The promise and the reality
The design subscription model showed up in the early 2020s with a simple pitch: stop paying agencies in big, irregular project chunks; pay a flat monthly fee, queue up requests, and get unlimited design output. For founders bored of negotiating scope and timelines on every engagement, it was instantly attractive. For agencies, it was a chance to escape the boom-bust cycle of project work and build something that looked more like a SaaS company.
Five years and a lot of operating data later, we can say with some confidence: it works, but only when the agency and the client both understand what they actually bought. Most of the friction in design subscriptions comes from a mismatch between what each side thinks the product is.
What clients actually buy
A design subscription is not "unlimited design". It is "a fixed amount of senior creative attention, delivered on a predictable cadence, with no per-project negotiation". That distinction matters. Clients who treat the subscription as a magic pipe that produces infinite output on demand burn out the relationship in a quarter. Clients who treat it as a continuous creative partnership — one in which they are responsible for prioritising and scoping their own requests — get extraordinary value.
The best clients we work with on subscription send us three to five focused requests per month. Each one is well-scoped, has a clear desired outcome, and has been pre-prioritised against other things they could ask for. The worst clients send us forty requests, expect them all to ship simultaneously, and treat the queue as a wishlist rather than a work plan.
What agencies actually sell
On the agency side the model only works when the team has the discipline to refuse certain types of work. We have learned to say no to:
Speculative pitch decks. Anything where the work is being created to win other work, not to serve the brand directly. These eat enormous time and produce no compounding value.
Multi-stakeholder projects with no DRI. If the client cannot tell us who has approval authority, we cannot ship efficiently. We will produce three rounds of work and watch the team disagree about every one.
Open-ended exploration without a question. "Show us what's possible" is a project, not a request. We will run an exploration, but it is scoped, and it has a finish line.
Engagements where the subscription is the only relationship. A subscription works best when there is also a quarterly strategy conversation, a roadmap, and a sense of what we are building toward. Without that scaffolding, the subscription becomes a feature factory.
The pricing question
We charge based on the number of weekly creative hours, not the number of requests. This is non-obvious and worth explaining. Counting requests creates a perverse incentive — clients trim their requests to fit the count, agencies pad simple requests to exhaust the count, and nobody is optimising for the actual outcome. Counting hours aligns both sides on the only thing that actually scales: senior attention.
Our standard tier is one senior designer-director for 15 hours a week, with a 30-day commitment minimum. The next tier up is 30 hours a week. We do not offer "unlimited" because unlimited is dishonest — there is no version of this work where attention is unbounded.
When subscription is the wrong choice
Subscriptions are not always the right model. We steer clients to project pricing when:
- They have a single, large, time-bounded deliverable (e.g., a complete rebrand with a hard launch date).
- The work needs deep specialisation we don't carry in-house (e.g., heavy 3D production).
- Their need is genuinely intermittent — one project a quarter, not continuous flow.
For all three cases, project pricing aligns incentives better than a subscription would.
What we wish we had known earlier
Three lessons from running this model for several years:
Communicate the queue. Clients who can see the queue trust the queue. Clients who cannot see it assume their request is at the bottom of it. We send a Friday email every week summarising what's in flight, what's queued, and what we shipped. It quietly removes 90 percent of "what's the status of X" anxiety.
Onboard like SaaS. The first 30 days set the tone for the whole relationship. We invest disproportionately in week one — kickoff workshop, brand audit, first ship within ten days — so the client sees the model working before the second invoice fires.
Cancel gracefully. Make it as easy to cancel as to start. Long lock-ins create resentment, not retention. The clients we want are the ones who choose to stay because the work is great, not because the contract trapped them.
Is it actually better than projects?
Honestly, for the right clients and the right agency, yes — measurably. Higher cumulative output, faster iteration cycles, less time spent on contracts and scope negotiation, and a creative team that gets sharper on a single brand over time rather than starting cold every quarter. But the model is not a free lunch. It demands more operational discipline from both sides than project work does. Get the discipline right and the subscription compounds; get it wrong and it just becomes a slower version of the project agency, with extra confusion.