Client count is evidence, and evidence is not a safe harbour
The number of clients on your invoices is the single most persuasive fact in a personal services business review, and it is still only a fact. The statutory question is whether you would reasonably be regarded as an employee of the person you are serving if your corporation did not exist, and nothing in that question counts clients. What several concurrent clients do is prove something the test cares about deeply: that you are in the market, selling your services to whoever will buy them, carrying the risk of finding the next engagement yourself. That is what being in business on your own account looks like, and it is hard to fake for long.
The reason it is not a safe harbour is that the same facts can be read the other way for any single relationship. A contractor with two clients who works four days a week at the first one's office, on their laptop, under their manager, with a set schedule and a day rate, has a small independent practice and a job. The second client is real evidence about the contractor; it says nothing about how the first relationship is actually run. Reviewers look at the day, not the ledger.
The four tests that decide each relationship, control, tools, chance of profit and risk of loss, and integration, are explained in what is a personal services business. This page assumes you know them and is about one recurring assumption we have to correct: that adding a second client answers the question.
The test runs relationship by relationship, so a second client does not launder the first
A corporation can be a personal services business with respect to one client and an ordinary business with respect to another, in the same year, on the same T2. The definition speaks of the person or partnership to which the services were provided and asks whether you would be their employee, so it is applied to each service relationship on its own facts. Income from the engagement that looks like employment is PSB income, taxed at roughly 44.5 per cent with almost no deductions; income from the engagement that looks like a genuine business is active business income, taxed at 12.2 per cent on the first 500,000 dollars with ordinary deductions. The corporation reports both.
That has two practical consequences. The first is that the dominant engagement does not become safe because a smaller one exists beside it; if the big client controls your work and supplies your tools, the revenue from that client is exposed regardless of the second invoice. The second is more useful: a genuine second client often changes the facts of the first relationship, not just the count. To serve two clients you need your own equipment, your own calendar, a contract without exclusivity and a real ability to decline work, and each of those moves the first relationship toward business; the protection, where it exists, comes from those changes, not from the second name on the invoice.
A corollary follows for how you run the corporation. If one relationship is exposed and the other is not, the compensation plan has to be built around the exposed one, which usually means paying that income out as salary while the rest can be retained. Blending both streams and treating the whole corporation as safe because it has two clients is how a mixed file becomes an expensive reassessment.
Token second clients, and why they can hurt
A second client added for the file rather than for the business tends to make the file worse, and reviewers are practised at spotting the difference. The classic tokens are an invoice or two to a friend's company, a related corporation, a spouse's business or a former colleague, for a small amount, once a year, with no proposal, no separate pricing and no evidence of work. Sometimes the corporation invoices a second entity in the same corporate group as the main client, which is one client wearing two names. None of it changes how the main engagement is run, which is the only thing under review.
| Marker | Token second client | Genuine concurrent client |
|---|---|---|
| How it was won | Offered by a friend or relative; no proposal | A pitch, a referral or your marketing, with a proposal you priced |
| Share of revenue | Small and flat, typically one or two invoices a year | Meaningful, or at least growing, with multiple invoices over the year |
| Pricing | Whatever was convenient | Your rate card or a fixed price you set and could lose money on |
| Timing | Started after you learned about the PSB rules | Overlaps with the main engagement on an ordinary business basis |
| Effect on the main client | None; the main engagement is unchanged | Forced changes: your own tools, no exclusivity, control of your calendar |
| Relationship to you | Related party, or an affiliate of the main client | Arm's length, found in the market |
The damage a token client does is threefold. It signals that you knew the test and dressed for it, which colours how every other fact is read. It is frequently inconsistent with your main contract, which may contain an exclusivity or non-compete clause you have now apparently breached, or a full-time hours commitment that leaves no room for the second engagement you say you served. And where the second client is related to you, the revenue can raise its own questions about whether the amounts were reasonable or real, so a weak second client is not a neutral addition; it is a fact that cuts against you.
What genuinely concurrent clients look like
A genuine second client is one you could describe to a stranger as a customer without embarrassment, and the description involves money you risked to win them. You marketed, or were referred, and you wrote a proposal with a price you set. You could have lost the pitch.
You deliver on your own tools, invoice on your own terms, chase your own receivables and carry the risk that they do not pay. If the main client called tomorrow and asked for five days a week, you would have to say no, or renegotiate, because you have another commitment. That last point is the one reviewers weigh most: concurrent clients constrain each other, and a constraint that shows up in the main engagement is the evidence that matters.
Concurrency is the operative word. Serving one client for two years and then another for two years is a sequence of engagements, and each is assessed on its own; a contractor who moves from placement to placement has one client at a time, however many names accumulate on the corporate history. Concurrent clients, even if the second is modest, demonstrate the capacity and the freedom to serve more than one buyer at once. A retainer with a small firm alongside the main engagement, a fixed-price project delivered on weekends, paid workshops, a productized service sold to several buyers, all count if the work and the money are real.
Size helps but is not decisive. A second client at 10 per cent of revenue that was genuinely won and is genuinely served tells a reviewer more than a related-party invoice at 30 per cent. What we look for in a file is a pattern that would persist if the PSB rules were repealed tomorrow, because that is the pattern a court would find credible.
The only bright line, and the agency trap
There is exactly one bright line in the personal services business rules, and it has nothing to do with client count: a corporation that employs more than five full-time employees throughout the year is not a PSB at all. The count is of employees, not contractors, it has to hold for the whole year rather than at year-end, and you can be one of the employees if you are genuinely employed full-time by your corporation. A related exit removes services provided to an associated corporation from the definition. Neither is available to a solo consultant, and neither can be manufactured; a business with six full-time staff is outside the regime because it plainly is a business.
The agency trap is the mirror image, two contracts that look like two clients but are one: an incorporated contractor placed by a staffing or placement agency has a contract with the agency and delivers services to the end client, and the question of whose employee you would be is asked about whoever actually receives the services, which in a placement is normally the end client. Invoicing the agency does not make the agency a client in the sense the test cares about, and a corporation placed by three agencies at three end clients in sequence has still had one client at a time. What the agency chain does add is a second set of contract terms, often never seen by the contractor, that tend to read as employment. If your situation runs through an agency, should an IT contractor placed through an agency incorporate deals with that fact pattern directly, including when the corporation is not worth having.
How to document independence honestly, and what changes the answer
Documentation is the record of a business that already exists, not a substitute for one, and it is only worth keeping if it is true. The file we build with clients has a fixed shape: every contract and renewal, with the substitution, exclusivity and termination clauses marked; proposals sent and pitches lost, because a lost pitch is proof you were in the market; invoices that describe deliverables rather than hours where the engagement allows it; proof of your own equipment, licences, insurance and business identity; and a dated one-page summary each year of which clients you served, how each was won and what changed. The wider programme for changing the facts themselves, contract by contract, is in how an incorporated consultant can reduce personal services business risk.
What you should not do is paper a relationship that does not exist, backdate anything, or describe a second client in the file in terms you would not use in front of them. PSB reviews happen years after the fact, and the file is read against your client's answers to the same questionnaire. A file that is modest and true beats a file that is impressive and contradicted.
The facts that decide whether your client mix actually protects you:
- Whether the clients are concurrent: served at the same time, so that each constrains the other, rather than one after another.
- How the second client was won: in the market, at your price, with a real chance of losing the work.
- Whether the main engagement changed: your own tools, no exclusivity, control of your calendar, a substitution right you would use.
- Who the counterparty really is: an arm's-length buyer, or a relative, an affiliate of the main client, or an agency standing in front of a single end client.
- Headcount: more than five full-time employees all year takes the question off the table; fewer leaves it on.
- How the exposed income is paid out: salary against the relationship that looks like a job caps the cost of being wrong.
We assess client mix as part of an annual PSB review inside Tax Planning & Advisory: which relationships are exposed, what the second client actually proves, and how much of the year's profit should leave as salary before year-end. If you have added a second client and are counting on it, a free 15-minute discovery call is enough to tell you whether it carries the weight you are putting on it.
