The daily routine and the five indicators

One day, four moments

A healthy operation is boring. The same actions, at the same times, in the same order. Here is the shortest form that works.

Morning: reread the night. Open the conversations that moved while you were away. Fix what needs fixing, take over where necessary, follow up on what was left unanswered. This is the most profitable moment of the day because night conversations are often the most engaged.

Midday: publish. The acquisition from module 4. Fixed volume, every day, no exception. Consistency beats intensity.

Late afternoon: the sample. Five to ten conversations picked at random, read in full. You are not looking for a specific error, you are looking for a drift: a price that slipped, a persona that changed register, a promise that was made. The industry’s job listings show that this work exists as a position in its own right, paid separately, to “audit live chatter logs”. In a small operation, that is you, and it takes fifteen minutes.

Evening: one line of numbers. The five indicators below, noted in the spreadsheet. No analysis, no conclusion, just the line. Analysis happens once a week.

The five indicators

1. First conversations. The number of people who wrote a first message. This is the exit of the module 4 funnel and the entry of everything else. If this number drops, nothing else can compensate.

2. Unlock ratio. The number of paid media opened per hundred sent, tracked per tier and not as an average. This is the selling metric defined in module 5.

3. Revenue per conversation. Net revenue for the period divided by the number of active conversations. It is always tracked together with indicator 2, because lowering prices pushes the unlock ratio up and this number down. Use net revenue, after the platform commission, which is 20% of the fan’s total payment on OnlyFans.

4. First reply delay. The time between a fan’s message and the reply. It is the only indicator that gets fixed faster than it degrades, and it is almost always the first to repair.

5. Conversation cost. The total cost of conversation divided by the number of conversations. The calculation is the one from module 5: pay and commissions, or subscription and tool commissions, plus supervision time.

Note what is not on the list: subscriber count, view count, number of messages sent, follower count. These numbers describe an activity, not a result, and none of them leads to a decision.

The order of correction

When several indicators are bad, there is an order, and it is not negotiable. Fixing things out of order means optimizing something the next step will invalidate.

First, the first reply delay. As long as it is bad, all the other numbers are wrong, because you are measuring conversations you have already lost. It is also the fastest indicator to fix, through coverage or through automation.

Then, the unlock ratio. Tier by tier. A weak tier 1 points to the teaser and the opening; a weak tier 2 while tier 1 is good points to the price or the pace. Those are the tests in the lesson Versions per audience and testing method.

Then, the price. Once the script is stable, and only then, move tier 2. Never before: a price test on an unstable script measures nothing.

Finally, acquisition. This is the slowest and most expensive correction. It should come only once the existing conversations are properly worked. Increasing traffic upstream of a conversation that does not convert is the most efficient way to lose money.

Three numbers that look useful and are not

Total subscriber count. It goes up when you add an acquisition account and down when you lose one, which tells you about your stock of accounts and not about your activity. Worse, it counts in the same way a subscriber who never opened a message and a subscriber who buys every week.

Number of messages sent. It measures effort, not result. An operation that doubles its messages sent and keeps its revenue flat has just halved its efficiency, and the indicator will show an improvement.

Monthly revenue on its own. It mixes everything: a price change, an acquisition spike, a big buyer who happened to pass by. Read on its own, it produces confident and wrong explanations. Read next to first conversations and the unlock ratio, it becomes useful.

The rule behind these three cases: keep a number only if a bad value tells you what to do next. If an indicator drops and you cannot name the action that follows, it is not an indicator, it is decoration.

The weekly review

Once a week, thirty minutes, three questions.

What moved? Compare the week to the three previous ones, never to the single week before.

Which test is running, and where does it stand? One variable, one written stopping rule, as in module 5.

What do I fix next week? One thing. An operation that fixes three things at once will never know which one worked.

This meeting with yourself is what separates an operation that progresses from an operation that fidgets. It fits in a text file, and it is more useful than any dashboard.