The Rejection Exercise

For years I’ve been making founders do 30 days of rejection, where they purposely go out and try to get rejected at least once a day. The purpose is to toughen yourself up so you’re not broken by hearing customers / investors / collaborators say “no.”

That’s the general format, but in recent years, I’ve done something different.

1) I do the exercise along with them.
2) I use AI to analyze the results (the recent group of 34 founders produced almost 1000 results).

These are the exact steps I used for this one.

1. Provide detailed instructions for how to pursue the rejections, including the document format to save them in.

Rejection Exercise
Rules:
– You must be rejected in person (not online, by phone or remotely) by another person at least once, every day for the next 30 days. Don’t just go through the motions on this. See how you can top yourself each day.
– A rejection counts if you are out of your comfort zone.
– A rejection counts if your request is denied.
– Try different types of rejections, not the same one again and again.
– Do not pursue ridiculous rejections, like “Can I have a trillion dollars?”
– At the time of rejection, you, not the respondent, should be in a position of vulnerability. You should be sensitive to the feelings of the person being asked.
– For this assignment you cannot involve any others in the group, including me!
– Fill out the rejection exercise template.

2. After 30 days, download the results.

3) The AI analysis.
Prompt: Read the attached zipped folder. Analyze the numbered (1 – 30) lists in the docs. Look at columns titled: “What you said”, which is what the people said to get a rejection and “Their response”, which is what the person they spoke to responded with. Make a summary of these exchanges. What patterns do you notice from the exchanges? Which people had the most extreme rejections? Add other things you notice that make sense in this context.

4) What to do differently next time.
– Provide analysis summary beforehand to inspire bigger asks and for context.
– Put instructions in template document (10% changed the filetype, which caused problems.)
– Reward the biggest asks and best rejections.

5) Output: Success rate for different types of asks