I've been fractional for a while now. These days I run Gev Design, fractional design and product work for founders, a small number of companies at a time.
So when a reader wrote me last week, I knew the conversation he was about to have.
His client had asked him to come on as a fractional partner. He wanted to know what to say. One day a week? Two? Three?
I told him he was answering the wrong question.
He'd done fractional once before and it went sideways. Main phase wrapped. Then the small asks started. Look at this. Jump on this call. Write this one page. He said yes every time. A few months in he'd become the guy who answers the inbox. He told me the word he used to describe himself was "vendor."
If you haven't done this before: fractional means you're a senior person working a few days a week for a company that can't or won't hire you full-time.
Situation 1: The client asks "how would you frame this?"
This is the most common opening and it feels like an invite. Treat it as a blank page instead, because they don't know either. If you answer with a number of days, you've priced a job nobody has any clue about.
What I do: I go back with a question instead of a number.
The message I send back: "Before I put a number on it, can you tell me which of these two things matters more right now?" Then I name the two jobs.
His client had floated two things in the same voice note. One was a seat at the table: sit in on the big decisions, tell us what you think. The other was hands-on: build the thing we don't have. Those are different jobs.
Don't answer a vague question with a number
Ask what they want you responsible for before you talk time
If they talk about two jobs, tell them that they're two jobs
Get the job description in writing before the rate conversation
Situation 2: They want you to name hours or days per week on the spot
I used to guess here. Now I don't, because the schedule comes after the work, not before it.
There are two kinds of fractional work and they run at different speeds. The first is advising. You sit in on meetings, you catch problems early, you're the second opinion. That works fine on one day a week. The second is building. Fix the system that stopped scaling. Three days, sometimes four, until there's something real to hand over.
What I do: I quote it in two stages, not one number. Heavier while I'm digging in and building. Lighter once the thing exists and the team can use it.
Advising work: one day a week, steady
Building work: three to four days a week, dropping off later
Agree when the days drop, and what has to be done first
Going from three days to two should feel like the plan working, not your scope shrinking
Situation 3: You don't actually know what's broken yet
This is the one I'd grab you by the shoulders for. The client says "can you build us a system for this?" and your hands want to start building. Don't. You haven't seen how they work yet. And when you look, it'll be four people doing it four different ways. Build for one of them and you've built the wrong thing for the other three.
What I do: I spend the first week doing nothing but figuring out what's actually going on. I talk to whoever's making the calls right now. I look at whatever they've already cobbled together. I hunt for the rules everyone follows and nobody wrote.
Talk to the people making the decisions today
Work out the rules backwards from what they've already built
Write up what you found and charge for it
Then you're arguing with facts, not feelings
Expect it to find two jobs where you thought there was one
Situation 4: The small ask lands three months in
Here it is. The one that got my reader last time. It arrives as "hey, quick one, could you take a look at this?" It sounds like a favour. It's phrased like no big deal. And it's outside what you were hired for. You want to be helpful. Being helpful is how scope creep starts.
You can't win this moment by having good instincts. You win it earlier, by writing down what you don't do, before anyone needs it.
Decide up front: what you decide yourself, what you check with them, what you don't touch
Put the no-list in the proposal, not in your head
Name the specific things: one-off requests, anything that doesn't move the main job forward
Do this before the first invoice goes out
When the ask comes, point at the doc, not at your feelings
I want to be clear about why this matters. The document does one job for you. When the small ask arrives, you're not negotiating your own worth in real time.
Situation 5: The work is winding down and nobody's saying it
Nobody says anything. There's less and less to do. You keep invoicing, and eventually someone feels weird about it. The engagement rots.
What I do: I put a ninety-day review in the original proposal. Both sides look at what got built, and what happens next.
Book the ninety-day review before you start, and put the date in the proposal
Ask yourself: if I disappeared Monday, could they use this thing?
Find out by handing it to them and watching them use it once, while you sit on your hands
At the review, say the three options out loud: fewer days, different job, or we're done
Leave on a date you both picked and they'll recommend you. Fade out and you'll get an awkward email in March
The days-per-week question does answer itself eventually. It just answers itself last, after you've figured out the job, worked out what's broken, written down what you don't do, and put a review on the cal.
My reader is drafting his proposal now. He told me the part that changed his thinking wasn't the pricing. It was realising he'd been treating "what I don't do" as rudeness, when the client would have read it as competence the whole time.
Thanks for reading, Gev
One more thing
Everything above assumes you can remember what you agreed to. Four clients, back-to-back calls, and the promise you made on Tuesday is in a notes app you'll never open again. The guardrails only hold if you can point at what was actually said.
So I've been building something for this. It's called Juggle. It listens to your client calls, files the summary under the right client, and turns the promises you made into tasks. No bot in the meeting, nothing in the participant list, and your data never trains anyone's model.
First group of beta testers starts this week. Early access is free while we're in private beta.
