Conveyor belt or workshop
The bespoke shape
In a bespoke service business, every client gets a custom design. New industry, new processes, new metrics, new dashboards each time. The operator becomes deeply skilled at improvisation. Margins can be high on individual jobs, but the founder is structurally trapped: hires take a long time to bring up to speed, no two clients are alike, and quality depends on the founder's personal judgement on each project.
The productised shape
In a productised service business, the offering is the same for every client. Same intake form, same timeline, same delivery method, same reporting pack. The operator gives up some headline rate per project, but gains the ability to hire, train, document, and scale. Quality stops being a daily founder problem and becomes a process problem, which is far easier to solve.
Why the choice is rarely accidental, and how to spot the trap
Almost everyone starts in the bespoke shape because the first few clients arrive with different needs and the founder takes whoever pays. The decisive move is to refuse the next custom request and instead concentrate on a single, repeatable form of engagement. Signs you are already in the trap: every new client requires fresh setup work, you cannot describe the engagement on a single page, hires keep breaking things, and the founder is the only person who can deliver consistently. None of these problems are solved by more effort; they ease only when the shape changes.
Productisation is also where defensibility lives
A productised practice is harder for a competitor to copy than a bespoke one, because the asset is not any single delivery — it is the documented process behind every delivery. Standard operating procedures, client-facing playbooks, trained classifiers, evolved scripts, and the tuning of which sequence works for which client profile: this is process intellectual property and it accrues quietly across many engagements. Combine that with switching costs (clients have many parts of their operation depending on you), accumulated client-outcome data (you know what works for sector X better than any newcomer can), and a brand that names a recognisable category — and you have what an analyst would call a moat. None of these defensibility sources exist in a bespoke shop, because nothing is repeated long enough to compound.
Real-world parallel — WP Curve and the productised exit
WP Curve, founded in 2013 by Dan Norris and Alex McClafferty, sold WordPress maintenance as a flat monthly fee. Customers received unlimited small jobs (each under thirty minutes) for one fixed price. The team refused custom development, framework migrations, and large redesigns — anything that did not fit the standard offering. Within roughly three years they had processed more than 105,000 jobs across over 1,000 paying members. GoDaddy acquired the business in December 2016. Norris has written that the moment the practice became scalable was the moment they started saying no to non-standard requests, even when those requests came with cash attached. The discipline of refusal — uncomfortable in the short term — is the defensibility argument made concrete.
Picture two operations: a small workshop in which every order is unique, and a small assembly line in which every order is the same. Which one are you building today? Which would you prefer to be running in three years?