A founder of a 19-person services firm at his Friday review with the operations director. Three weeks ago he told her she could own customer onboarding decisions for thirty days. He has reviewed every onboarding decision she has made. He has asked clarifying questions on most. He has overruled her on two of them.
She has noticed. The trial period ends in nine days. She is preparing to hand the work back. He thinks the experiment has failed because she has not stepped up. The experiment has failed because he has not let her.
Why does informal delegation almost always fail?
Most owner-led delegation is informal. The founder says “you can run this for a while“ without specifying what success looks like, what alarm thresholds should trigger an escalation, when the review will happen, or what the audit will measure. The role holder takes the work and proceeds with their best guess.
The founder remains in the loop through informal channels and intervenes when something looks wrong, often before the role holder has had time to surface the issue. At the end of the unspecified period, the founder concludes either that the role holder is not ready or that the work cannot be delegated. Both conclusions are usually wrong. The actual problem is that the experiment was not designed; the founder was running an audit and the role holder was running a job, and the two were not the same job. The 30-day structure exists to align them.
The framework
Pick a single category of work currently owned by the founder. Choose something important but not the most critical work in the firm. Name a single role holder who will own it for thirty days. Define what success looks like and what failure looks like, written down. Define alarm criteria: what red flags should trigger an immediate escalation.
Schedule a fifteen-minute check-in at day five (clarity confirmation, not intervention) and another at day fifteen (mid-point pulse). Plan a full review at day 30. The structure is non-negotiable. Without success criteria written down, the day-30 review becomes subjective. Without scheduled check-ins, the role holder is alone with the work and the founder is alone with their anxiety; both lose. With the structure, both parties know what they are doing and the experiment produces real data regardless of whether the outcome is good or bad.
Good candidates for the first experiment
Customer onboarding decisions: who gets accepted as a customer, on what terms. Financial reporting: the founder receives a standard summary monthly rather than reviewing raw data weekly. Hiring of individual contributors within a defined team and budget. Vendor relationships: the operations manager owns the relationship with main vendors. Proposal approval: the founder approves proposals over a certain value, and below that threshold the team decides.
The first experiment is deliberately not the most critical work. It is important enough to matter but not so critical that failure is catastrophic. The pattern that works: pick something the founder spends three to five hours a week on, where the team has the relevant knowledge, where the boundary conditions are easy to write down, and where a 70 percent outcome is acceptable for the first cycle. After the first experiment runs cleanly, more critical categories become possible.
The explicit handover paragraph
The role holder is told explicitly: “I am delegating this category to you for thirty days. I trust you to own this. Here are the success criteria. Here are the alarm thresholds. I will check in briefly at day five and day fifteen, and we will do a full review at day thirty. Please do not wait for me to ask how it is going; if an alarm threshold is triggered, escalate it immediately.“
The text matters. It is the contract; it removes the ambiguity that defeats most delegation attempts. The role holder reads it twice and confirms they understand. The founder commits to it in writing. Once the contract is in place, the role holder has clean authority for 30 days and the founder has clean criteria for the audit. The relationship is operational, not informal.
The four audit criteria at day 30
Did the role holder get it right? The work met quality and timeliness standards. Did they ask for help when they needed it? Escalating appropriately or suffering in silence. Did they raise the right red flags? Problems flagged early or surprises at month-end. What would the founder have done differently? How much divergence between the role holder’s approach and the founder’s preferred approach. The audit converts subjective impression into specific feedback.
The fourth criterion is where the cognitive trap lives. The founder will inevitably find places where the role holder did things the founder would have done differently. The audit measures whether the outcome meets the success criteria the founder defined upfront, not whether the role holder did the work the way the founder would have. The role holder’s divergent approach might have produced a better outcome, the same outcome, or a worse one; the audit treats the outcome as primary.
The four common failure points
Under-defined success criteria. The founder and the role holder have different ideas of what success looks like, and at day 30 they discover they were playing different games. The antidote: spend time upfront ensuring success criteria are crystal clear, agreed in writing with the role holder before day 1.
Founder shadow-checking. The founder says they are not intervening but actually reviews everything the role holder does and subtly corrects things, which the role holder perceives as interference and gives up on genuine autonomy. The antidote: commit explicitly to not reviewing or correcting work in progress.
Role holder under-resourced. The role holder is given the task but not the information, authority, or time required to do it well, and the experiment fails because of lack of support, not lack of capability. The antidote: do a resource review before day one. Does this person have access to all the information and people they need? Does this person have the authority they need to make decisions, or do they have to ask for approval anyway?
Time horizon too short. Thirty days is sometimes too short to see the full outcome of a decision category. The role holder might do fine with day-to-day decisions but miss the longer-term implications of a contract negotiation or customer selection. The antidote: for some categories (hiring, contract negotiation, strategic vendor relationships), extend the experiment to 60 or 90 days so the long-term outcomes can be assessed.
The first experiment as the prototype
The first 30-day experiment is the prototype. Once it has run cleanly, the next category is easier. The founder has built the habit of writing success criteria, of resisting the urge to shadow-check, of using the day-5 and day-15 check-ins as listening rather than redirecting moments. The role holder has built confidence that the experiment is real, the authority is real, and the audit is fair.
After three or four cycles the founder has built the habit of structured delegation. After eight or ten cycles, the operational layer of the firm has been redrawn. The experiment is also a diagnostic for the founder. A founder who cannot resist shadow-checking the first experiment is not yet ready for larger delegation. The experiment surfaces the gap. The fix is not to run a more critical experiment; the fix is to run more cycles of small experiments until the discipline of not intervening has actually been built.
What to do this month
Pick one category of work you spend three to five hours a week on and where you trust the role holder enough to attempt the experiment. Block 30 minutes to write the success criteria, the alarm thresholds, and the day-5, day-15, and day-30 plan. Send the explicit handover paragraph to the role holder. Confirm they have read it and that they understand the criteria.
Then do not intervene for the next 30 days. Resist the urge to check in informally between the scheduled check-ins. At day 30, run the four-criteria audit honestly. Whatever you find is the data; the data informs whether to extend the delegation, run another cycle, or redesign the role.
If you would like a second pair of eyes on the success criteria for your first experiment, book a conversation.
