Solo UnicornHow I Run Three Businesses Solo — My Time Management System
How I Run Three Businesses Solo — My Time Management System
Last week, a community member asked me: "How do you manage three businesses by yourself? I can barely handle one."
My first instinct was to say "I'm barely managing either," but that's not the most honest answer. The truth is: I am indeed running ArkTop AI, JewelFlow, and the Solo Unicorn Club simultaneously, with zero full-time employees. But the reason it works isn't because I'm more disciplined or hardworking than anyone else — it's because I designed a time system specifically built to sustain this kind of operation.
This article lays out that system in full. The framework, the tools, the real numbers, and the mistakes I made along the way.
2024: Everything Was Competing for Attention
Let me start with context so you know where I started.
By mid-2024, ArkTop AI was still in its early days. Client projects came in one after another, and I was essentially in a "take project — deliver — take next project" loop. JewelFlow was just getting started, eating up huge chunks of time each week for product discussions and technical implementation. The club had just launched, member messages were flooding in daily, and I was replying to nearly every single one by hand.
Those months, my calendar was a genuine disaster. On any given day, I might have an ArkTop client call, a JewelFlow bug fix, and club topic curation all stacked together. The three businesses constantly interrupted each other, and nothing was ever finished in a state of focus.
I remember one Wednesday afternoon: I was writing ArkTop's quarterly report, got pulled away to fix a JewelFlow API issue, then went to reply to thirty-something club messages. All three things got done — but the report was rough, the bug fix left an edge case I'd later discover, and some genuinely valuable questions in the club got half-hearted responses.
That evening, I thought about it for a long time. The problem wasn't that I didn't have enough time. The problem was that my time had no structure.
The Time Allocation Framework: Business Isolation
I spent about three weeks designing the framework I use today. The core principle is one single rule: different businesses cannot compete for my attention on the same day.
Weekly Calendar Architecture
I divide my week into three types of workdays:
ArkTop AI Days (Monday, Thursday) Client communication, solution design, deliverable reviews — all concentrated on these two days. On these days, I mute notifications from my other businesses entirely. Once I sit down to work on ArkTop, my attention is fully locked in.
JewelFlow Days (Tuesday, Friday morning) Product decisions, feature planning, tech debt assessment happen here. Friday afternoons I reserve for code review — the quietest block of my week, perfect for tasks that need sustained concentration.
Club + Content Day (Wednesday) Community operations, content creation, external communications — all on Wednesday. On this day, I essentially don't touch the other two businesses.
Saturday, I spend 2–3 hours on a weekly review — no new project work. I check each business's status, update task priorities, and anticipate potential blockers for the coming week. Sunday is completely off.
This framework sounds rigid. In practice, I can stick to it about 75% of the time. The remaining 25% gets disrupted by client emergencies or production issues. I accept that ratio. The point of the framework isn't zero exceptions — it's making exceptions the exception, not the norm.
The First 90 Minutes
Every workday from 9:00 to 10:30 AM, I take no meetings and answer no messages. Those 90 minutes are reserved for the single most important task of the day.
On an ArkTop day, that might be drafting a proposal. On a JewelFlow day, it might be solving a design problem. On a club day, it might be outlining an article. Whatever it is, those 90 minutes go to one thing only.
I've tried "making use of scattered time" many times. The conclusion is always the same: scattered time is fine for execution tasks, not for thinking tasks. Reserve your best block of the day for the task that requires the most judgment, and everything else becomes more efficient.
AI-Delegated Scheduling: Removing Execution Tasks From My Calendar
The time framework solved "where does my attention go," but there's still a huge volume of execution tasks that need handling. This is where AI systems come in.
I categorized the execution tasks across all three businesses. The breakdown looks roughly like this:
ArkTop AI
- Client data report generation: runs automatically each week, I review for 15 minutes
- Meeting notes: recording transcription → Claude API extracts key points → pushed to Notion
- Competitive intelligence tracking: custom RSS + summary Agent, sends weekly digest to my email
JewelFlow
- User feedback aggregation: form data → classification Agent → sorted by priority
- Error log monitoring: Railway alerts connected to n8n, automatically determines whether I need to step in
- Weekly report generation: key metrics auto-pulled, formatted, and sent to me
Solo Unicorn Club
- New member onboarding: automated questionnaire + welcome message, fully automated
- Content scheduling: weekly topic prompts sent on a timer
- Highlights archiving: valuable discussions automatically organized into the knowledge base
Combined, these remove roughly 18–20 hours of execution time from my calendar each week. It's not that these tasks don't matter — it's that they don't need me to do them.
Here's how my actual weekly time breaks down now:
| Business | Weekly Focus Time | Primary Activities |
|---|---|---|
| ArkTop AI | 14–16 hours | Client communication, solution design, delivery review |
| JewelFlow | 12–14 hours | Product decisions, code review, technical planning |
| Solo Unicorn Club | 6–8 hours | Content creation, community judgment calls, external comms |
| Weekly review | 2–3 hours | Status sync, priority rebalancing |
Total: roughly 35–41 hours per week. Lower than my 2024 numbers. Back then, execution tasks consumed enormous amounts of time, and execution tasks have no natural boundary — they only multiply.
Tool Stack
Tools chosen specifically for time management:
| Use Case | Tool | Monthly Cost | Why |
|---|---|---|---|
| Calendar management | Google Calendar + Calendly | $0 / $12 | Color-coded for business isolation, Calendly controls meeting slots |
| Task management | Linear (ArkTop/JewelFlow) + Notion (Club) | $8 / $16 | Two separate systems by design — prevents tasks from bleeding across |
| Time tracking | Toggl Track | $0 | Monthly audit of actual time allocation, catches drift |
| Meeting notes | Fireflies.ai | $19 | Transcription + key point extraction, feeds into Claude API for further processing |
| Workflow automation | n8n (self-hosted) | $5 | Central hub connecting execution tasks across all businesses |
| Attention management | Opal | $9.99 | App blocker — on ArkTop days, club notifications are locked out |
A note on Toggl: I initially thought time tracking was a vanity exercise. After three months, I changed my mind. The gap between actual time data and my subjective perception was striking. One month I was convinced I'd spent a lot of time on JewelFlow. Toggl showed that 40% of my "JewelFlow time" had actually gone to handling miscellaneous ArkTop issues. That kind of drift, left undetected, compounds into serious problems.
Lessons From Mistakes
Mistake #1: Turning "business days" into "standby days"
In the first few weeks of the framework, I'd still instinctively check JewelFlow's Slack on ArkTop days. The moment I saw a message, I'd think "it'll only take a second to reply." That "second" broke whatever I was writing, and getting back into flow cost me another 15 minutes.
The fix was using Opal for hard isolation — on ArkTop days, JewelFlow's workspace is locked behind a password. That friction of "needing a password" is just enough to make me pause and ask whether the thing is actually urgent. 90% of the time, it isn't.
Mistake #2: Letting the weekly review become another workday
I scheduled my weekly review on Saturday. It kept turning into "catching up on what Friday didn't finish." The review became an excuse to extend work hours, and the reflection and planning that should have happened never did.
My adjustment: the Saturday review answers only three questions, written on paper, finished in 15 minutes:
- What was the most important progress in each business this week?
- What went wrong, and what do I change next week?
- What is the single most important thing for each business next week?
Three questions, three businesses, nine items max. Write them down, then stop.
Mistake #3: AI automation created new time debt
After building the automation system, I went through a phase of "the work is done" and let my guard down on quality checks. One auto-generated ArkTop report had a data methodology error that I didn't catch before sending it to the client. The client noticed and emailed me. Handling the fallout cost me half a day.
The time savings are real, but they depend on maintaining the habit of review — not eliminating it. My rule now: any AI-generated output that goes external, I spend a fixed amount of time reviewing it, no matter how much I trust the Agent.
Advice for People Considering Multiple Businesses
If you're running a single business right now, you don't need a system this elaborate. But if you're thinking about a second or third, here are the three things I consider most important:
1. Figure out where your time actually goes right now.
Before adding a business, spend two weeks tracking your current time allocation. Many people feel like they don't have enough time, but in reality, huge amounts of it leak into things they're not even aware of — random message replies, frequent task-switching, doing things that "feel useful but produce no direct output." Until those problems are solved, a second business will only make things worse, not better.
2. Only add a business where AI can take over at least 50% of the execution work.
This isn't about AI replacing people. It's about this: if a new business's execution tasks require you to handle them entirely yourself, it will directly squeeze your time for your other businesses. ArkTop AI's report generation, JewelFlow's user data processing, the club's onboarding flow — all of these are fully automated. If 80% of a business's tasks need your hands on them, it's not suited to be your second business — unless you're prepared to hire someone for it.
3. Design the boundaries between businesses as physical barriers, not willpower challenges.
I don't rely on self-discipline to maintain business isolation. I enforce it with tools. Separate task management systems, color-coded calendars, Opal's app blocker — these all turn "don't switch" into a physical constraint. Relying on willpower to maintain attention boundaries is unsustainable long-term. Environment design beats willpower every time.
Where Things Stand Now
March 2026: three businesses running simultaneously, no full-time employees, weekly work hours stable under 40.
This isn't the optimal solution, and it isn't the end state. If JewelFlow's user base keeps growing, I estimate I'll need to bring on a first technical hire in the second half of this year — otherwise tech debt will start dragging down product velocity. If ArkTop needs to serve more concurrent clients, the reporting system will need an upgrade too.
A solo company doesn't mean one person forever. It means using systems to replace headcount for as long as possible. Right now, I use systems for execution and reserve my own time for judgment.
Many members in the Solo Unicorn Club are running two or even three parallel projects. Everyone's framework is different, but the underlying logic is remarkably consistent: know what only you can do, hand off everything else, and protect the time for "what only you can do" from interruption.
Where's your biggest time bottleneck right now — too much execution, or too scattered decision-making?