A long-time client asked for a deliverable I could ship in roughly 30 hours of my own work. I had 12 hours that week. The decision was not whether to find time. The decision was whether to refuse the work, hire a contractor for the capability, or stretch myself thin enough to hand back a draft I would not be proud of. This is the log of how I decided, the screening process that made the third option real instead of theoretical, and the markup math that determines whether a contractor relationship survives past the first invoice.
The fork: hire a contractor, refuse the work, or absorb it
Most write-ups about contractors in a fractional practice present a two-option fork. Hire or do it yourself. The honest fork has three branches, and the third one wins more often than operators expect. The third branch is refusal. Saying no to a piece of work that does not fit the week, the practice, or the capability set, and keeping the engagement clean rather than packing it with a deliverable nobody will be proud of.
The wrong answer compounds. Absorb a 30-hour deliverable into a 12-hour week and the next two clients get your tired-Wednesday work. Hire an unscreened contractor and the deliverable arrives at quality you cannot ship without rebuilding it yourself. Refuse the work and you protect the practice, but you also surface a conversation with the client that some operators avoid for years because the conversation is uncomfortable.
Option A: refuse the work and stay capped
The first branch is the one I default to when the answer is unclear. Refuse the work, keep the scope intact, and absorb the conversation with the client. The phrasing I use is not "I cannot do this." It is "this is outside the engagement we scoped, and adding it now would compromise the quality of the work we agreed on. Let's scope it as a follow-on." That frames the refusal as a structural decision instead of a personal limitation.
The win is clean. The week stays inside its capacity envelope. Other clients on the calendar get the rested, on-context version of you. Quality holds because you are doing only the work you scoped. Management surface is zero.
The cost is a soft revenue ceiling. Work you would have billed if you absorbed it stays unbilled, or slides into a future quarter where it might compete with newer engagements. Some clients hear "follow-on" as a polite no and quietly route the work elsewhere. Most accept the framing if you have built credibility; some will not.
This is the right answer most of the time. The work that prompts the contractor question is rarely so unique to the moment that it cannot be deferred or rescoped. The instinct to skip past refusal is usually the instinct of someone who has not yet learned that a refusal can be a feature of the practice rather than a failure of it.
Option B: absorb the work yourself
The second branch looks easiest in the moment and is almost always the most expensive across the quarter. Absorb the new work into your own week. No contractor risk, no markup math, no review tax. Full quality control because every keystroke is yours.
The win is short-term continuity. The client gets what they asked for, on the timeline they wanted, from the operator they hired. Margin is the highest of any branch because no contractor fee leaves the system. If the work is genuinely a one-off, learnable inside your existing context, and small enough to fit the week without displacing other commitments, this is the right answer.
The cost is your week, your other clients, and your sleep. Stretching a 12-hour week into a 30-hour week pulls the extra hours from the same pool you use for everything else. Other engagements get the leftover attention. The next time something breaks in production, your reaction speed is slower because the buffer you used to absorb chaos is already spent. The failure mode that absorbs every piece of stretch work into the operator's own hours is the same failure mode I described in how a fractional operator weekly plan needs to survive a bad week, and the bad week is what eventually exposes the choice.
The honest test is short. Recurring work or one-off? Capability adjacent to mine, or a specialty where someone else is faster and better? Does the work fit inside the time I already have, or am I borrowing from the next two weeks? If two of those three answers point toward "structural, not incidental," option B is the wrong answer dressed as the easy one.
Option C: hire a contractor for the capability
The third branch is the one this article is structurally about. Hire a specialist for the capability, ship the work without diluting your own quality, and keep the client relationship intact. The branch only works if the specialist is actually screened, the contract is right, and the markup math survives the review tax.
What it gives you is real. Capacity that does not consume your week. Specialty depth on capabilities where someone who does this work every day will be faster and cleaner than you are. Predictable turnaround when the relationship is mature. The ability to say yes to work that would otherwise have been refused under option A or absorbed badly under option B.
What it costs is screening time, contract overhead, review tax, and a margin haircut. The screening time is the largest cost in the first six months of any contractor relationship. The contract overhead is real but bounded. The review tax is the recurring cost that determines whether the markup math works. And the margin haircut is the line where the unit economics either survive or quietly turn negative.
The branch is right when three conditions hold. The capability shows up in more than half of your sprints. The specialist is meaningfully faster and better at it than you are. The work survives a 40 percent markup over the contractor's rate without the client feeling the price. If any one of those is missing, the contractor relationship is more likely to dilute the practice than scale it. The wider architecture for keeping the client relationship clean while running specialists underneath sits in the subcontractor playbook for a solo practice, which this article narrows into the screening and unit-economics layer.
What I chose, and how hiring contractors in a fractional practice actually screens
For the deliverable that prompted the fork, I chose option C and dispatched it to a specialist I had already screened across two prior engagements. The work shipped on the original timeline at a quality the client signed off without revisions. The decision worked because the screening had already happened. If the relationship had been new, I would have refused the work under option A and scoped a follow-on, because starting a contractor relationship under deadline pressure is the cleanest way to ruin both the relationship and the deliverable.
The screening process I run has three steps and takes between three and six months from first contact to a relationship I would dispatch real work into. Skipping any step has hurt me, and I no longer skip them.
The first step is a paid trial deliverable. Eight hundred to two thousand dollars, fixed scope, two-week timeline, paid up front in full at the contractor's rate. The work is real but the blast radius is bounded. The deliverable is something I would have shipped myself if the contractor were not in the loop, so I have a clear baseline to compare against. I am evaluating four things and only one of them is the final output: communication across the two weeks, whether they surfaced risks early or stayed silent until the deadline, whether the final deliverable matched the scope or quietly rewrote it, and whether the work needs more than light editing to become shippable.
The trial yield is roughly one in seven. Six out of every seven trial engagements end with a thank-you and no continuation. A handful produced fine output but the process was wrong. Others communicated well but the output needed a full rewrite. A few hit on output and process but missed the timeline by enough to surface a different problem. The yield sounds harsh until you compare it to the cost of skipping the trial, which is hiring a specialist into a live client engagement and discovering the gap when the deliverable is due to the client. The trial fee is the cheapest insurance I buy in the practice.
The second step is a medium-sized engagement inside a real client sprint where the work matters but the failure mode is bounded. The work is real, the deadline is real, but I have built enough buffer into the sprint that a poor delivery from the contractor would not blow up the client relationship. This step screens for behavior under live conditions: scope drift, response under pressure, willingness to ship a draft early so the review loop has space to work.
The third step is a regular partnership with a standing rate, priority access, and a recurring cadence of work I dispatch with confidence. By this point I have seen the contractor handle three or four engagements at increasing levels of stakes, and the relationship is durable. I tell the contractor explicitly that they have priority on dispatch and that I will protect their pipeline as long as they protect my deadlines. The relationship moves from transactional to structural at this step.
I keep paying for the trial step even when a contractor comes in pre-vetted from a referral source I trust. The pre-vetting always covers technical capability. It rarely covers process under client deadline pressure, and process is what the trial is actually testing. Skipping the trial because someone vouched for the contractor is one of the few reliable ways to discover, two months in, that the relationship is not going to work.
The unit economics of contractor markup
The markup is where the relationship either pays for itself or quietly bleeds. The 40 percent floor I run is not aspirational, it is structural. Three costs sit underneath it that operators new to contractor work tend to underestimate.
Review time is the largest. Every deliverable from a contractor passes through me before it reaches the client. The review involves reading every line, running every script, opening every component, and editing where needed. Review time runs roughly 20 to 30 percent of the contractor's own hours on the work. The first hour of any contractor relationship is review-heavy because I am still calibrating their ceiling; mature relationships drift to the lower end of that band.
Rework risk is the second cost. Even with a strong contractor, roughly one in five deliverables comes back needing meaningful revision. Some of that is genuine miss; some is shifting requirements I underspecified. Either way, the rework lands in my hours, and the markup needs to absorb at least one round of rework per engagement without going underwater. A 40 percent margin survives one round of rework. A 25 percent margin does not.
Opportunity cost is the third and least visible. Every hour I spend dispatching, briefing, and reviewing a contractor is an hour I am not spending on the work the client is paying my full rate for. If a contractor is saving me 10 hours of execution but costing me 4 hours of dispatch and review, the actual time saved is 6 hours, and the markup needs to clear what I would have billed for those 6 hours of my own work, not the 10 the client perceives.
“A 40 percent margin survives one round of rework. A 25 percent margin does not.
”
The math gets cleaner when the work sits inside a productized sprint structure. A fixed-scope, fixed-price deliverable means the contractor's piece can be priced at a fixed amount and the markup baked into the sprint's price tag from the start. The client sees one number; the operator knows the contractor's number is contained inside the sprint's economics; the contractor knows the deliverable boundary because it was set at signature. Hourly billing breaks this structure. When the client is paying by the hour, every hour the contractor adds requires a separate margin calculation, and the dispatch cost compounds invisibly. I have stopped taking hourly engagements partly because the contractor math falls apart inside them, alongside the productized audit pricing math that governs the rest of the product ladder.
Two contractor relationships I run today have been live for over a year. One is in deep Shopify Liquid work, where the specialist ships a full theme section in a third of the time it would take me. One is in BigQuery dashboard work with dbt models underneath, where the specialist's familiarity with the tooling produces cleaner code than I would write. A third relationship, in creative production at volume, is mid-trial as I write this. Three relationships sounds small compared to the size of the practice, and that is the point. Most capabilities I dispatch internally because the dispatch overhead would eat the saving. Three is the right number for a solo practice running four to five concurrent engagements.
What I would revisit
The trial size, looking back, has probably been too small at the bottom of the band. Eight hundred dollars buys you a small enough deliverable that some contractors can ace it on talent alone without surfacing process problems that show up at higher stakes. The next contractor I onboard will start at fifteen hundred, with a scope that requires at least one mid-trial check-in so I can observe how they handle a mid-flight conversation.
The 1-in-7 yield is harsh and could plausibly be 1-in-4 with better filtering upstream. Most of the wash-out happens at the process layer, not the capability layer. If I can screen process before the trial - reference checks targeted specifically at deadline behavior, specifically asking for the worst miss in the last six months - the trial may end up doing less work and burning less of both our time. I have not yet rebuilt the upstream filter, and the next quarter is the right window to test it.
The capabilities I have not hired for and probably should: deep email lifecycle work in Klaviyo, specifically flow logic and segment math, where the specialist work is sharply faster than mine. I have absorbed it into my own hours for two quarters now and the absorption cost is starting to compound. The fork I described in the opening paragraph applied directly there, and the answer was option B for too long. The screening for that capability is the next one I plan to start.
FAQ
When should a fractional operator start hiring contractors?
When a single capability is showing up in more than half of your sprints, you have refused real revenue at least twice because you could not absorb the work, and the practice has enough buffer to spend three to six months on screening without that screening creating a financial squeeze. Earlier than that, the contractor overhead is more likely to dilute the practice than scale it.
What is the right size for a paid trial deliverable?
Mine has been $800 to $2,000 over two weeks, with a fixed scope and a real review at the end. I am moving the bottom of that range up because $800 buys too small a deliverable to surface process problems. Fifteen hundred is the floor I am moving toward, with at least one mandatory mid-flight check-in inside the scope.
Why is a 40 percent markup the floor instead of the goal?
Forty percent covers review time, one round of rework, and the opportunity cost of the dispatch hour. Below 40, the first time a deliverable comes back needing real revision the engagement goes negative. Above 40 the relationship is sustainable. The number is structural, not greedy. Higher margins are fine when the work supports them, but the floor is the thing that prevents quiet bleeding.
What does the client see when work is dispatched to a contractor?
They see the operator. The contractor never appears in client meetings, in client Slack channels, or on client invoices. The deliverable is signed off by the operator, not the contractor. This is the structural difference between a solo practice with contractors and a small agency, and it is enforceable as long as the contract terms make it the contractor's incentive too.
What if a referral comes in pre-vetted from someone I trust?
Run the trial anyway. The referral pre-vets capability. The trial pre-vets process under deadline pressure, which is what actually determines whether the relationship survives its first real engagement. The two are different layers, and skipping the trial because someone vouched is one of the few reliable ways to discover, two months in, that the relationship is not going to work.
How many contractors should a solo fractional practice run?
Three or four for a practice running four to five concurrent engagements. More than that and the dispatch overhead starts eating the saving. Fewer is fine if your specialty mix is narrower. The number is downstream of how many capabilities recur across your sprints, not a target to grow toward.
Sources and specifics
- The trial deliverable size of $800 to $2,000 over two weeks is the band I have run across roughly a dozen contractor screenings since 2024.
- The 1-in-7 yield is from my own trial-to-partnership ratio and is not a public benchmark.
- The 40 percent markup floor is the minimum I run; actual margins on productized sprints land higher because the sprint's price tag was set independently.
- Three live contractor relationships are active in the practice as of Q2 2026: deep Shopify Liquid work, BigQuery and dbt dashboard work, and creative production at volume (mid-trial).
- The structural decision frame in this article is grounded in roughly two and a half years of operating a fractional practice with contractor capacity layered underneath.