The Internet, best product ever!
This topic is a passion of mine and the conclusions I draw in this post are after over 20 years and working on perhaps hundreds of projects, some hugely successful, some, overpriced and woefully poor.
There are a lot of ground rules I am confident are so solid that I can see red flags within minutes of evaluating any initiative that I think is missing the core characteristic of success. This post is just going to look at one of these, and that is how deliverables are defined. Because anyone who thinks a product is just one big thing is holding a red flag.
To explore this I am going to simplify 1 successful product.
The Internet
I'm not sure there is any product more successful, I'd certainly like to hear one if you can think of one. Now, let's imagine that we were given the story card for The Internet.
As a user, I want to transmit (send and receive) text and images between computers in different locations.
Before I go into how this product was initially delivered and why even 40 years later it is still one of the most sustainable products on earth, I want to speculate as to how an engineering-minded techie might wish to deliver this. In a nutshell, they would want to do it all. I know enough, and all techies love nothing more than building stuff themself, but as they say, you can't boil the ocean and so to facilitate delivery you need to split the product up and that's what we call a feature backlog.
The Internet** is split by something called the OSI model, this is 7 features or layers, and back in the early 1980s, it was agreed that to deliver the internet, all the 7 layers could be developed by totally separate parties. For simplicity, I'll cover just 5 of the 7, rolling 3 into the top layer. The deliverables were, and the features today still are,
- Presentation
- Transport
- Network
- Data
- Physical
** More specifically, this is just networking and doesn't include servers.
Even if you don’t wish to know how the internet works, I’m sure you understand that you interact through the browser or email client and that between your device and one of the people which you communicate with, there are wires and stuff. Well, what you perhaps also understand is that the wires and stuff are provided by a telecoms provider, while the routing of information is provided by another service provider that uses devices predominantly supplied by one of the planets biggest ever companies, CISCO. I’m positive you’ve heard of IP addresses, and maybe even TCP/IP. All these are protocols and they provide connectivity, while HTTP and HTTPS are more protocols that present and encrypt information. Right now you are likely either bored, confused, overwhelmed or possibly all three. And that’s my point!
That’s as much technical information as I’m going to cover here, but the point is that every single technical solution mentioned in the above paragraph just works. Each is a feature of the internet, and at a point in the early 1980s, each was a concept and I expect was a deliverable on a backlog that was planned to be combined to deliver what we know today as a single product. The Internet.
What is most impressive in the life of the Internet is the foresight to make the internet incredibly sustainable. This is achieved because each layer in the OSI model is independent of the next. This was enabled by creating acceptance criteria for any solution that wished to deliver a layer/feature. Even today, engineers are looking to roll out 5G, but they don’t look to change the browsers for instance, or our beloved TCP/IP, all they will do is engineer the Physical Layer. What’s more, in order for it to work the MVP will need to at least meet the original acceptance criteria, but the gorgeous part is that they can also make marginal gains and evolve the product. What this means is that the Physical Layer can make small enhancements to the interface with the Data layer above, these enhancements are transparent and documented for everyone, and if agreed will be added to the acceptance criteria for the feature. Enhancements to one layer don’t necessarily have to be implemented immediately, but the providers of the interfacing layers can choose to either adopt or otherwise. Those that adopt success enhancements succeed, those companies which don’t evolve tend to lose market share and die. This is the wonderful evolution of the product which doesn’t belong to any single company and yet is the most successful product the world has ever seen.
It is my opinion that if you want rapid and sustainable delivery then these principles can and should be applied to all products.
Where did I begin?
You can’t boil the ocean. Well, maybe I mean, you can’t boil it alone. But if you build on a foundation of good principles and with the right people, good planning and solid backlog with agreed acceptance criteria, you can watch a cat fall into a bath in China from an airport lounge in Argentina.
Please discuss these ideas with me on Twitter @alexcrossley