Founders ask Collective 54 this once in our records, and that was in 2026. The full-time versus contractors, offshore and when to hire answers on this site cover employment types and timing; this page covers a capability the firm does not have yet and whether to buy it or build it.
The full-time versus contractors answer on this site splits work into three kinds: billable core delivery, non-billable overhead and variable surge capacity. As an inference, the same split answers most of this question. A capability you lack falls into one of two groups. Either it is part of what clients pay you for, or it supports the firm without being sold.
The IT essay in the newer book is clear about the second group. It says IT is non-billable by design and that, as with finance and HR, the most disciplined firms keep it lean, fractional and outsourced, because full-time internal roles add fixed cost and management complexity without adding revenue. Support capabilities like these should usually be rented indefinitely.
The replication chapter of the 2020 book says boutiques are hired because they are experts in an area, and that if the founder is the expert, clients will require the founder on every project, which does not scale. It says the expertise must be replicated in the staff, that the firm needs many experts, and that growing your own is more profitable than seeking hired guns.
The HR essay in the newer book explains the economics. It calls talent the raw material, the inventory and the delivery mechanism of a services firm, and describes an eight-stage talent supply chain: recruit, select, onboard, deploy, develop, retain, promote and succession. It says development compounds skills, expands capacity and reduces dependence on any single person, and that promotion builds middle management depth. As an inference, an outside expert can be deployed, but only your own people move through the later stages where the value compounds.
The service offering chapter of the 2020 book describes the choice SBI faced when clients asked for help it did not yet offer, such as sales territories, quotas and compensation. It could refer the work to partners for a quick referral fee, or develop the capability in-house, costly in the short term and lucrative in the long term. It chose to build, grew to more than one hundred offerings and sold for about 30 percent above comparable firms. As an inference, the cost of building is real, but so is the cost of telling a client you cannot help with something they need now.
As an inference, for a core capability the practical answer is usually both, in a set order. Bring in outside help for the first engagements, so clients are served well while you learn. Pair that help with the people who will own the capability later, so they learn on real work rather than in theory. Capture what the outside expert does as method, templates and examples, so the knowledge stays when they leave. And set a point at which your own people lead and the outside help moves to review, then leaves.
The replication chapter describes the tools for the build: postmortems of recent engagements, an inventory of the knowledge and skills each task requires, and certification of people at 101, 201 and 301 levels, with learning paths that move them up. It recommends renting an instructional designer to turn tribal knowledge into learning content. As an inference, the same approach turns borrowed expertise into your own.
The HR essay describes what a failed hire costs a small firm: recruiting spend, onboarding time, lost productivity, team disruption and margin erosion, and when a hire fails the firm pays twice, once for the mistake and again to replace it. It says selection errors are expensive and difficult to unwind. The replication chapter admits that on a single project it is always less efficient to use people who are still learning, because they need supervision and take longer. As an inference, building a capability is slower and riskier in the first year than renting it, and cheaper every year after that if the work keeps coming. Compare both paths over three years, not one engagement.
The IT essay describes a division of labor that applies beyond technology. Boutique firms now need CTO-level thinking but cannot hire that talent, because it is too expensive, too scarce and rationally attracted to technology firms. The answer it gives is not to build an internal IT team but to own the design and outsource the execution. As an inference, when a capability is too specialized or expensive to grow, keep the decisions about how it is used inside the firm, and rent the hands.
The operations essay adds a warning about the rented option. It says fractional arrangements fail without clear role design, because fractional does not solve ambiguity. As an inference, before bringing in outside help, write down what they own, what they may decide and what they must leave behind.
The finance essay in the newer book says every hour now carries a dollar cost and lists the decisions that follow: automate, shift to AI, move offshore, give to a more junior role, or keep with a senior expert. As an inference, part of a missing capability may be work AI can now carry, with a person supplying judgment and review. The AI tool versus hire answer on this site covers that comparison.
The exit essay says buyers pay for transferable engines, and the full-time versus contractors answer notes that a rented delivery bench costs you at exit, when buyers discount for talent that does not transfer. As an inference, a firm whose core service depends on outside specialists is harder to sell than one that has grown the skill, even if the work looks the same to clients today.
Collective 54 publishes no rule for when to stop renting a capability and no ratio of employees to outside help. The published positions are replicating expertise in staff, growing your own over hired guns, talent as the raw material and inventory of the firm, the eight-stage talent supply chain, the SBI choice to build rather than refer, certification and learning paths, overhead functions kept fractional, owning the design while outsourcing execution, fractional help failing without role design, and every hour carrying a cost.
If demand for the capability is uncertain, as an inference, rent it until you have seen it sold several times; building for one client is a favor, not a strategy.
If the skill changes faster than your people could learn it, outside help may stay the better option even for client work.
And if a client deadline is close, get the help now and plan the build later; the client should not pay for your learning curve.
Ask whether the capability is part of what clients pay you for. If it supports the firm, rent it and keep renting it, as the IT essay says of overhead functions. If it is core, the 2020 book says growing your own experts is more profitable than seeking hired guns, and SBI built rather than referred. Borrow outside help now so clients are served, pair it with the people who will own the capability, capture the method, certify your people and set a date for the outside help to step back. Own the design and rent only the execution.
The 2020 book says growing your own experts is more profitable than seeking hired guns. As an inference, borrow outside expertise for the first engagements and use it to train the people who will own the capability.
The IT essay says non-billable functions should stay lean, fractional and outsourced. As an inference, rent capabilities that support the firm and build the ones clients pay for.
As an inference from the 2020 book, pair them with your own people, capture the method as templates and examples, and certify your staff on the knowledge and skills the work requires.
As an inference from the exit essay, yes if they deliver your core service, because buyers pay for transferable engines and discount talent that does not transfer with the firm.
Sources: Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 16 for replicating expertise in staff, juniors being less efficient on a single project, the founder as the expert on every project, growing your own over hired guns, postmortems, knowledge and skills inventories, certification at 101, 201 and 301 levels, learning paths and renting an instructional designer; chapter 19 for the SBI choice to build capabilities rather than refer them for a fee, more than one hundred offerings and the price about 30 percent above comparable firms. Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI HR Manager for talent as raw material, inventory and delivery mechanism, the eight-stage talent supply chain, and the cost of a failed hire; The AI IT Manager for overhead functions kept lean, fractional and outsourced, CTO-level thinking boutiques cannot hire, and owning the design while outsourcing execution; The AI Operations Manager for fractional arrangements failing without role design; The AI Finance Manager for every hour carrying a dollar cost and the decisions that follow. Greg Alexander, Why Some Boutique Firms Exit Cleanly and Others Never Really Do (Collective 54), for buyers paying for transferable engines. Related Collective 54 answers on this site: should I hire full-time employees or use contractors and fractional talent; should we use contractors or offshore talent, and how do we manage that; how do I know when and who I need to hire; should we hire an operations person or vendor, or build this ourselves; is it cheaper to use an AI tool than to hire someone for this. Note on scope: Collective 54 publishes no rule for when to stop renting a capability. The core versus support test, borrowing now and building on purpose, the four steps of the transfer, applying own the design and rent the execution beyond technology, and the flips are inferences used here to organize the source material rather than published Collective 54 positions.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.