She was standing in her sister's kitchen at half past eleven at night, on the second evening of a four day family visit, answering a billing question on her phone. Her sister asked if everything was okay. She said yes, this will take one minute, and then spent another thirty five minutes replying to eleven other things because once she had opened the app she could not stop.
That was the night she started drafting the message that begins "I have made a difficult decision about the future of our community."
She did not send it. What happened instead took about a year, cost her real money, felt worse before it felt better, and is the reason she is still doing this work. I want to walk through it properly, because delegating for the first time is one of the loneliest decisions in this business and almost nobody describes the middle part honestly.
This is a composite, stitched together from a pattern I have watched play out in a number of small memberships. The details are generalised on purpose. The shape is the useful bit.
Where she was
Roughly three hundred members. A community for people in a specialised professional field, the kind where the members are competent adults with demanding jobs and very specific questions. Four years old. Profitable, which is worth saying, because this was not a failing business. By every external measure it was working.
She was doing all of it. Every welcome, every question, every billing hiccup, the monthly call, the weekly post, the resource index, the tax filing, the sales page, and a rolling background process of noticing which members had gone quiet.
The number that mattered was not revenue or churn. It was that she had not had a day without opening the community in fourteen months, including the day of a funeral.
The thing that actually broke
Not her health, though it was heading that way. What broke was her judgement.
She started reading ordinary messages as attacks. A member asking whether the call time could move became, in her head, an accusation of laziness. She drafted a reply that was cold, sat on it, and later could not understand why she had written it that way. She snapped at a newcomer for asking something answered in the welcome post, then spent the evening feeling sick about it.
This is the real signal of being past your limit, and it arrives before exhaustion does. Not tiredness. A loss of generosity. When the work you love starts producing irritation instead of warmth, you are not being difficult, you are over capacity, and no amount of better systems will fix a capacity problem.
What she tried first, which did not work
She tried the three things everybody tries.
She tried batching. All community work in two blocks a week. It lasted nine days and collapsed, because a billing failure does not wait for Thursday and she could not stand the thought of a newcomer's introduction sitting cold for four days.
She tried cutting delivery. She cancelled the weekly post. Nobody complained, which stung slightly, and it bought her about ninety minutes a week. Real, but not the problem.
She tried a stricter boundary announcement. She posted that she would reply within two business days rather than immediately. Members were completely fine with it. She still checked at midnight, because the issue was never the members' expectations. It was hers.
That is the useful discovery in this whole story. She did not have a systems problem or an expectations problem. She had a single-person problem, and the only fix for a single-person problem is a second person.
Who she asked
Not the most enthusiastic member. This is the instinct and it is usually wrong.
She asked the member who had been quietly answering beginner questions for two years, accurately, without ever making anybody feel stupid, and who had never once tried to build a following inside the community. Mid volume. Reliable. Slightly formal. Not the person she would have picked as a friend, which turned out to be an advantage.
She asked in a private message that was specific from the first line. Not "would you like to help out." Something closer to: here are four specific things I need someone to own, here is roughly how many hours a week I think they take, here is what I can pay, would you like to talk about it.
The specificity mattered enormously. Vague requests for help get vague yeses that quietly evaporate. A defined job gets a real answer, and if the answer is no you learn quickly and can ask somebody else without either of you feeling awkward.
The arrangement, on paper
They wrote a one page document. It took forty minutes and it is the single highest value artefact in the whole story.
It named four owned tasks: welcoming every new member within a day, first response to questions in the main space, keeping the resource index tidy, and flagging anything that looked like a conflict rather than handling it.
It named a rough time commitment: around five hours a week, with an explicit line saying that if it consistently ran over, they would revisit the arrangement rather than quietly absorb it.
It named money. She paid an hourly rate, monthly, as a contractor, and looked up what that meant for her own tax paperwork before the first payment rather than after. Not a free membership. Not "we will figure it out." An actual rate, paid on a date.
It named the limits of authority. The helper could nudge publicly and message privately. Warnings, pauses, and removals stayed with the founder. Nothing about pricing, refunds, or anything financial.
And it named the access. The helper got her own account with her own credentials and only the permissions the job required. No shared login, not once, not for an afternoon. No access to payment details, which is a place a helper should never need to go. She wrote down, explicitly, that private member messages and anything a member had shared in confidence were not to be forwarded, screenshotted, or discussed outside the two of them.
That last paragraph is the part people skip, and it is the part that protects everybody. Members had joined a private space. Adding a second person to that space is a real change in who can see their words, and doing it carelessly is a breach of the thing they actually paid for.
The first six weeks were worse
Here is the part nobody warns you about.
She spent more time, not less. She reviewed replies she had not been asked to review. She rewrote a welcome message that was perfectly good because it was not in her voice. She answered a question forty minutes before the helper would have got to it, then apologised for it, twice.
She also felt strange. Slightly displaced. There is a specific grief in watching somebody else do the thing that had become your entire identity, and it does not help to pretend it is not there. A member thanked the helper for something warm and she felt a small unattractive pang, and then felt worse about having felt it.
If you are in week three of a handoff and thinking this was a mistake, this is normal. The cost front loads and the benefit lags by about two months.
What changed at week eight
Two things flipped at roughly the same time.
The helper started doing things she would not have thought of. Reorganised the resource index by problem rather than by date. Started a small monthly thread that turned out to be popular. That was the moment the arrangement stopped being a copy of her and became an addition to the place.
And she went away for a weekend and did not check. Not heroically. She just forgot, for about a day and a half, and when she remembered, nothing was on fire and eleven new members had been welcomed properly.
What it cost, honestly
Money. A meaningful monthly amount that came straight out of what she used to take home, and for the first quarter she felt it.
But look at what it bought against the alternative. She had drafted a shutdown notice. The comparison is not "spend money versus keep the money." It is "spend money versus have no business at all," and framed that way it stopped being a difficult decision.
By the end of the year the community had grown modestly, mostly because welcoming people properly turns out to matter more than any marketing, and the extra revenue covered a decent chunk of the cost. That is a pleasant outcome rather than a promise. It might not have. She would still have been right to do it.
What she would do differently
She would have asked a year earlier. Every founder says this.
She would have started with two tasks instead of four, because the review burden in the early weeks scales with surface area.
She would have written down her own tone in advance. A short page of examples: here is how I welcome somebody, here is how I decline a request, here is a phrase I never use. She ended up writing it in week five, after correcting the same thing three times. Written first, it would have saved both of them a lot of small friction.
And she would have told the community sooner and more warmly. She introduced the helper in a slightly apologetic sentence, as if bringing in support was an admission of something. Members were delighted. They had been able to see how thin she was stretched for months.
The transferable part
You do not need three hundred members or a specialised niche for any of this to apply. What generalises is the sequence.
Notice the loss of generosity, not the tiredness. Accept that no system fixes a single-person problem. Pick reliable over enthusiastic. Write the one page. Pay properly. Separate accounts, never shared credentials, minimum necessary access. Expect six weeks of feeling worse. Then let them change something without correcting it.
Do this this week
Write down the four things you do every week that absolutely do not require you specifically. Be honest. Welcoming people is almost never on the list of things only you can do, though it feels like it is.
Name one person who is reliable rather than loud.
Draft the one page. Tasks, hours, money, authority limits, access limits. Forty minutes, no more.
And separately, whether or not you are ready to hand anything over, spend one evening this week not opening the community at all. Just to see what it feels like. That is where she started, in a kitchen, at half past eleven, and the good news is that the next chapter of that story is a much better one than the message she almost sent.