What would someone smart — like Elon Musk or Jeff Bezos or Mark Zuckerberg or even Seth Godin — do in this situation? Would they quit and say, fuck it, this is never going to work? Would they raise prices? Would they start a product?
Or would they do a little bit of raising prices combined with releasing a product? Or would they say, let’s not do a few of these things — let’s go deeper in these non-commoditized things and get a couple of dozen clients, giving them newsletters with deep analytics? What would they do?
Let me cut straight through your list, because the way it’s laid out hides the real answer.
None of those four would walk away from a 20-year asset the week before fixing the thing that’s depressing its value. Covered that already.
Repricing is not a strategy, it’s hygiene. It’s correcting an error. Bezos wouldn’t dignify “we finally charged what the work is worth” as an answer to “what would you do.” Do it this week, bank the 30%, and don’t mistake the housekeeping for the plan.
This is the one option all four would actively reject, and Zuckerberg most violently. A product built at 20% effort dies. His entire operating signature is the opposite: pick one bet and put the whole weight of the company behind it. “A bit of everything” is how you do all of it at 60% and win at nothing.
So the real question is your option 5 versus a product. And here’s what the smart operators would see that the list obscures: those two are not a fork. Option 5, followed all the way down, is the product.
Look at what you actually said in option 5 — “go deeper in the non-commoditized things.” That instinct is correct. It’s pure Seth Godin: don’t be a commodity, find the smallest viable audience, be remarkable to them. And option 5 is smarter than it looks for a reason you may not have fully clocked — newsletters are the one service line your own constraint document says already runs without Arseni, because there’s an SOP and a repeatable process.
So option 5 isn’t “more clients like David Drebin.” It’s “more clients of the one thing that’s already a working system.” That’s genuinely the best agency-side idea in your list.
A newsletter is a commodity — Mailchimp, every agency, table stakes. Analytics reporting is a commodity too — every email platform has it. The only genuinely remarkable, non-commoditized thing in option 5 is the intelligence: telling a gallery “call this collector this week, show them this piece, here’s why.”
And here’s the trap and the key in one sentence: intelligence delivered by people is a service with a ceiling; the exact same intelligence delivered by a system is a product. The deeper you take option 5’s own instinct, the more it stops being a newsletter service and becomes software. Option 3 isn’t a different choice from option 5 — option 3 is just option 5 followed to its logical end.
Which means the whole decision collapses to one design choice, and it’s the choice that separates a smart operator from a busy one. You are going to serve those couple-dozen clients. The only question is what you scale with.
Hire more designers, run a bigger newsletter shop — and you have built a nicer, larger version of the exact cage you’re in now: costs rising with revenue, you still the quality gate.
Build the newsletter-and-analytics intelligence as a system, and have people operate the system rather than do the work — and the same two dozen clients now fund you and hand you a sellable product. Same option 5. Two completely different destinies. One design decision.
That tool-first version is what all four would actually do — and it’s the same move under four different names: Bezos calls it AWS (infrastructure you built to do the work becomes worth more than the work). Musk calls it the algorithm (option 5 even starts correctly, by deleting the commodity lines before scaling). Zuckerberg calls it focus (one bet, go deep and narrow first). Godin calls it refusing to be a commodity. They all point at the same thing: serve the two dozen clients with a system, not a staff.
Bank the 30%. It’s just error-correction.
The criterion is brutally simple: shortest path to ten paying customers with the least founder-dependence. This is where option 5 genuinely earns its place — you have roughly two dozen past newsletter clients, a warm base, proven demand. Upselling them an intelligence-grade newsletter product is a faster, warmer path to first revenue than iBrain’s colder market of other agencies.
That’s a real argument, and it’s worth saying plainly: option 5 may have just identified a better first product than the one I’ve been pushing. Build that one, tool-first. Sell its output as the managed service to the two dozen — that is your cash and your proof at the same time. Then the tool becomes the thing you sell.
iBrain doesn’t disappear; it becomes the second arrow, later — never a parallel one.
Which answer is “right” depends on what Arseni actually wants — and that connects straight to the hostage conversation.
The labor version of option 5 can reach $20K/month and a calmer life; if that is the real goal, it’s a legitimate, lower-risk choice and nobody should sneer at it. But it does not make the business sellable and it does not end the dependency — it just builds a roomier cage.
Only the tool-built version delivers the thing you were actually circling last time: freedom, and an asset you could one day walk away from with money in hand. The four operators would all choose that one — but they were all empire-builders. You’re allowed to choose your own size.
What you’re not allowed to do, if you want out, is the labor version and call it the exit.