Product Land (Part 1)
You can’t build what you can’t think of in the first place.
This is the first of a couple of posts about why I think we need better ways for thinking about the design of digital products.
Thinking in a linear way, in simple hierarchies or in timelines, is pretty simple for us; our brains evolved in a world with 3 dimensions (plus an additional one for time), but for more complex systems, systems with many dimensions, we tend to need tools and concepts to help us.
As an experiment, think about evolution for a few seconds. What picture do you have in your head?
Was it this one?
That image, from a 1965 popular science book, goes by the name of ‘The March of Progress’. It shows the supposed progression of the evolution of humans.
Or maybe the first picture that jumped into your head was of a fish hauling itself out of the sea using its fins as legs?
The idea of linear progress, of directional change, is a stong meme, but in evolutionary terms is also not very helpful. In Wonderful Life, a book that is part story of early life on earth, part the detective story of recreating the bizarre world of the Burgess Shale from 2.5 dimensional fossils, Stephen J Gould dedicates a whole chapter to detailing the damage it has done.
Maybe you had a branching tree in your head, with species on each branch? Though slightly closer to reality, it is still far from an accurate representation. Genetic diversity and evolution by natural selection are inherently multi-dimensional and as a result we need tools to think about them properly.
In The Blind Watchmaker, Richard Dawkins uses things called ‘biomorphs’ to do just this.
Biomorphs are computer generated critters created under a simulated evolutionary pressure. They can vary from each other in 9 ways (things like width, branching, height, colour).
Biomorphs exist in ‘biomorph land’, a many-dimensional space that contains every potential biomorph. Similar biomorphs are clumped together; radically different ones more distant. It is a space you can take a mental walk though understanding how patterns change over multiple axis. In turn, biomorph land is a tool for understanding a more complicated space:
There is another mathematical space, not filled with nine-gened
biomorphs but with flesh and blood animals made of billions of cells,
each containing tens of thousands of genes. This is not biomorph space
but real genetic space. The actual animals that have ever lived on
earth are a tiny subset of the theoretical animals that could exist.
These real animals are the products of a very small number of
evolutionary trajectories though genetic space. The vast majority of
theoretical trajectories though animal space give rise to impossible
monsters. Each perched in its own unique place in genetic hyperspace.
Each real animal is surrounded by a little cluster of neighbours, most
of whom have never existed, but a few of whom are its ancestors, its
descendants and its cousins.
Sitting somewhere in this huge mathematical space are humans and
hyenas, amoebas and aardvarks, flatworms and squids, dodos and
dinosaurs. In theory, if we were skilled enough at genetic
engineering, we could move though the maze in such a way as to
recreate the dodo, the tyrannosaur and trilobites. If only we knew
which genes to tinker with, which bits of chromosome to duplicate,
invert or delete. I doubt if we shall ever know enough to do it, but
these dear dead creatures are lurking there forever in their private
corners of that huge genetic hypervolume, waiting to be found if we
ever had the knowledge to navigate the right course though the maze. We
might even be able to evolve an exact reconstruction of a dodo by
selectively breeding pigeons, though we’d have to live a million years
in order to complete the experiment. But when we are prevented from
making a journey in reality, the imagination is not a bad substitute.
Biomorphs are used as a stepping stone to paint a picture of a huge hanger containing every actual and potential organism suspended in the air. A hanger you can take a mental walk though, a couple of axes at a time.
It works with other concepts too. Try and imagine a ‘musical-instruments-with-strings hypervolume’, imagine it has an axis for number of strings, one for volume, sound, methods of playing, tone, colour, country of origin and so on. We can only glimpse a couple of the axes at any one time, but along them we might see all the guitars and banjos in a clump and the bowed instruments as a cluster with the hurdy gurdy sitting just out from them, along with all the other instruments that have never been lying inbetween, some viable, some that would make your ears ache.
For a reverse example, try and imagine all the axes that existed in Jeremy Deller’s History of the World before it was squished into two dimensions (geography, musical genre, political systems, historical events, synthesisers).
What does this have to do with digital products though?
My proposition is that digital products are also inherently complex and inherently multidimensional, that design is too often constrained by our methods of thinking about them and too often risk being either derivative or simple iterations of variants as a result; or worse, user needs are never met as well as they could be because we are looking for solutions in the wrong place.
Products can vary along many axes - the degree to which they are active or passive, centralised or federated, specific or generic, solitary or social; they can vary in the technology used to build them and on which they are consumed, the power which they create or remove from users and the organisations we put around them. The number of potential products dwarfs the ones we ever conceive of, while many will be nonviable, some will meet user needs. We need ways of exploring the potential space products could occupy, tools for embracing the margins of potential products, tools for walking though product land.
Things like the much-tweeted Spotify diagram of how they build products are just too simple, too linear; they are the equivalent to the March of Progress (and almost certainly doesn’t do justice to the thinking the team actually put into their products).
Even the standard build-test-iterate-repeat loop diagrams feel unhelpful, since it is too easy to fall into the trap of picking one thing, one broad concept and iterating the hell out of it, making sure every feature is as perfect as it can be, without ever again questioning the core approach (or maybe that you need two or three coexisting products or 2 equally good ways of addressing the same user need, just as convergent evolution has solved flight in pterodactyls and parrots, fruitflies and hummingbirds). Product land remains unexplored, user needs unmet.
Some digital industries seem to be particularly stuck, shuffling around a single idea, the equivalent in animal space to assuming the only animal that could exist was a hyena (albeit with slightly different markings).
My favourite examples of this are the digital products we have at our disposal for finding work and finding a holiday (though there are many others). In both of these industries almost all the products seem to be clumped around one small area of product land. Some tools may be slightly better than others, but they are all essentially variants the same product.
Finding a holiday online almost always conforms to the pattern of choosing a departure airport, a destination and a price range, then seeing a list of results. So you are pretty well served by existing products if you know where you want to go from and to (and want to do it by plane).
Finding a job online consists of entering a search term into a centralised system and getting some results back in a list. The search is generally free text and searches against a subset of the jobs available online, the ones the product knows about, which is in turn a subset of the available jobs in society at the moment, since many (most?) jobs never make it online. There is a bit of variation, but most products conform to this pattern.
Better products for people to find the best holiday might not be that important in the great scheme of things (unless you run an online holiday finding service), but better tools for people to find work is good for almost everyone (people find jobs and get paid, companies find employees, government pays less in out of work benefits and the economy generally functions better).
You are currently pretty well served by existing products if you know the type of job you want and where you want to work, which generally means having some employment history. In short, if you don’t fit that description, the market fails you.
Part 2 will try and imagine some alternative viable products in these spaces and the sort of tools we can use to find the edges of product land.