Marc Andreessen's 2007 blog contained some of the best writing on business and startups, and this piece on retention is a perfect example. The article addresses a challenge every growing company eventually faces: how to keep great people when growth slows and the stock price stalls.
His central insight flips conventional wisdom on its head: retention isn't really about retention at all—it's about winning.
"Companies that are winning—even really big, old ones—never have a retention problem. Everyone wants to stay, and when someone does leave, it's really easy to get someone great to take her place."
This explains why all the perks, raises, and HR initiatives in the world won't fix a retention problem if the company isn't winning. The $6,000 heated Japanese toilets and Olympic lap pools are band-aids. Great people want to work at winners, period.
Key Strategies
• Companies don't have retention problems—they have winning problems. Great people naturally want to work at winning companies, and when growth stalls or stock prices tank, retention becomes an issue that can only be truly solved by returning to winning.
• Focus retention efforts on "magnets"—the key architects, managers, and leaders who attract and retain other great talent. Without these critical people, you'll lose everyone else.
• Clean house aggressively by removing "vesting in peace" employees, summertime soldiers who joined when times were good, and mediocre performers. This improves morale for remaining talent and frees up resources.
• Restructure for clarity and ownership by promoting your best people, simplifying org structures, eliminating matrix reporting, and giving stars direct responsibility for critical missions with smaller, higher-quality teams.
• Change the narrative through bold moves—reorganizations, acquisitions, new product initiatives, or transformative deals—to reintroduce energy and curiosity while actively recruiting new talent and re-recruiting departed stars.
What Resonates
The "nuke all matrices" advice is particularly striking. Matrix organizations sound sophisticated but they diffuse responsibility and slow everything down. Giving your best people clear, direct ownership—even if it means tolerating overlap—is worth the apparent inefficiency.
The section on talking people out of joining startups is also gold. The key is showing them how staying for two more years will make them better prepared for eventual startup success. Most people aren't good at picking startups, and the experience of running larger operations is invaluable.
What Andreessen doesn't say explicitly but implies throughout: if you're not willing to make dramatic changes—fire mediocre people, promote aggressively, restructure radically—you're not serious about fixing the problem. Half-measures won't work.
The article is part of a larger series on big companies that's worth reading in full at pmarchive.com.