How to stop working on the wrong things as a founder

Executive overview

Most founders mistake busyness for effectiveness — coding all day, tweaking designs, doing support — precisely because these tasks feel comfortable. The certainty vs. uncertainty framework fixes this: do work where the outcome is uncertain, delegate work where the outcome is certain.

Uncertain work creates value. Certain work can be handed off.

Uncertain vs. certain work

• Uncertain work (do it yourself): customer development calls, positioning, messaging, new marketing channel tests, product roadmap decisions, finding product-market fit
• Certain work (delegate): customer support, social media posting with templates, admin, specced-out coding tasks, email management, data entry
• Uncertain work always becomes certain over time — once you find a repeatable pattern, templatize and delegate it
• Your job is to fire yourself from certain jobs one at a time and move to new uncertain work

Caveats and when to hold off

• Do your own customer support for the first 50–100 customers to learn patterns
• If you can't afford delegation yet, make it the goal — not the permanent state
• Delegating doesn't mean ignoring — monitor quality and watch for product signals in support issues

Overcoming delegation objections

• "Nobody does it as well as me" — you're not delegating your best work, you're delegating specced-out execution; save your expertise for uncertain problems
• "Training takes too long" — a few hours of screencasts and an AI-generated knowledge base recovers 10–40 hours a month
• "I tried it and it failed" — check whether you delegated uncertain work or gave insufficient context; delegation is not abdication
• If you can't document a task well enough to train someone, you don't understand it well enough yourself

The calendar audit exercise

• Open your calendar and go through last week hour by hour
• Mark each block R (right — uncertain outcome) or W (wrong — certain outcome)
• Most founders find 50–70% of blocks are W
• Pick one W, look up real rates on Upwork, and calculate whether the cost is worth recovering 10–40 hours per month