Politely disagreeing with the 'stay manual as long as possible' line above, because it depends what you are selling. If the value is judgement, manual teaches you the product. If the value is speed and self-serve, manual teaches you very little and actively hides the failure mode: people will happily wait four hours for a human and will not wait four seconds for software. I ran a concierge version of a scheduling tool for two months, everyone was delighted, and the automated version got taken apart on exactly the dimension the manual one was papering over.
This. I delegated the floor and immediately filled the space with admin I could have batched into one Friday morning. Nobody asked me to do it at 11pm, I just did.
This is right but the churn assumption is doing a lot of work. 5% monthly on a self-serve $19 tool is optimistic-to-normal depending on who you sell to. Freelancers churn when their project ends and that is not your fault and you cannot fix it with features. Ask yourself whether the job your tool does is recurring for the customer or one-off, because that decides your churn floor more than anything you build.
What I did wrong was treat the tiredness as a discipline problem. Caffeine from six in the morning, a five-day gym programme to 'build resilience', and I crashed harder in October than I had in the spring. Adding load to a system that's already failing to recover just moves the collapse. When I finally let myself do genuinely nothing for six weeks, sleep started changing within about a month.
The tool payback maths is more favourable than people expect, and it is mostly one job. A chain tool, a cassette lockring tool, a chain whip and a gauge is a small one-off spend, and replacing a chain yourself takes about fifteen minutes once you have done it twice. If your shop charges labour for a chain swap three times a year, the tools pay for themselves inside the first year on that single job. Tyres and tubes need no tools at all beyond levers. Where I still pay a shop is anything involving bearings and anything hydraulic, because the tools there are genuinely expensive and the jobs are rare.
Money was maybe 60 percent of it. The other 40 is that I've asked for a specific project three times and been told "next quarter" three times. Reading this back, that's the answer isn't it.
The six-month review is a real lever and it is also the one most likely to evaporate when your manager changes. Get it in the offer letter or treat it as a nice sentence.
I white-knuckled it for fourteen months on exactly your reasoning, no savings, can't stop, keep going. Then I got signed off for three months anyway, and the three months I 'couldn't afford' happened at the worst possible time and on someone else's terms. I'm not saying stop tomorrow. I'm saying the maths of I-can't-afford-a-break assumes you get to choose when the break happens, and I didn't.
My expensive lesson was rolling what I still owed on the old car into the new loan because the monthly number still looked fine. It looked fine because the term was longer. I was underwater on that car for years and could not sell it when I needed to. If anyone is doing arithmetic that keeps your payment the same while the total goes up, that is the moment to slow down. I am not a finance person, and honestly my credit union talked me through it far better than the dealer did.
I bought a shop on the seller's word and the declared takings were about thirty percent above what the statements showed once I finally saw them. My fault entirely - I liked the owner and I didn't want to look suspicious. It cost me two years of overpaying on a loan for goodwill that wasn't there. Being unpleasantly thorough is not an insult, and any decent seller expects it.
Politely disagreeing with the level-not-salary advice, at least as a first move. Re-levelling usually means going back through a committee, which can take weeks and occasionally reopens things you thought were settled. If you like the team and the gap is a few thousand, ask for a signing amount to bridge it and a written review at six months. Faster, lower risk, and the compounding difference is smaller than people assume at this level.
Don't drop a client without a date and a number written down first. You'll fill the free time with product work that doesn't compound and then panic in month three.
Volunteer for the thing nobody wants. Fastest route out of the flat phase I know, because it puts you in rooms you'd otherwise wait a year for and people remember who put their hand up.