How a Concept Package Is Put Together

A concept package is frequently treated as a small rendering job with several outputs, and that framing produces something that looks fine and fails at its actual purpose. It is closer to designing an experiment than to producing imagery, because the object is to let a group of people compare several things fairly and reach a decision they can defend.

This guide describes how one is assembled, in order, and what each decision determines. The technical production is the least interesting part and the structural decisions are where the value is.

Step one: reduce the options before producing anything

The most common state at the start of this process is too many schemes. Five, seven, sometimes more, because generating variants has become cheap and nobody has been forced to eliminate any.

Producing imagery for all of them is expensive and, more importantly, it does not help. A group presented with seven options does not compare seven options. They form an impression from two or three and ignore the rest, and which two or three depends on the order they were shown rather than on merit.

The reduction to two, three or at most four is a judgement about the decision rather than about the design, and it usually has to be made by whoever owns the project rather than by the design team, who are reluctant to kill their own work.

That step costs nothing, saves the most money of anything in this process, and is skipped more often than any other.

Step two: decide what question each image has to answer

Every image in a concept package should have one job, and writing that job down before production changes what gets produced.

Common jobs: how tall does this feel from the street, how does it sit against the neighbour, how much of the site does it consume, what happens at the corner, what does the approach look like.

An image without a stated job tends to become a general view, which answers everything vaguely and nothing specifically, and general views are what most concept packages consist of.

Naming the job also determines the viewpoint, which is the next decision and the one that most affects whether the package works.

Step three: choose one setup and hold it

This is the technical decision that makes the comparison valid, and it is a single commitment applied everywhere.

One camera position. One camera height, at eye level unless there is a reason otherwise. One time of day and one sky. One material treatment. One level of surrounding context. One level of detail.

Everything that varies between images should be a difference between the schemes, and nothing else. If option B is shown from a slightly better angle, or in slightly warmer light, or with slightly more landscape, the package has made the decision before the meeting started.

Holding that discipline is harder than it sounds, because each individual deviation is defensible on its own terms. It is only in aggregate that it becomes a distortion, and by then nobody can see it.

Step four: build the context once

The surroundings are shared across every option, so they get built once and reused, which is both a saving and a consistency guarantee.

What the context needs is correct height and position rather than detail. Neighbouring buildings as accurate volumes, the street, the ground with its real levels, and anything that will remain.

It is the item most often trimmed for budget and it is the one that carries the most explanatory weight, because the question at concept stage is nearly always relational. A building shown in isolation cannot answer how it sits against anything.

Where the context is only approximately known, establishing it properly at this stage pays for itself repeatedly, since the same context carries into entitlement imagery and eventually into marketing.

Step five: model each option to the same standard

The schemes are built to the level of detail that has actually been decided, which at concept stage is usually massing, orientation, approximate height, the relationship to the street and the broad arrangement.

What is not decided is facade treatment, materials, window rhythm, entrance detail and landscape. Those get represented as categories rather than as choices: a glazed area rather than a specific curtain wall, a masonry mass rather than a particular brick.

Getting this level right is the hardest judgement in the whole process, and both failure directions are common. Too plain and the audience cannot read the building. Too resolved and they read commitments nobody made.

The working test is whether somebody could point at any element in the image and ask when that was decided. If the honest answer for any element is that it was not, that element should not be there.

Step six: light it neutrally

Light does more persuasive work than anything else in an image, which is exactly why it should be flattened at this stage.

Neutral overcast or soft even daylight, the same across every option, with no golden hour, no dramatic shadow and no reflective glass catching a sunset.

This is frequently resisted, because the resulting images look less impressive than the client expected. That reaction is worth anticipating and explaining before delivery rather than defending afterwards.

The explanation is short: the images are for comparing, and warm light makes every scheme look good, which removes exactly the discrimination the package exists to provide.

Step seven: add scale, and only scale

People, vehicles and trees belong in a concept image for one reason, which is to let a viewer judge size. They do not belong to create atmosphere.

A small number of figures at realistic density, positioned where people would actually be, does the scale job completely. A crowded plaza implies an activation nobody has demonstrated and quietly sells an outcome rather than describing a building.

The same applies to vegetation. Mature trees drawn at full size are a landscape decision and a fifteen year assumption, and showing them at concept stage commits to both.

Step eight: assemble the comparison, not the gallery

The package is presented as a set rather than as a sequence of individual images, and how it is laid out affects the decision as much as the images themselves.

Options side by side, at the same size, in a stable order, with the same annotation on each. Whatever varies between them stated in text as well as shown, because a viewer who can read the difference numerically will trust what they see.

Presenting them one at a time, in sequence, produces a recency effect and a first impression effect, both of which are avoidable by simply putting them next to each other.

The order the images are shown in, which nobody controls and everybody should

A concept package is almost always delivered as files and then presented by somebody who was not involved in producing it, which means the sequencing decision is made by whoever happens to open the folder.

That matters more than it sounds. Presentation order produces a first impression effect and a recency effect, and in a group setting both are strong enough to swing a close decision between schemes that are otherwise evenly matched.

The defence is to deliver the package as a single comparison layout rather than as separate files, so the options are seen simultaneously rather than sequentially. If separate files are unavoidable, deliver them with a stated order and a note explaining why the order is arbitrary.

It costs nothing to do and it removes a distortion that is otherwise invisible to everybody in the room, including the person presenting.

Annotating what the viewer cannot see

An image communicates volume, relationship and character. It communicates almost nothing about the numbers that actually differentiate the schemes, and a concept package that relies on the image alone is asking the audience to guess.

Each option should carry the handful of figures that define it: approximate floor area, unit count or key programme quantity, storeys, and whatever the specific trade off is between the schemes.

That combination is what makes the comparison rigorous rather than aesthetic. A viewer who can see that option B gains twelve units and loses the corner outlook is comparing a trade off. A viewer with only images is comparing preferences.

It also protects the package from a common failure, which is a group choosing the scheme that looks best and then discovering it was the one with materially worse economics.

The viewpoint that answers the question, chosen rather than inherited

Most concept images are made from whichever angle shows the building best, and that is a marketing habit imported into a stage where it does not belong. The viewpoint should be selected because it answers the question the group is actually arguing about.

If the argument is height, the viewpoint is from where the tallest existing neighbour is visible in the same frame, at eye level, so the comparison is direct rather than implied. If the argument is bulk on the street, it is from the pavement at the point of closest approach. If it is about how much of the site is consumed, an oblique from across the road showing the whole frontage does more than a three quarter hero view.

Choosing this way frequently produces an unflattering angle, which is uncomfortable and correct, because a flattering angle chosen by default answers a question nobody asked and hides the one they did.

It also makes the package defensible later. A viewpoint selected because it addresses the disputed issue is a decision that can be explained, whereas one selected because the building looked good from there cannot.

Two viewpoints per option beats one, and beats four

There is a practical sweet spot in how many views each scheme gets, and it sits lower than clients expect and higher than budgets usually allow.

One view per option is not enough, because a single angle can flatter or penalise a scheme in ways nobody can detect without a second reference. A tall thin building looks slender from one direction and dominant from another, and with one image the group is deciding on whichever direction happened to be chosen.

Four or more views per option produces volume without insight. The additional angles rarely reveal anything the first two did not, and they multiply the production across every scheme in the set.

Two is the number that works: one addressing the disputed question directly, and one from a different approach to confirm the scheme reads consistently. Beyond that, additional value comes from adding an option rather than an angle, and usually not even that.

What to leave out deliberately

Interiors, unless the internal arrangement is the thing being decided, because interiors at concept stage are almost entirely invention.

Night views, which are pure atmosphere and answer no comparative question.

Aerials as the primary image, because nobody experiences a building from the air and an aerial answers the arrangement question while skipping the experience question.

Branding, signage and tenant identity, none of which have been decided and all of which distract from the comparison.

And anything animated, because motion is persuasion and this stage needs the opposite.

The one exception worth making

There is a case for producing one deliberately attractive image alongside the plain comparison set, and it is worth naming because the request will come.

If the concept decision has to be socialised beyond the immediate group, to a board, a partner or a funder who is being asked for support rather than for a technical choice, a single aspirational image serves a genuine purpose.

The rule is that it is produced after the decision, of the selected scheme, and labelled as an impression rather than a proposal. Produced before, of one option among several, it corrupts the comparison it sits beside.

How the package should be reviewed

Review the reduction first: are these the right options, and has anything survived that should not have.

Review the setup second, on one option only, because if the viewpoint or the light is wrong every subsequent image inherits it.

Review the level of detail third, using the test of whether anything shown was not actually decided.

Review the set last, side by side, and specifically look for any way in which one option has been advantaged. That review is best done by somebody who does not have a preferred scheme.

What a good package feels like

It is slightly boring to look at, which unsettles people who commissioned it and is the correct outcome.

The differences between the options are immediately visible and everything else is identical.

Nobody in the room asks what material that is, because the images do not claim to know.

And the conversation that follows is about the schemes rather than about the pictures, which is the only measure of whether the package worked.

Frequently asked questions

How many options should a concept package show?

Two to four. A group presented with seven does not compare seven, they form an impression from two or three depending on presentation order. The reduction is a judgement about the decision and it usually has to come from whoever owns the project rather than the design team.

Why must every option use the same setup?

Because anything that varies other than the schemes decides the outcome. One camera position, height, time of day, sky, material treatment and level of detail. Each individual deviation is defensible alone and in aggregate becomes a distortion nobody can see.

How much detail should each scheme have?

Only what has been decided, which at this stage is usually massing, orientation, approximate height and relationship to the street. The test is whether somebody could point at any element and ask when it was decided. If it was not, it should not be there.

Why light concept images neutrally?

Because light does more persuasive work than anything else, and warm light makes every scheme look good, removing exactly the discrimination the package exists to provide. Expect the flatter result to be questioned and explain it before delivery.

Should the package include one attractive image?

Only after the decision, of the selected scheme, and labelled as an impression. Produced beforehand alongside several options it corrupts the comparison. Afterwards, for socialising the decision to a board or funder, it serves a genuine purpose.