What does "unlimited design" get you in a month? Queues, parallel tasks and real throughput

The short answer: "unlimited" describes how many requests you may file, not how many get finished. Every unlimited design plan is a queue with a fixed number of tasks in progress at once. What comes out in a month is decided by four numbers: tasks worked in parallel, time to a first concept, revision rounds per task, and the days the team works each week. In the illustration below, the same month finishes three tasks at one end and twenty-five at the other, on plans that all say "unlimited".

The arithmetic in this guide is an illustration with round numbers. It is not a promise from any provider, including us. Provider figures are quoted from their own sites as we read them on 12 October 2026.

What does "unlimited" mean on a design plan?

Unlimited design is a pricing model in which you may file any number of requests for a flat monthly fee, and the provider works a fixed number of them at a time. The limit moves from the invoice to the calendar.

Providers say this themselves, usually in the FAQ. Designjoy's reads: "you're able to add as many design requests to your queue as you'd like, and they will be delivered one by one." Duck.Design's reads: "The only limitation is turnaround time, as our designers are real people doing the work."

Delivery teams have precise words for this. The Kanban Guide defines work in progress as "The number of work items started but not finished" and throughput as "The number of work items finished per unit of time". A design subscription sells you a work-in-progress limit. Throughput is what you are trying to buy, and no pricing page states it.

Which four numbers decide what you get?

Tasks worked at once. The number of requests in progress at the same moment. Everything else waits.

Time to a first concept. The hours from filing a task to seeing something. Check what the clock measures: a first update or finished work, business hours or real ones.

Revision rounds. Unlimited revisions cost nothing in money and a day or two each in time: the designer's turnaround, plus however long you take to reply.

Days worked each week. A 48-hour turnaround on a task filed on Friday afternoon arrives on Sunday from a team that works every day, and on Tuesday from a team that works Monday to Friday. Duck.Design's FAQ, for example, says "Designers work on Monday to Friday, from 09:00 to 18:00 (GMT+2)."

The table shows how four providers state these numbers.

Provider	Worked at once	First turnaround	Days, as stated
Designjoy	"One request at a time"	"Avg. 48 hour delivery"	"two business days or less" on average
Penji	2, 3 or 4 "Active Workstreams", by plan	"1 day turnaround" or "Same-day turnaround", by plan	Comparison table reads "1 business day"
Teamtown	2, 5 or 8 "projects at a time", by plan	"24-48 hour turnaround"	"24/5 customer support"
Recast	Five	First concepts in 48 hours	Every day of the week

How much is that in a month? An illustration

Take one ordinary task: a product screen at every state. Assume a first concept in two days, then two revision rounds of two days each, one day for you to comment and one for the designer to revise. That is six working days from start to done.

A 30-day month has about 22 working days for a team that works Monday to Friday, and 30 for a team that works every day. Each slot finishes one task every six days.

Set-up	Working days	Finished per slot	Finished in the month
One task at a time, five-day week	22	22 ÷ 6 = 3.7, so 3	3
One task at a time, seven-day week	30	30 ÷ 6 = 5	5
Five in parallel, five-day week	22	3	5 × 3 = 15
Five in parallel, seven-day week	30	5	5 × 5 = 25

Three things to take from it. Every row describes a plan sold as unlimited. The spread between the first row and the last is about eight to one. And the model flatters all four: it assumes every task is the same size, your comments arrive within a day, and nothing waits on a decision. The seven-day rows also assume your feedback keeps pace at weekends. If you only reply on weekdays, the real figure sits between the two.

Why do five parallel tasks not mean five times the output?

Three limits apply.

Your own review time. Five tasks in progress means five threads waiting on your comments. If you take three days to answer, each round is four days, not two, and the six-day task becomes ten.

The hours behind the slots. A slot is not a person. Some providers publish the conversion. Penji's pricing lists "2 Active Workstreams (Equal to 2-4 hrs daily)" on its $995 plan and "4 Active Workstreams (Equal to 10-14 hrs daily)" on its $4,500 plan. Kimp's pricing FAQ says the team "will allocate 2-2.5 hours each business day to completing your requests". Teamtown describes its model as "an hourly-based subscription". Where hours are stated, use them: a task that needs four hours of design takes two days on a plan that allots two hours a day, whatever the turnaround line says.

Task size. A slot holding a design system is busy for weeks. The other four keep moving, which is the real argument for parallel work: one large job does not stop the small ones.

How do you size a task?

A task is one deliverable that a designer can take from brief to first concept in one turnaround. If the first concept cannot arrive in the stated time, it is more than one task, and it will be split whether you split it or not. Designjoy's FAQ says larger requests "are broken down on Designjoy's end" and that you should "expect to receive a reasonable amount of work every 24-48 hours until the entire request is done."

Split it yourself, because then the order is yours:

• Small, hours: an empty state, a settings row, a revised table header, six more icons in an existing set.
• Standard, one turnaround: one screen at every state, or a flow of two or three screens that only make sense together.
• Large, split it: "redesign onboarding" becomes the sign-up form, the verification step, the first-run checklist and the empty dashboard, filed in the order your engineers will build them.
• Programme, weeks: a design system or a full product audit. Give it one slot and keep the others for everything else.

One screen or one flow per task, with a brief of five short parts. How to brief a product designer has the template.

How many parallel tasks do you need?

Work backwards from your own month. Count the tasks you would file, multiply by the days each takes end to end, and divide by the working days.

An example, with the same assumptions as the table. A team that ships every two weeks files eight standard tasks a month. At six days each, that is 48 task-days. On a five-day week, 48 ÷ 22 = 2.2, so three slots. On a seven-day week, 48 ÷ 30 = 1.6, so two.

The same sum tells you when a subscription is the wrong tool. Forty tasks in a month is 240 task-days, which is eight slots on a seven-day week. That is a project team, and an agency is the right recommendation.

Before you subscribe anywhere, ask for the four numbers in writing, and ask what the turnaround clock measures. How to check a provider's scope covers what the work includes, and how to test a design partner covers trying it first. For the four numbers side by side, see Recast compared with the alternatives.

Where Recast fits

Recast's plan is $3,900 a month, billed monthly, with no annual contract. Unlimited tasks, five worked in parallel, the rest queued. First concepts in 48 hours, and because the team works every day of the week, those are 48 real hours. Revisions are unlimited.

That puts the plan in the last row of the table as a set-up. It does not put a number of finished tasks in your month, because that depends on the size of your tasks and the speed of your replies.

The quickest way to find your own number is to measure it. The $490 pilot is one week with up to three active tasks. Or book the free 20-minute pilot call, or read the plan.