Founders ask Collective 54 this 2 times in our records, 2 of them in 2026. The productize, bundling and dropping service lines answers on this site cover packaged offers, combining services and retiring them; this page covers starting from the problems clients bring and building the service portfolio around them.
The service design essay in the newer book gives the core insight of the productization movement plainly: custom work does not scale. It says firms had treated customization as a virtue, with every client unique and every engagement bespoke, which worked while firms were small and founder-led and failed when growth introduced complexity. Productization, it says, forced clarity around what was being sold, reduced delivery variability, enabled fixed and recurring pricing, improved margin predictability and made services easier to explain and replicate.
It also lists the symptoms when service design is weak: sales struggles to explain what is being sold, delivery teams reinterpret scopes on the fly, margins erode project by project, custom work multiplies faster than headcount, growth feels fragile, and exit conversations stall over founder dependency.
The essay explains how a portfolio loses its shape. Service designers in earlier eras rarely had real market intelligence; feedback arrived late and filtered through sales conversations. So firms tweaked services based on anecdote rather than evidence, added customization to save deals and paid for it later in margin, and expanded offerings to satisfy loud clients, then struggled to support the complexity. It adds that brilliant experts naturally gravitate to what is interesting or technically elegant, while markets reward what is urgent, fundable and reliable to deliver.
As an inference, that is why standardizing around common problems works better than standardizing around what the firm likes to do. The problems are the evidence.
The essay calls the first category of service design work market truth. It includes mapping how frequently specific client problems appear across conversations and engagements, distinguishing urgency from curiosity, identifying who experiences the problem versus who controls the budget, accounting for substitutes such as in-house teams, delay and competing firms, detecting real willingness to pay rather than stated interest, and assessing the timing pressures that make a problem actionable now. It says these signals can now be observed continuously from sales calls, lost deals, client questions, delivery friction, pricing resistance and renewals.
As an inference, start with the last two years of engagements and proposals, won and lost. For each, write the problem the client described in their own words, not the service you sold them. Let AI cluster them. Most boutiques find that a handful of problems account for most of the revenue, and a long tail accounts for most of the complexity.
The replication chapter of the 2020 book describes a practical method that fits here. Take a representative sample of recent engagements, break each one down at the task level, and inventory the knowledge and skills each task required. The chapter uses this to build a certification program, so that junior staff can do work the owners have been doing, and it says the unit of profit in a healthy boutique is the project.
As an inference, the same breakdown shows what is common across engagements for the same problem. The steps that repeat become the standard method; the steps that differ show where customization really earns its keep.
The essay sets out how a service should be architected. Define the offer in terms of outcomes, not activities. Set explicit scope boundaries, what is in and what is out. Design repeatable methodologies that can be taught and enforced. Identify components that can be reused across clients and services. Define rules for customization rather than allowing ad hoc variation. And clarify the role the client must play for the service to succeed.
It adds that each service should answer what problem it solves and for whom, why a buyer will fund it now, how value is measured, and how it is delivered consistently without heroics. As an inference, if two problems on your list share most of their method, they may be one service with two entry points; the bundling answer on this site covers combining components.
As an inference, standardization fails most often at the moment a deal is at risk and someone offers to tailor the work. Decide in advance which parts of each service can flex, such as data sources, the depth of one phase or the reporting format, and which cannot, and what a change costs. The essay says to test pricing sensitivity to discounting and scope creep before a service goes to market; the scope changes answer on this site covers handling requests once work is underway.
The replication chapter is blunt that if the expert must be involved in every project, the firm cannot scale, and it recommends converting what owners know into learning content with knowledge and skills levels for staff. The essay adds that services depending on a single mind are fragile assets. As an inference, a standard service is only standard when someone other than its designer can deliver it to the same result.
The essay also explains why this matters at a sale. It says that in the exit stage buyers do not pay for heroics; they pay for transferable engines, and services that are clearly designed, economically sound and independent of a single individual command higher multiples and smoother transitions.
The essay describes AI continuously monitoring market signals, detecting patterns across sales, delivery and pricing, and triggering redesign when a service loses relevance. The 2020 book warns that clients who keep being offered the same thing become fatigued. As an inference, review the problem counts every quarter. A problem that is growing may deserve its own service; one that is fading belongs on the retirement list, which the dropping service lines answer covers.
Collective 54 publishes no service catalog template, standard number of offerings or customization limit. The published positions are custom work not scaling, what productization solved, the symptoms of weak service design, drift from anecdote, deal-saving customization and loud clients, experts favoring the interesting over the fundable, market truth and its components, the task-level breakdown of representative engagements, the project as the unit of profit, the service architecture principles, pricing tests against discounting and scope creep, single-mind dependence, continuous monitoring, and client fatigue.
If your clients are in industries that adopt AI slowly, as an inference, a traditional productized portfolio may be all you need; the essay says productization remains a valid and often sufficient step there.
If almost every engagement is genuinely different, standardize the components and the method rather than the offer.
And if one problem dominates your revenue, standardizing around it is right, but watch concentration.
Let the evidence pick the problems. The service design essay says custom work does not scale and that portfolios drift when firms follow anecdote, save deals with customization and chase loud clients. Map how often each client problem appears in your engagements and proposals, and keep the ones that are urgent and funded. Break down representative engagements, as the 2020 book recommends, to see which steps repeat. Design one service per problem with a clear outcome, scope, reusable components and written rules for customization, train others to deliver it, and let AI keep counting so the portfolio changes when the problems do.
The service design essay says to define the offer by outcomes, set scope boundaries, build a repeatable method, reuse components, write customization rules and clarify the client role. The productize answer on this site covers the full approach.
As an inference from the service design essay, review two years of engagements and proposals, record each problem in the client words, and cluster them. The essay calls this market truth.
The service design essay says custom work does not scale, and that customization added to save deals is paid for later in margin erosion and complexity.
Collective 54 publishes no number. As an inference, offer one service per recurring, funded problem, and retire offers whose problems are fading.
Sources: Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Service Design Manager for custom work not scaling, what productization solved, the symptoms of weak service design, drift from anecdote, deal-saving customization and loud clients, experts gravitating to the interesting rather than the fundable, market truth including problem frequency, urgency, budget ownership, substitutes, willingness to pay and timing, continuous signals from sales calls, lost deals, questions, delivery friction, pricing resistance and renewals, the questions every service must answer, the service architecture principles, pricing tests against discounting and scope creep, single-mind dependence, continuous monitoring and redesign, and productization remaining valid for slow-adopting industries. Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 16 for breaking down representative engagements by task, inventorying knowledge and skills, certification, experts required on every project, and the project as the unit of profit; chapter 19 for clients becoming fatigued when brought the same thing. Related Collective 54 answers on this site: how do I productize our services into repeatable, packaged offerings; how do I bundle or pair complementary services into one offering; how do I decide which service lines to drop or phase out as we reposition; how do I manage scope changes without letting them blow the budget; how do I define clear deliverables so clients know exactly what they are buying. Note on scope: Collective 54 publishes no catalog template or number of offerings. Problems as the evidence, the two-year review and clustering, using the task breakdown to find the standard method, one service per problem, written customization rules, quarterly problem counts, 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.