16 product things I learned at Imgur

  1. Every interface can be made simpler. One useful technique is to have a separate meeting focused just on cutting and simplifying. You can also appoint an individual ‘cutter’ to take this role outside of the meeting, but only after brainstorming.
  2. Every feature you launch is a feature you’ll need to support with users, infrastructure and development. So launch as few things as possible.
  3. Data has two roles to play in the product process: discovery and validation. Use data to guide you to what you should be building. Then use data to make sure you were right.
  4. Like any kind of habit-building, you’ll never use data successfully if you treat it as an all-or-nothing proposition. First use data at the highest leverage point and continue to expand.
  5. People doing creative work together are going to have tension. Make sure that tension is healthy, productive and about the work, not about the people. It will feel like it’s about the people a lot.
  6. Try to create egoless ideation, where folks separate their self-worth from their ideas and are quick to support other’s (good) ideas. Strong opinions, weakly held.
  7. Our natural inclination is to assume our past success is because of our past actions. This can create very bad product dogma if left unchecked. What were the right things to do in the past might not be the right things to do in the present.
  8. Even if past actions did lead to success, remember that everything else has changed. New platforms, new users, new dynamics. Re-apply your strengths in your new environment.
  9. Find the one thing you need to get right and spend most of your time on it. You can screw up basically everything else.
  10. You can A/B test individuals, but it’s nearly impossible to A/B test communities because they work based on a mutually reinforcing self-conception. Use a combination of intuition (which comes from experience), talking to other community managers and 1:1 contact with a sample of your community. But you’ll still be wrong a lot.
  11. Public forums are not a useful way to get feedback, but are a useful way to get buy in from the community. This is true of both online and offline communities.
  12. Communities are unpredictable. Don’t take the community’s criticism too personally or you’ll become afraid of change and slow. Instead, be open, thoughtful and move quickly.
  13. Look at any product of sufficient complexity and there will be a ten things you’ll want to say “if only they moved this button here…” and there will be a very good reason you’re wrong. Sometimes you’ll be wrong because of implications you haven’t considered, but sometimes you’ll be wrong because there was a product evolution that made perfect sense every step of the way. Separate the two. Don’t get bogged down in the history of the second kind when making new decisions.
  14. Use your first few weeks at a new company to understand your product’s history and match it against reality. Ask questions. Write everything down. Don’t imply incompetence. Don’t make rash judgements.
  15. Some people want to see the big picture, some people want to see just the next step. Try to provide both, and connect the two clearly.
  16. Most partnerships are a waste of time. All partnerships take a lot of time.