Designing an app that never shipped
Most portfolios contain unshipped work and are quietly vague about which pieces those are. Being precise about it is what makes the rest of the page believable.
Most design portfolios contain work that never shipped, and most of them are quietly vague about which pieces those are. The vagueness is understandable and it is a mistake — it makes a reader discount everything on the page, including the parts that did ship.
The better position is to say plainly which work is a concept, and then be clear about what a concept can and cannot demonstrate. That turns out to be a short and useful list.
What a concept cannot show
Starting here, because it is the part usually skipped:
- That anyone wanted it. No one downloaded it, because there was nothing to download.
- That the interface works under real use. Nobody used it tired, one-handed, on a cracked screen, with a year of accumulated data in it.
- That it could be built at that quality. Every unshipped design is, to some extent, a promise about engineering that was never tested.
- Anything about outcomes. No retention, no engagement, no conversion. There is no measurement, and there cannot be.
Any portfolio claiming otherwise about unshipped work is describing an intention as a result. It is also usually detectable, which is the practical argument against it: a reader who spots one inflated claim stops trusting the ones that were true.
What it does show, which is not nothing
A concept is a real piece of thinking made concrete, and concrete thinking is checkable in ways a description is not.
Whether the problem was understood
The structure of an interface encodes a claim about what matters. What is on the first screen, what is one tap away, what is buried — those are arguments, and a reader can evaluate them without any usage data at all.
Whether the hard parts were faced
This is where concepts most often fail, and where an honest one earns its place. Every product has screens nobody enjoys designing: the empty first-run state, the settings, the permission request, the moment the data is wrong. Concept work that shows only the beautiful middle of the product has skipped the part that determines whether it would survive contact with reality.
Whether it holds together
A single screen proves very little. Twelve screens that share a spacing scale, a type ramp, a navigation model and a consistent idea of what a card means demonstrate a system — and a system is the thing that transfers to other work.
A concept on this site
Saying that costs nothing, and it is what makes the rest of the page readable as what it is: interface and flow decisions, presented as decisions rather than as validated outcomes.
How to make concept work actually useful
If you are producing self-initiated work — to explore an idea, to keep a skill sharp, or for a portfolio — a few things separate the pieces that are worth showing from the ones that are decoration.
- Pick a real constraint and keep it. "A sleep app" is a mood. "A sleep app for someone who checks it half-awake at 3am in a dark room" is a constraint that decides type sizes, contrast, tap targets and what the first screen shows. Constraints are what make design decisions checkable.
- Design the unglamorous screens. Empty states, errors, permissions, settings, the first run before there is any data. These are where the concept either becomes a product or stays a poster.
- Use realistic content. Real-length names, real numbers, a week where the data is missing. Placeholder text that always fits is how a layout gets approved and then breaks.
- Write down what you decided and why. The reasoning is the transferable part. A reader can disagree with a decision they can see; they cannot engage with one they have to infer.
- Label it honestly and move on. One line saying it is a concept, stated early, and then no hedging for the rest of the page.
Why this is not a weakness to manage
There is a worry underneath all of this: that admitting work did not ship makes a studio look less capable. In practice the opposite reads better, for a reason worth stating directly.
A portfolio where everything is implied to be shipped client work is not more impressive — it is less specific, and specificity is what a reader is actually assessing. A page that distinguishes commissioned work from shipped work from exploration is telling you the author knows the difference matters. That is the signal.
It is also the only version that survives being checked. Anyone deciding whether to hire a studio will click the live links, and a claim that does not hold up costs far more than the claim was ever worth.
If you have an idea you want explored properly before committing to building it, tell us about it.