To help people and organizations to better internalize some key concepts and opportunities in digital economy, we are doing small series of posts on the following concepts; API Economy, Data Economy and Platform Economy.
We start with API economy and explain the concept using few physical world analogies, starting with vending machine business.
While reading this post, imagine vending machines being pretty much like any software application. Also it does not really matter if an application would be running as local application on a PC or provided as an internet application (SaaS) at this point, as that’s more about how these different application are made available to you.
Now, just like in any application, in a vending machine you also have a digital or physical User Interface (UI) with different buttons and instructions that you interact with, to make request for product (value) for what you want the vending machine to provide for you.
But different to software application alone, since vending machine have physical limitations, it can only have limited selection of products you can buy from it.
This selection of products is typically chosen based on the location where the vending machine is, what type of customers are passing by and what type of “wants and needs” these customers are expected to have. If the location is high traffic with many types of people passing by, the products can be pretty much anything, while the overall quantity of product that can fit inside is still a limiting factor. Also depending if there are other vending machines in close proximity, then additional things factor in as well for what the selection should hold, and what should be the prices of products in the selection.
Other factors may weight in as well, especially if there are same product available in same price, most importantly how the vending machine works and how quick & easy it is to use (user experience), what payment options are available and how are the products served. Ie if you are buying a coke, you most likely want it to be served ice cold. More competition there are, the more these things matter.
Vending machines, just like any software application, are automatic “machines” serving customers 24/7, but where both still are owned and need to be operated, maintained and developed by someone. Including fulfilling stock, making adjustments to the product or feature offering, maintenance and improvements for the system etc.
All applications aimed for people (physical or digital), have a user interface that is used to access “the product”, being that digital service, features, knowledge, information, candy, coffee, song, movie etc. In case of digital, where not having same physical limitations like vending machines for products that could be offered, there are still some other reasons that can limit the overall product offering or categories of product offerings that one single “application”. - Mainly due user experience (how to communicate the offering, logical context for what are relevant, how to access the value etc.) along with related costs associated to making those items available, Ie. building the features, designing user experience (UX), maintaining the offering etc. especially in cases like music or movies, where application maker also have limitations to how to get some of this content into their applications offering from the owner of the rights of those items.
Now, having established this general analogy between software application and vending machine, let’s explore the API Economy in the context of vending machine business.
Let’s start by imagining two vending machines by two different owners being side by side in same location, like at a central railway station. One is selling cold sodas and another one is selling chips.
Now if a customer, let’s call him Mike, would like to get both chips & coke, Mike would have to make two different purchases using two different vending machines.
He needs to do two full separate transactions taking at least twice as long. If Mike is using both machines the first time, he also needs to learn how to use both of these two different systems, potentially even having different payment options as well (ie two different user experience).
This scenario make it very easy to understand why Mike would rather just use either one of the machines to get both, chips & coke. Also if he already knows how to use one of the machines already, that he would most likely buy both items from that machine if possible. Or from either one of them, due reasons like; one have someone else using it and another one have not, available payment options, providers brand, loyalty program, better user experience etc.
But to make this happen from vending machine providers perspective, it would most likely mean getting a more expensive vending machine with abilities to provide both. And regardless due physical limitations, this would impact the size of inventory per item and/or number of variations of products it can carry. It would also lead to shorter refill cycles that would increase overall business costs.
Now, what if the soda machine could directly offer and sell chips from the other vending machine? It’s pretty obvious that a machine being able to sell both items with one transaction, would gain more sales from customers wanting to buy both or additional available items, without losing any of those customers who only want those items already available. In case both machines would sell both machines products, then the one with better user experience or some other similar factors could potentially win more customers.
It's also quite easy to see that overall both vending machine owners would benefit, simply due overall time saved by all customers, where more customers could be served faster with single transactions for combo purchases. Also, now both could also promote these combo deals for increasing their total sales.
In addition, both vending machine owners would now also get more information about their customers selections beyond their own product selection. Ie from other machine and also from combinations of purchases made that could also help optimize own product offering or introduce and test new product offerings, by analysing this data.
Now - if we only think about these two vending machines and their owners having understood these benefits, between the two, they could quite easily agree on business terms, on how the money between sales would be divided. Like sharing 10% of the revenue from sales via other vending machine sales.
Also, it would be quite easy for them to technically connect their vending machines to work with the other machine. In such simple setting between only two of them, they could simply choose to create digital connection that would work in their setup.
But let’s imagine an owner of a third vending machine comes along and sets up one more vending machine along side with the other two machines. This time selling candies.
After having understood how the soda and chips vending machines have been connected to work together, the owner of the candy machine would also like to connect his candy machine with the other two machines. It could also be that before having even realized the setup of the two other vending machines, one of the owners of other machines approach to suggest that candy machine could join with their setup as well...
Or perhaps all sides would be sceptical of actual benefits. Especially depending on the revenue share model:
But again, - it’s pretty easy to see the benefits purely from customers perspective and also from perspective of overall sales, additional information, etc.
Chips and soda machine owners could also conclude, that starting to sell candy in one location, where the candy machine vendor is already specialized in candies and perhaps already have multiple locations. So all things considered, learning the candy business, getting access to best wholesale deals etc. is most likely not worth their efforts. So better option would be to work together.
Finally they all agree to connect the third vending machine with the two others as well.
However, now instead of the previous case in connecting two machines, where each machine had to make just one connection to make it work, - now both soda and chips machine needs to have one additional connection and the candy machine owner needs to create two separate connections, one with each machine.
But having completed their tech work, it's now possible to buy chips, coke and m&m's or any other similar or partial combo, from any of the three vending machines and simply pick items from each vending machines outputs.
However, now each vending machine owner also need to maintain two custom connections each. Six in total when combined.
I guess by now you already have figured out how this story would continue…. Yep, an ice cream vending machine comes along, next comes coffee machine and so the story could continue.
With each new vending machine coming along, the number of connections to be created and maintained between all vending machine would multiply exponentially. And so would the complexity of deals to negotiate and connections to maintain.
Finally, together vending machine owners start to think - “there must be a better way, how these connections should be done? We should agree on standard interface and terms to be used in connecting with our machines (applications) and how new machines can be connected with other machines, simply by their owners agreeing on these standard terms and model of connecting”.
And that is exactly what application programming interface (API) is, what API's enable and how those are used in business context.
But to make APIs really work in business, - just like the user interface and user experience are just few key parts of a puzzle for a normal customer; a more complete set of things need to be created to support application programming interfaces (API’s) as well.
In this vending machine business scenario, it would mean things like; documentation explaining the use and functions of the APIs, data models about the information that is exchanged via these APIs, etc. But also the business “rules”, general terms and conditions for using those APIs & related information, business terms like cost of use, revenue share of product sales, along with payment terms, etc.
And finally, also a basic agreement between connected vending machines owners, on connecting a new vending machine with existing vending machine(s).
All these elements combined, enable broad networked value exchanges between API enabled applications globally, - also referred as API economy.
But even after all this, - after all the vending machines in this story would be using the exact same APIs, related rules etc. to connect with each other, due the network effect, there still would be another problem to fix. - The exponentially growing number of connections between each new vending machine added to the network.
So again, a better model is needed to make it really useful and scalable. A model where each vending machine would only need to enable one connection, to be able to then connect with any number of other vending machines.
To make this scenario really work, a separate “switchboard” application would be needed, that would then make it easy to manage connections between all connected vending machines. A system where with single connection, any number of connections with other connected vending machines could be enabled and managed, and transactions followed.
That would be something like “system and operator of vending machine connections”.
In this setup, it would be quite natural for each connecting vending machine owner to pay for the operator to not have to worry about all existing and future connections. It’s also quite clear that it would cost much less for any of the vending machine owners compared to building and maintaining all those connections individually. Perhaps owners of the vending machine businesses would even collectively decide to create a new jointly owned entity to create such operator, if none that would fit exists.
It’s also logical that such operator business would no longer be really a vending machine business, but something quite different. Where in addition to knowledge about vending machine business, additional knowledge about running an API connectivity business and technical operations is becomes very important. In ways, the more complex the industry is where such connectivity operator would enter, the more relevant the combined knowledge and connections in related businesses are.
While this connected vending machines example have worked quite well so far, due to familiarity and simplicity, it does not really scale far enough due to physical limitations. So let’s continue with another example to explore how the slot machines, “one-armed bandits” have evolved over time.
Initially slot machines were single machines operating independently, limited with their winnings, due to physical limitations like how much money in and money out those could hold. In any business it naturally does not make sense to pay out more than how much money is put in, but also physically they needed to have the money in to pay out the wins.
However step by step over time, both the limitations of payouts disappeared by other means of payment but also by starting to connect the slot machines with each others in a casino setup, next between casinos and eventually even beyond. Making it possible to build up huge jackpot wins by taking small percentages from each slot machines incoming money, that could then be won from any one of the slot machines that were part of the system.
There are no actual business or technical reasons limiting how far such network of connected slot machines could be expanded. I.e. to even start connecting these networks of connected slot machines and/or even other types of systems, like online slot machines and other gambling systems. The limitations are more on the regulatory side.
While regulation is used to limit some businesses to protect consumers, - in traditional banking on the other hand, - regulator in Europe have actually had to step in to regulate so that banks have to start providing public API’s to force banks to open up, to help improve innovation and customer experience in banking industry and to make it develop faster (I.e. PSD2 regulation by European Commission).
At the end of the day, in these examples of vending machines, slot machines and digital applications like virtual slot machine. They are all applications with user interfaces, where customers access the value provided by each application, but also where the collective value of these application can be expanded by network effect and extended to all customers, utilizing API’s.
And this is exactly what the opportunity is in the API economy, for all types of businesses and developers of existing and new applications and services.
When APIs exist, and when access, terms, business rules and documentations are made publicly available, connections and even totally new applications can be designed and build on top of such API’s. By having increasing volume of items, functions and data available behind available API’s things can be more easily utilized for innovating new value and services.
It’s common practice for innovators and developers to mix APIs from multiple different systems and applications, to build whole new services and applications based on APIs even cross industries and/or even between countries. Often in ways that would not had occur to those providing the APIs in the mix. But since having existing business rules as part of using their APIs, they also gain new business, revenue and growth from their own application as well (depending on terms created).
Just as in case of the first two connected vending machines, as the overall value of the business grows, so does the value of each of the vending machine added to connected network. But most importantly, so does the value that customers receive as well. Except perhaps in case of slot machines. - Where, while the overall value does grow, gambling is still not really a good deal for players, but that does not stop them seeing the value as something better ;)
Hope this post was helpful in understanding the APIs and the broader API economy. While these were potentially useful examples, these only cover part of the benefits in the API economy.
With more connections and more information about all the individual transactions in each connected application in each network, - combined with related information like locations, user profiles etc. - not only is this data very useful for developing new features or applications. - But this data is also very valuable, reusable and growing digital asset. An entirely new “raw material” that can be also packaged and sold in many different ways to enable totally new data revenue streams and even new data businesses. And this is what a data economy is about. But that is a topic for a whole other post.