Not every idea needs to become a business immediately. Not every project needs a business plan, an investor deck and a precisely calculated ROI. Sometimes it is worth making something simply because the idea is interesting, strange, beautiful, funny, technically challenging, or because it refuses to leave you alone.
From a business point of view this can look unproductive. Why invest time if you do not know whether anyone will pay? Why build an experiment that may be used by only a few hundred people? Why learn a technology that has no obvious place in your main work today? Because the value of an experiment is often impossible to predict in advance.
A side project can become a new company, the first prototype of a larger idea, a marketing channel, the beginning of an audience, a job or collaboration opportunity, a portfolio piece, a way to learn a new skill, training for creative thinking, or simply a place where you can test ideas without a client's restrictions. Sometimes it becomes none of those things and remains a small, finished experiment. That can still be valuable.
What is a side project?
A side project is something you build alongside your main job, business or studies. It can be a one-evening prototype or something that develops over many years.
It might be a small web tool, an app, a game, an artwork, a newsletter, a podcast, a data visualisation, an AI experiment, an open-source library, a community project, a new service model, or a completely absurd idea that does not yet have a category.
The main difference from a conventional business project is freedom. A side project does not have to prove on day one that it can make money. You can begin with a simpler question: what would happen if I actually built this?
In business we often try to reduce uncertainty. In an experiment we deliberately enter uncertainty because that is where new information appears.
Why companies need experiments
As companies grow, they naturally begin protecting what already works. Budgets, processes, managers, targets and accountability appear. Every new idea may need several approvals. This reduces mistakes, but it can also suppress ideas that initially look too small, too strange or too unrelated to the core business.
Experiments serve a different purpose. They let a team test an assumption quickly, try a new technology, observe unexpected user behaviour, explore ideas outside the current roadmap, give initiative to people closest to a problem, and learn with relatively limited risk.
The goal of an experiment is not always to “win”. Its goal is to create new information. If five out of six experiments go nowhere but the sixth reveals a new category of project, the combined investment can still be worthwhile.
Google's “20% time”: not quite Friday, but the right principle
Google is often described as having special Fridays when employees could work on their own projects. A more accurate reference is the company's famous 20% time.
In Google's 2004 founders' letter, the company wrote that employees were encouraged to spend part of their time - conceptually around 20%, or roughly one working day a week - on projects they believed could most benefit Google. Gmail and Google Desktop Search were among the examples associated with that culture.
It was not simply a paid hobby day. The ideas still had to be relevant to Google's possible future, but employees had more freedom to choose what to investigate. The important part is not the exact number “20%”. Not every company can afford a full day each week, and in practice such programmes are never as simple as the mythology around them.
The useful principle is this: if all available time is reserved for today's tasks, there may be no room left for tomorrow's ideas.
A small company or a single person can adapt the principle: one afternoon a week, one experimentation day a month, a 30-day prototype cycle, a fixed budget for small ideas, two finished experiments a month, or 10% of working time reserved for ideas without an immediate commercial target. The exact formula matters less than protecting the time consistently.
Side-project success stories
Not every side project becomes a large company, and that should not be the requirement. Still, there are many examples of small experiments that outgrew the thing that created them.
Unsplash: ten free photographs
Unsplash began very simply: a Tumblr page with ten high-quality photos that people could use for free. It was a side project of Crew rather than a grand platform launched with a complicated business plan. The project attracted attention, built a community and eventually became independent.
The lesson is not that every free resource becomes Unsplash. The lesson is that a tiny project can have a clearer proposition than a large company, a useful free resource can become marketing for its creator, and an audience can sometimes grow faster than the original business.
Twitter: a project inside a company doing something else
Twitter emerged inside Odeo, whose main business was connected to podcasting. The first short-status prototype was a small internal project rather than the company's central strategy. When it became clear that the idea had potential, it received more attention and later became a separate company.
The important point is organisational permission. The team had space to build something outside the core product. If everyone had been limited to improving only what the company already did, that direction might never have appeared.
Gmail and the culture of experimentation
Gmail is often cited as an outcome of Google's internal experimentation culture. The useful lesson is not “every side project should become Gmail”. It is that the first version of an important future project may look like a small experiment that does not fit neatly into today's roadmap.
You can only connect the dots looking backwards
In his 2005 Stanford commencement speech, Steve Jobs described taking a calligraphy course after dropping out of college. At the time it had no obvious practical application. Years later, that knowledge influenced the typography of the Macintosh. His point was that you cannot reliably connect the dots looking forward; you connect them looking backward.
Side projects create those dots. One project teaches you payments. Another teaches you to write clearly. A third teaches video. A fourth forces you to research biology, logistics or human behaviour. A fifth introduces you to someone with whom you build something entirely different two years later.
Individually, none of these experiences looks like a plan. Together they can become an unusual combination of skills. There are many programmers, many designers and many people with business experience. Someone who can combine technology, business models, conceptual thinking, communication and the ability to build a working prototype alone is much rarer.
My side project: absurd.website
My own example is absurd.website, an independent internet-art and experimentation practice that I started in 2020. I describe it with the phrase “Startups as Art. Products as Concepts.”
I build functioning internet projects in which startups, digital services, platforms, games, interfaces and business models become artistic material. It is not merely a portfolio of images or ideas. If I invent a fictional company, I build its website. If I invent an impossible service, it may have an order form. If I invent a speculative institution, it gets a system, rules and infrastructure.
Some works are internet art. Some resemble startups. Some are games, public experiments, AI projects, strange tools or projects that are difficult to place in a single category. Absurd.website is not a conventional startup, but it uses startup logic. It is not a traditional art portfolio, although art is at its core. It sits somewhere between a creative laboratory, an internet archive, a project studio and a long-running artwork.
The practice now contains more than fifty finished projects, with two new works produced each month - one public and one members-only. The important thing is not only the individual works. It is the rhythm and the archive. Every new project adds another dot, another skill and another possible future direction.
What has absurd.website given me?
1. I make projects I would never create if money were the only filter
If every idea first had to answer “who will pay for this?”, many of the most interesting ones would die before the first prototype. In a creative practice I can begin with different questions: Is the idea interesting? Does it say something new? Does it change how we look at technology or society? Can it be made functional? What happens if people can actually participate in it?
That freedom takes me to places a purely commercial process rarely reaches.
2. Every project teaches something
One project requires learning an AI API. Another requires payments, print production, data visualisation, game mechanics, video processing, automation or work with a physical object. Even if the project never becomes popular, the skill remains.
I also learn more than technology. A conceptual project often forces me into a new field: science, security, education, advertising, death, fashion, autonomous systems, artificial intelligence or human behaviour. Over time those areas form a wide network of knowledge and references.
3. I train the ability to finish
Most people do not lack ideas. They lack finished work. A rhythm of two projects a month teaches me not to wait for the perfect moment. I have to choose scope, make decisions, remove unnecessary parts and release the work.
Creativity without discipline often remains in a folder called “someday”. A publishing rhythm turns making from a mood into a practice. You learn to reduce an idea to something achievable, distinguish the core from decoration, accept imperfection, finish before the energy disappears, show the work publicly and start the next project without becoming trapped in the previous one.
4. I train non-standard thinking
In a conventional project the question is often “how do we do this better?” In an experiment you can ask different questions: What if we do the exact opposite? What happens if a business model is taken literally? What would be a project nobody should create? Can a service itself be an artwork? Can a fictional institution begin behaving like a real one?
This kind of practice later helps in commercial work too. When every competitor offers similar features and similar advertising, technically correct thinking is not enough. You need an idea that people notice and remember.
5. I learn things I would not learn otherwise
Learning in the abstract is difficult. A specific project creates a need. I do not learn a technology because it might be useful one day; I learn it because by the end of the month I need a specific idea to work. That makes learning faster and deeper because it has context, a deadline and a visible outcome.
6. An experiment can transform into a real project
Not every absurd.website work is intended to become a business. But some can reveal a practical application, an audience or a separate project direction. An experiment can evolve into a SaaS project, brand campaign, education tool, physical installation, research project, company collaboration, exhibition, paid digital project or a standalone startup.
Unlike an idea in a presentation, a functioning experiment already shows what the idea looks like in reality. People can use it, criticise it, share it or reinterpret it.
7. The archive becomes a portfolio, reputation and idea library
One experiment can disappear. Fifty experiments show a method. An archive communicates more than “I can build a website”. It shows how I think, which subjects interest me, that I can create original concepts, turn them into functioning objects, finish regularly and combine technology, design, writing, business and art.
8. A side project can become an audience
An audience often forms around repeated output rather than a single launch. People may first arrive for one strange tool, one useful article or one visual idea, then stay because they understand what kind of work they can expect next. That audience can later support completely different projects.
Side projects as marketing
A side project can market you without looking like an advertisement. A calculator, dataset, visualisation, game, public experiment or useful tool can travel because it has value of its own. People share the thing, and only afterward discover who made it.
This is often more powerful than saying “we are innovative”. Instead, you make something that demonstrates the type of thinking and execution you want to be known for.
Side projects as market research
Experiments can also reveal demand. You can observe who uses the project, what people ask for, which part gets shared, what they misunderstand, whether anyone returns and whether people request a more practical version.
That is real market information. Sometimes the experiment confirms that an idea should remain art. Sometimes it reveals a narrow group willing to pay. Sometimes it shows that the real opportunity is not the original project but one small mechanism inside it.
Creativity comes from volume, not only inspiration
It is tempting to imagine creativity as waiting for one exceptional idea. In practice, producing more finished attempts creates more chances for an exceptional idea to appear. The quality of the archive grows partly because there are more experiments, more combinations and more feedback loops.
The goal is not to publish random low-quality work. The goal is to make enough serious attempts that you stop demanding that every single one justify your identity or career.
Why experiments matter even more in the AI era
AI makes many forms of prototyping dramatically cheaper. A single person can create interfaces, code, copy, images, analysis and small automations faster than before. That means the scarce resource is shifting from the ability to produce a first version toward the ability to choose good questions, test them and recognise what deserves further investment.
When building becomes cheaper, experimentation becomes more valuable. The advantage is not having one idea and protecting it for a year. The advantage is being able to complete many meaningful tests and learn faster than someone who only plans.
When does a side project become avoidance?
Side projects are useful until they become a way to avoid uncomfortable work in the main business: talking to customers, selling, maintaining existing systems, finishing an important launch or making a difficult decision.
A useful distinction is whether the experiment produces information, skill, output or genuine energy - or whether it only creates the feeling of being busy. Ten unfinished prototypes can be avoidance. One small published experiment with a clear lesson can be progress.
How to start an experimentation practice
1. Choose a question, not a grand company
Instead of “I will build the next global platform”, begin with “can this interaction work?”, “will people use this strange service?”, or “what happens if I combine these two mechanisms?”
2. Set constraints
Give the project a time limit, budget limit or technical constraint. Constraints stop the experiment from expanding into a six-month product build.
3. Build the functioning core
Make the smallest version that demonstrates the idea for real. If the concept depends on interaction, make it interactive. If it depends on a transaction, simulate or implement the transaction. Do not hide the central mechanism behind a slideshow.
4. Publish
An unpublished experiment teaches only the maker. A published one also teaches you how other people interpret it.
5. Write down what you learned
Record what worked, what surprised you, what people did and which parts might be reusable later. This turns isolated projects into cumulative knowledge.
6. Build an archive
The archive is where the compounding happens. Years later you may discover that an old prototype contains the missing piece for a completely different project.
Do not force every idea to prove its value immediately
Business discipline is useful, but applying it too early to every thought can kill exploration. Some ideas need to be tested before their category, audience or business model is visible.
The point is not that economics do not matter. The point is that timing matters. A project can begin as a question, become an experiment, reveal an audience and only then deserve a business model.
Conclusion: if an idea will not leave you alone, build it
Not every side project will become a company. Most will not. But a side project can still produce skills, code, knowledge, relationships, reputation, an audience and future options. Those outcomes are difficult to model in advance, which is exactly why experimentation needs space.
If an idea keeps returning, the best way to understand it may not be to analyse it for another month. Make a small working version and see what happens.