When does creative technology earn its place?

An object follows the cursor. A headline breaks into particles. A photograph bends as the visitor scrolls. A page opens like a machine, a room or a landscape. The technique is impressive. For a few seconds, attention is guaranteed.

Then comes the uncomfortable question: what did the interaction actually communicate?

Creative technology has expanded the language of the web. We can manipulate images in real time, compose scenes with shaders, build spatial interfaces, let visitors reshape a narrative and make information behave rather than merely appear. It would be a mistake to dismiss these possibilities as decoration.

It would be just as mistaken to use them indiscriminately.

The interesting work begins when the behaviour of a digital experience becomes part of its meaning.

At SO.., we’ve been exploring this question through the design of our own online portfolio. We wanted the case studies to demonstrate something about each engagement, not simply present a sequence of beautiful assets with a technical flourish at the top.

An important distinction: these interactive case-study presentations were developed later, for the SO.. website. They are not being presented as digital installations or interactive deliverables produced for the original clients. The campaigns themselves had their own original creative concepts, imagery and formats. The website is our subsequent interpretation of that work.

That separation lets us ask a useful design question: what kind of interaction genuinely helps a visitor understand a piece of work?

First, identify what should change for the visitor

Before choosing a technology, describe the intended change in plain language.

Not “the user will see a cool distortion effect”.

Instead: “the visitor should understand that a harmful action can leave consequences after the original moment is no longer visible”. Or: “the visitor should discover that a single label conceals many different lives”. Or: “the visitor should see how an observation becomes a strategic recommendation”.

These statements point towards different digital behaviours.

One asks for persistence over time. Another asks for comparison or revelation. A third asks for a relationship between pieces of information.

Only after defining the desired experience should we ask whether that calls for WebGL, CSS, SVG, video, a carefully composed still image or a simple change of words.

Technology has no inherent obligation to be spectacular. Its obligation is to be appropriate.

A shadow that refuses to leave

The #SadZnaš campaign, developed for the 2025 edition of 16 Days of Activism against Gender-Based Violence, used shadows to communicate aspects of digital harm: anonymity, hidden behaviour and consequences that persist beyond the screen.

The original campaign expressed that concept through photography, film and graphic materials.

When we later developed the campaign’s portfolio page, a conventional gallery would have shown that material accurately. But it would not necessarily have conveyed the central metaphor as an experience.

So the website explores light and persistent shadows. A visitor can influence what becomes visible. A person fades; a shadow remains. Other figures enter, and their traces accumulate. The interaction is built around a temporal idea: something stays after its apparent source disappears.

This is a reasonable use of real-time rendering because persistence is the argument. The visitor doesn’t merely read about an afterimage; the interface gives the afterimage a duration.

There are limits. The shadows are conceptual, not a suggestion that the real campaign participants are perpetrators or survivors. Visitors should not be made responsible for the violence being addressed simply because they move a cursor. And nothing essential should depend on owning a powerful device or performing a precise gesture.

Those ethical and functional constraints belong to the idea itself. They are not an afterthought added by an accessibility checklist.

The clothesline does not need a shader

The previous year’s campaign, #SadZnam, used domestic objects and public-space photography to question the expectation that gender-based violence should remain hidden. Its clothesline motif was physical, recognisable and located in the streets of Sarajevo.

The online case study takes a different technical approach. A line travels through the page. Photographs appear as prints attached to it. Elements respond with a little physical movement. At moments the line disappears and the original street photography takes over.

A heavy three-dimensional scene would not automatically improve that experience. It could undermine the documentary material by placing the technology between the viewer and the photographs.

The better choice was to let layout, typography, scroll and restrained motion carry the physical metaphor.

This matters because designers sometimes confuse innovation with technological escalation. If one page uses a shader, the next must use two; if one interaction bends an image, the next must bend an entire room.

There is no reason for that progression. Different subjects require different levels of mediation.

The strongest technical decision may be the one that allows an original photograph to remain entirely itself.

Sometimes the interaction is the method

For the OSCE case study, a cinematic metaphor would have been less helpful. The engagement concerned a digital-ecosystem assessment, communications recommendations and staff training. Much of the underlying analytical report was not appropriate for public disclosure.

The challenge was therefore to show strategic judgment without publishing privileged findings or pretending that invented examples were real client data.

The portfolio experience responds through an interactive reasoning model. Visitors can move between observation, implication and strategic response. The relationship between these steps is the content.

Here, interaction does not simulate an actual OSCE dashboard. It makes a way of thinking visible. The examples are illustrative, and the page says so.

We can learn something from that distinction. A transparent reconstruction can be more honest and useful than an attractive screenshot stripped of the context necessary to understand it.

It also creates room for a quieter technical vocabulary. The animation can be precise because the ideas being compared are precise. There is no need to make information spin through space merely because a browser can render it.

Sometimes the best effect is a better sentence

The UNFPA Not Inactive case study begins with a labour-market label: inactive. The page then introduces the women who are too easily reduced to that category, including twenty-five original participant statements from the campaign.

Its essential move is editorial. It changes what we are looking at, and therefore what we are being invited to consider.

A label can be crossed out. But a caregiving role such as mother or housewife should not be theatrically erased. The problem is not the existence of those identities. The problem is treating them as evidence of diminished worth or opportunity.

No algorithm solves that judgment. A complicated animation might even make it harder by giving equal visual treatment to things that carry different meanings.

The useful interaction lets people encounter the voices, filter by themes without hiding anyone, and see the original cards. The content has authority because the words belong to real participants and remain legible.

Sometimes the most advanced part of an interaction is the decision not to interfere with what a person said.

An interaction should pass five tests

When an idea is technically exciting, we find it useful to separate ambition from justification.

1. Meaning. Can the team explain in one sentence what the interaction helps visitors understand, notice or remember? If the only answer is “it looks impressive”, the idea is not finished.

2. Integrity. Does the experience respect the material? Faces, testimony, statistics and documentary images are not interchangeable graphic textures. Some can be manipulated symbolically; others should remain unchanged. Consent and context matter.

3. Agency. Is the visitor’s involvement meaningful? A cursor effect that does the same thing regardless of input is often merely animation disguised as interaction. If choice matters, the effect should respond clearly to it.

4. Access. Can someone reach the same essential understanding with a keyboard, on a smaller screen, without precise pointer movement or with reduced motion? The alternative need not be visually identical. It must preserve the substance.

5. Cost. What is paid in loading time, battery, memory, maintainability and attention? Is the gain worth that cost? Could a video, diagram or better composition achieve most of the effect with fewer consequences?

These questions are not a scoring system. They are a way to keep the technology in conversation with the idea it was supposed to serve.

Performance is part of art direction

A scene that looks beautiful on a development machine can feel broken on a mid-range phone. A pinned section that seems immersive with a trackpad can become a trap on touch. A motion effect that is charming for five seconds may become exhausting when it repeats through a long page.

The artistic question is therefore not just what happens at the ideal frame rate. It is what happens when the environment is ordinary.

There are practical ways to protect the experience: pause rendering when the scene is offscreen, cap pixel density, compress assets, avoid loading multiple heavy films at once, make interactions optional rather than compulsory, and supply a meaningful non-WebGL version.

Some compromises are technical. Others are editorial. If a section needs an excessively long pinned scroll to explain one idea, the problem may not be the animation engine. It may be that the scene is too long.

Reduced motion should not mean reduced meaning. A well-composed still sequence can communicate the same conceptual progression without replicating every movement. A text transcript can preserve a film’s argument even when playback is unavailable. The test is whether the visitor can still understand the story.

Accessibility and performance are often described as constraints on creativity. In practice, they ask creative teams to clarify which part of an experience is essential.

That is valuable discipline.

The portfolio is not the project

There is a temptation, particularly in agency work, to make a portfolio page more elaborate than the original project and then let the two blur together.

A beautifully constructed interactive archive can demonstrate the agency’s design and development capabilities. But it should not retroactively claim that an original social campaign included a real-time installation, that an institution used a simulated dashboard, or that a client commissioned a website experiment that was created much later.

We think that distinction makes the portfolio more interesting, not less. There are two layers of craft on display: the work originally delivered for a real brief, and the editorial decisions made later to help a different audience understand it.

The two can speak to each other without being misrepresented as the same thing.

Know when to stop

Not every page needs a signature effect. Not every section needs to react when the cursor moves. And not every strong idea becomes stronger when converted into a miniature application.

A page has a rhythm. There are moments for surprise, moments for reading, moments for evidence and moments for stillness. If everything moves, movement stops being a meaningful distinction.

The ambition should be to give each story the form it needs. Sometimes that calls for real-time image rendering. Sometimes it calls for a physical scroll motif, an interactive analytical ledger, a transcript or a large photograph with nothing on top of it.

We do not become more creative by rejecting technology. We also do not become more creative by defaulting to it.

Creative technology earns its place when removing the interaction would remove part of the idea.