A founder's note by Rajiv Sankarlall, editor of The Founders Report.
Split your attention across three active projects and you do not get a third of your best work on each one. You get a fifth. Software management researcher Gerald Weinberg's task-switching heuristic, cited by Scrum.org, holds that one project gets 100% of a person's effective time, two projects drop that to 40% per project (20% lost to context switching), and three projects drop it to 20% per project, with 40% of total capacity burned on the switching itself. I ran more than twenty active projects at once before I did that math. The number that stopped me was not the twenty. It was the realization that past two or three, the marginal project is not additive. It is subtractive.
The Apple Precedent
When Steve Jobs returned to Apple in 1997, the company was carrying roughly 350 products and was, by most accounts, about 90 days from running out of cash, after a fiscal year that closed with a $1.04 billion loss (Inc.com). Jobs is reported to have stopped a product review meeting and said, "Stop! This is crazy," before drawing a two by two matrix, consumer and pro on one axis, desktop and portable on the other, and cutting the line down to four quadrants (Inc.com). Ten products replaced 350. One year later, Apple posted a $309 million profit (Inc.com).
The lesson founders usually take from that story is cut hard. The more useful lesson is procedural. Jobs did not cut by rank ordering 350 items on a spreadsheet. He built a small structure, four boxes, and only what fit inside the structure survived. Twenty projects cannot be triaged one at a time in your head at 11pm. They can be sorted against a frame in an afternoon.
The One Idea Your Mind Actually Runs
Paul Graham's argument in "The Top Idea in Your Mind" is that a person has exactly one dominant idea their thoughts drift back to during unoccupied moments, in the shower, on a walk, half awake before the day starts. Whichever idea holds that seat gets the deep, undistracted thinking. Everything else gets whatever attention is left over, which after project four or five is very little. Graham's warning is blunt: "it's a disaster to let the wrong idea become the top one in your mind."
The American Psychological Association's research on multitasking explains why the wrong idea is so easy to install. Each individual switch between tasks costs only a few tenths of a second, small enough to feel free. But that cost compounds when switching happens repeatedly across a day, and people who multitask heavily show measurably more difficulty filtering out irrelevant information than people who do not. Running twenty projects does not just cost time per switch. It degrades the filter that decides what deserves attention at all, which is the exact mechanism that lets a low value project occupy the top slot Graham describes.
The problem with twenty projects is not that you cannot get to all of them eventually. It is that you lose the ability to tell which one deserves to be first.
Cutting Is Continuous, Not a Crisis
Founders treat killing a project as an admission of defeat, which makes them delay it well past the point where the delay costs more than the project was ever worth. Large, well resourced companies do not operate under that shame. The Killed by Google tracker lists 299 products, services, and hardware devices that Google has shipped and then discontinued since 2006, an average of roughly one every three and a half weeks for two decades. Google has more capital and more engineers than almost any company alive, and cutting is still a routine part of how it runs a product line, not a signal that something went wrong.
Twenty projects taught me that the cut works better as a standing discipline than as a single dramatic move. Jobs's 1997 purge reads, in hindsight, like a crisis event. Google's graveyard is the more honest model: cutting on a rolling basis, before the crisis, as ordinary maintenance.
What to Do This Week
- Count your active projects and run the Weinberg curve against your own list: two projects, roughly 40% effective time on each; three, roughly 20% each; the per project share keeps shrinking while the share lost to switching keeps growing.
- Do the Jobs exercise. Not a ranked list, a small forced structure, a two by two or something similarly tight, and see how few of your projects actually land inside a box.
- Whatever survives the grid earns the top slot in your own mind, the one Graham describes. Everything else goes on a graveyard list, reviewed on a fixed schedule, cut without ceremony.
Twenty active projects is not a portfolio. It is twenty ways to make sure none of them gets the thinking it actually needed.