real life it isn't. Every time you come back to a project, you spend time reloading it: finding the doc, rereading the Slack thread, remembering where you left off. The more projects you juggle, the longer that takes. Click Our team's real result under the slider to see what happened when my team was juggling about 5 to 10 projects each.
Gerald Weinberg's rule of thumb: juggling 2 projects costs about 20% of your time, and 5 costs about 75%. Work block is how long you stay on one project before switching; half a day means two projects a day.
Here's what my team shipped over 17 months, straight from our project tracker. We put a WIP limit of 2 in place at the start of February 2016.
| Before | After | ||
|---|---|---|---|
| Median days, start to finish | 17.5 | → | 4 |
| Projects shipped | 30 | → | 94 |
| New projects started per month | 3.4 | → | 13.6 |
Each line is one project, from start to finish. If anything, this flatters the "before" side. We kept pushing start dates back until someone finally got a few hours of real focus on a project, even though the meetings and Slack threads about it had started weeks earlier. The team also grew from two people to three or four over this period, so the gain per person is smaller than the raw count, but still more than double. Small requests under about 45 minutes aren't shown at all (see rule 8), and one outlier that sat open for 13 months is left out.
The math only works if you hold the line, and holding the line means rules your stakeholders will find irritating. These are the ones my team ran on. Each one comes with the thing people heard us say a hundred times, and the reason behind it. One role matters most: a manager who owns the queue and takes every conversation about future work, so the people building never have to. On my team, that was me.
We aren't working on that yet.
A limit with exceptions is a suggestion. The charts above are what you give up every time you squeeze one more in.
You'll get me, not the person who's going to build it.
Every "quick question" about a project that hasn't started is a context switch for someone deep in one that has. The manager absorbs those so the builders don't.
Please put it in the project, not in my DMs.
A project that hasn't started has no status to report. Asking anyway just interrupts the projects that do.
Add it to the project item.
Stakeholders can write requirements, examples and screenshots into the project whenever they like. Writing interrupts no one, and when the project starts, everything is already there. This is the release valve that makes rules 2 and 3 bearable.
Why does this matter more than what's next in line?
When a slot opens, the manager picks the next project from the ones with a clear case. Priority is earned by impact, not by who asks loudest or most often. It also takes "playing favorites" off the builder's shoulders.
What can we cut?
A fixed size forces the scope conversation up front, while it's cheap. Whatever doesn't fit becomes its own project and competes for its own slot.
What happens if this lands a week later?
We honor dates where late really means too late: a renewal, a launch, a compliance date. We don't need pressure to move fast. Focus does that. Artificial urgency just reshuffles the queue and adds switching.
It's on the mini-request list. Friday afternoon.
Anything under about 45 minutes, like a page layout tweak or a new user, skips the WIP line and goes into a separate queue. We burned it down on Friday afternoons and in the gaps: between projects, right after lunch, or the morning after a company happy hour that ran a little late.
Inspired by Joel Spolsky's Human Task Switches Considered Harmful (2001), Paul Graham's Maker's Schedule, Manager's Schedule, and Kim, Behr, and Spafford's The Phoenix Project.