Listen to the audio version on Spotify and Apple Podcasts.
Let us start with examples of very common types of questions when analyzing innovations:
(1) Hundreds, perhaps thousands, of employment models for persons with disabilities are launched every year, independently and without awareness of each other. What do successful models have in common?
(2) Are all development aid projects that support entrepreneurship for persons with disabilities similar, or are there substantial differences?
(3) There are many guidebooks aimed at making the built environment more accessible. Can they be treated as a single, consistent method, and should online guides also belong in this cluster?
It is all about clusters
For more than ten years now, the work of the Zero Project Team has been all about clustering. Clusters are deeply embedded in basically everything that we do: Defining what we are looking for in our nomination process. Allocating experts as competent peer-reviewers. Programming the Zero Project Conference Agenda. Structuring our publications.
We do not always call it cluster. It is not always the same process. And we do not always do it for the same reasons. But there is one unifying reason behind it: Finding these clusters, and getting them right, is crucial for understanding innovations and their potential to scale or be replicated.
Think of the Zero Project as a machine that is built for scaling in this regard: if the Zero Project cannot support the innovations that we selected, and we cannot get our community interested in them, there would be no reason to run this machine, and the whole team should do something else instead.
The underestimated part of every innovation-research
Clustering is rarely identified as a crucial component in innovation processes; again, not even by the Zero Project Team. Above all, the Zero Project Team looks at unique features, which we refer to as innovation, and we try to understand what is truly innovative in each context. We look at features that solve real problems, which is what we call impact. And we look at whether they are relevant for the work of others in our community; this is what we call scalability (as a broader term, covering both the ability to grow on its own, to be replicated by others, or simply to inspire others).
So, why is there a need to cluster innovations, and why is it a key component of innovation processes? In other words: Why do we want to understand what they have in common, what makes them different, and what makes them outstanding?
Starting again with the three examples used in the beginning: Why is it not enough to be sure that one specific employment model is extraordinary in itself – let us say this great job platform for persons on the autism spectrum in the UK? Or this wonderful microfinance support for women in Burkina Faso?
I give you a list of eight reasons, but I could give you many more.
1. Categories as a baseline for comparison
Innovations need categories (characterizations, properties, features, labels, criteria, tags, etc.) to describe and compare them. Which categories to choose, and which narrative – these two decisions are not only the baseline for clustering, but for all types of analysis of innovation processes.
2. Where the wheel is invented more than once
Clusters give indications as to whether innovative models are invented more often than once, and independently from each other. For decision-makers, this is important information. If a broader cluster exists, this is an indication that something powerful is going on here. In other words: the wheel is invented repeatedly here, and there is a reason for that. Decision-maker: Go, look and understand this cluster, and try to understand how you can use and adapt it in your context!
That may sound counter-intuitive: shouldn’t we instead be desperately trying to find the opposite —the unicorns? Those groundbreaking, game-changing, disruptive innovations? Clustering also helps identify these unicorns, which are usually technology-related: they are the ones that do not fit into meaningful clusters.
Of course, they are of great interest —but usually only to smaller groups, such as investors and early adopters.They are not yet ready for mass scaling, for various reasons.
3. Clusters are evidence, individual models are not
Individual models use their own language and their own arguments to describe who they are and what they are here for. They use a narrative to attract as much attention of their relevant stakeholders —customers, beneficiaries, service providers, policymakers, funders —as possible, because they have to. Innovators have to compete for attention, business, or funding to survive. Their narrative must persuade, because without objective clustering, good storytelling, marketing, luck, or high-level access often decides who succeeds.
Only with clusters of solutions, trends can be found and analyzed. For decision-makers to work beyond persuasion, luck and “old-boys networks”, they need clusters and also trends, even if if they may not even be aware of it. A politician is much more inclined to promote a solution if it has been deployed successfully in other countries, so the risk of misallocation of taxpayers’ money is lower. A manager is much more inclined to invest if she or he spots clusters or trends, because it means a growing business opportunity.
4. Innovating requires collaboration, and collaboration needs clusters.
Any organization, any individual with an interest in innovating has a need —and usually an endless number of opportunities —for collaboration. But only a few of them can be managed and an even smaller number makes sense, given the limited resources that always exist. But how to choose? Clusters help define both peer groups and market niches within existing products and services, which are the baseline for good collaboration.
5. Government support needs clusters.
Especially governments and their areas of responsibilities, such as licencing, tax-incentives, public funding agencies, and watchdogs, cannot work with individual innovations and small organizations. For a government to support innovations it needs an evidence-based understanding of new developments that improve the status quo. They need to hear a narrative based on clustered solutions (i.e. new developments) that are better than the existing clusters of solutions (i.e. status quo).
6. Good strategies need clusters.
For innovators themselves it is essential to understand their peer group, their market niche, and their own unique features (in advertising it is called USP, Unique Selling Proposition). Evidence-based clustering is a prerequisite for that. In early phases of development, with local stakeholders to persuade, it may still work to have one’s own unconnected narrative that does not refer to any clusters.
But at a certain level, when large and experienced stakeholders need to be convinced, that lack of evidence in describing the existing ecosystems —in other words: cluster-based analysis of challenges, markets, and potential —will be a glass ceiling for many promising innovations.
In my next blog post, I will explain the principles of clustering innovative solutions.
Follow on LinkedIn
This blog is available as bi-monthly LinkedIn Newsletter. Subscribe and engage in the conversation!
How this post was made:
This post was written, edited, and fact-checked entirely by Michael Fembek. To enhance the experience, we used Zero Project Responsible AI for data analysis and ChatGPT for final proofreading. The visuals feature original photography from the Zero Project Conference 2025, edited by Alessandro Gobello, and the audio version was created using an Eleven Labs AI model of Michael’s voice.