The most consequential pricing shift in the services economy of the past decade is not, in our reading, the move to retainers. The retainer pattern is decades old. The shift worth paying attention to is more recent, less commented on, and more structurally interesting. A small but growing class of operator-grade services companies has stopped pricing on hours, retainers, or even per-engagement scopes, and started pricing on a model that looks much more like software pricing. Tiered service plans. Defined unit-of-work caps. Subscription-shaped contracts with software-shaped terms of service. The convergence is, in our reading, the right structural answer to a problem the services economy has not historically been good at solving. The piece below is the argument for why.
A note on framing. We are not talking about agencies that have built a software product on the side. That is a different and well-documented pattern. We are talking about agencies whose primary product is still services, but whose pricing surface — the way the buyer interacts with the price — has been deliberately redesigned to look like a software pricing surface. The distinction matters because the second pattern is the more disciplined one. It requires the agency to commit to a defined unit of work and to deliver against it at a defined price. That is, in our reading, the right discipline.
37signals and the priced-product agency
37signals is, in some senses, the original example. The company has, for the past two decades, operated on a model that mixes a software product with a small-scale services practice. The services practice is, in the way 37signals has run it, priced on a defined-package model rather than on hours. The buyer chooses a level of engagement and pays the priced level. The agency commits to the work in scope at the priced level. The pricing surface is, in effect, a software-style tiered plan with a services-style deliverable.
The interesting thing about 37signals's approach is that it predates the broader convergence we are writing about. The company arrived at this pricing posture not because of a structural shift in the services economy but because of the founders' own positions on how to run a business. Jason Fried and David Heinemeier Hansson have, in their public writing, been consistent that pricing should be predictable for both sides of the transaction. The defined-package model is, in their framing, the right answer to that predictability question. The model has aged well.
Basecamp's services-and-software hybrid
Basecamp is, structurally, a software product. The agency-style services that have, over the past decade, been bolted onto the Basecamp pricing surface — the implementation packages, the workshop and training tiers, the priced advisory engagements — are, in the way they are priced, examples of the convergence we are describing. The buyer does not pay for hours. The buyer pays for a defined tier of engagement that comes with a defined unit of work. The model lets Basecamp's services side of the business operate with the unit economics of a software product, even though the work being delivered is delivered by humans.
For an operator-class founder running an agency, the Basecamp services pattern is, in our reading, the more directly transferable. The Basecamp services packages are simple enough that an operator running a small agency can rebuild the same structure inside their own pricing surface in a quarter. The structural commitment — defined tier, defined unit of work, defined price — is the right commitment.
Linear's services tier as an in-product upsell
Linear's services tier is the version of the pattern that is integrated into a software product the buyer is already paying for. The buyer who is using Linear and who wants implementation support, training, or other priced services can buy the priced tier inside the same surface where they pay for the software. The pricing surface is, by deliberate design, the same as the rest of the product's pricing surface. The services are, in effect, a higher tier of the software product.
The Linear model is, in our reading, the version of the pattern that is most likely to be widely adopted by software companies over the next several years. The structural argument is that the buyer who is already paying for the software has a clear path to paying for the implementation, training, or advisory work that makes the software more valuable. The pricing surface does not require the buyer to switch contexts. The work is priced on the same scale as the software. The deliverable is defined by the same tier system.
For an agency competing against the Linear pattern, the response is, in our reading, to ship the same shape of pricing surface around a deliverable that the software company cannot ship. That is the structural opening for the operator-class agencies that are doing the convergence on their own terms.
The structural argument for the convergence
The structural argument for why agencies should price like software is, in our reading, three-part.
The first part is the predictability argument. An hourly or per-engagement scope pricing model produces variance for both sides. The buyer does not know what they will spend. The agency does not know what they will earn. The hours-based model amortizes the uncertainty over the engagement and resolves at billing. The defined-tier model resolves the uncertainty at the start. Both sides know what the engagement will cost and what it will produce. The predictability is, in our reading, the single largest improvement on the working day.
The second part is the discipline argument. An agency that has to price on a defined tier has to commit to a defined unit of work. That commitment forces the agency to scope the engagement carefully, to refuse work that does not fit the tier, and to deliver against the tier on a defined timeline. The discipline of pricing forces the discipline of delivery. An agency that prices on hours can absorb scope creep at the buyer's expense. An agency that prices on tiers cannot. The tier model produces, on the evidence of the agencies that have made the transition, more disciplined engagements.
The third part is the leverage argument. An agency that prices on hours is, structurally, a labor business. The agency's revenue scales with the agency's headcount. An agency that prices on defined tiers can, with the right delivery infrastructure, deliver the work at a higher leverage ratio than the headcount would suggest. The structural opening for the agency to deploy agentic stacks, internal tooling, and other forms of leverage inside the delivery work is, on the priced-tier model, structurally larger than on the hours model. The hours model penalizes the agency that becomes more efficient. The tier model rewards it.
The risk on the other side
The risk on the other side of the convergence is, in our reading, the discipline of refusal. An agency that has priced on a defined tier has to be willing to say no to work that does not fit the tier. That is harder than it sounds. The buyer who wants a bespoke engagement is, often, the buyer with the largest budget. The agency that says yes to the bespoke engagement is, in effect, breaking the tier model for the rest of its buyers. The agency that says no is, in effect, walking away from the largest available revenue.
The operators who have built priced-tier agencies in 2025 and 2026 are, almost without exception, the operators who have learned to say no. That is the structural skill the convergence requires. An agency that has not learned to say no will not, in our reading, be able to maintain the priced-tier model for long. The bespoke engagements compound. The tier model erodes. The agency reverts to the hours model.
Why we are writing this now
We are writing this now because the convergence is, in our reading, in the middle of its inflection. A meaningful share of the operator-class services economy is moving toward defined-tier pricing. The next two years are, in our reading, the window in which the pattern becomes the default for the segment of the services economy that is structurally able to commit to it. The agencies that make the transition early will compound on the operational benefits of the tier model. The agencies that resist the transition will, in our reading, be left running the labor business while their competitors run a structurally cheaper and more disciplined business at the same revenue level.
The buyer side of the transaction is, in our reading, ahead of the agency side. Buyers have, over the past several years, become accustomed to software-shaped pricing surfaces. The transition to buying services on the same shape of pricing surface is, on the buyer's side, a small step. The agencies that ship the pricing surface the buyer is already expecting will be the agencies the buyer chooses.
Operator Press will continue to track the agency-economics convergence through the rest of 2026. Operators running tier-priced services businesses who want to be considered for future coverage can write to our editorial desk at editorial at operator.press.