🐈

Upgrading the big site is hard for multiple reasons. One is that every new feature has to be upgraded to the same level as the other existing features. E.g. if I had a basic feature at level 500, if I activated a level 1 feature user happiness would drop.

Reminds me of how everything Apple does has to be Apple quality. (also, innovator's dilemma)

Another reason it's hard is because the chain of requirements for the modules necessary for the more advanced features (for more potential users) is difficult to track. I ended up outsourcing huge blobs of it so I wouldn't have to trace through what each employee/group was doing

Even though every employee was kept busy, the final output was dependent on the bottleneck. The bottleneck is VERY difficult to track. It moves all over the place from team to team, department to department, especially as things change.

The managers in the game r stupid. I could only give them production targets instead of more high level goals like "Make stuff for feature A". I'd have to comb through production plans to track down the bottleneck and often find one time making hundreds of unnecessary stuff.

Marketing is very slow relative to the world population. E.g. something like 20k users a day is huge from a absolute sense but slow at the millions/billions of users level.

Marketing is like watching a plant grow. The user base grows over years and years. This was actually surprising to me. I had a mental model in my mind where getting front page on reddit or something would max you out on all potential users.

Between marketing (actual users), development (potential users), and sysadmins (hosting) I felt like I was managing "streams of growth". Marketing always had to have room to grow so development had to keep pace. Hosting had to keep pace with usage.