Where would we be if Linus wrote an MS-DOS clone 20 years ago?

In response to my discussion on the Trouble with OpenStack, I get a lot of questions of the kind "Where would we be if Linus wrote an MS-DOS clone 20 yrs ago" or "Yeah, Linux should copy Windows APIs" etc.

Ok, I'm a bit tired of repeating the same old stories for what has been 6 years now but ... one last time.

You're mixing the world of products with the world of utility.

They are not the same.

The game is not the same.

The power of ecosystems is different.

With products you can't use consumption to determine novel and diffusing acts that have been built on your product, you have to use market research which is costly and slow. With utilities, you can use consumption and this is a huge sodding deal.

With consumption information, you can run an ILC model (innovate-leverage-commoditise) by being a first mover to industrialise an act, getting everyone else to innovate on top of it, leveraging consumption of the service to spot diffusion, being a fast follower to any successful change you've spotted and then commoditise to new components to grow the ecosystem.

By doing so, you can gain :-

• highly stable, high volume, increasingly low margin revenue
• maximum wealth generation by being fast follower to novel
• high rates of innovation (as everyone else is doing it for you)
• high rates of customer focus (by leveraging data in the ecosystem)
• high rates of efficiency (by focusing on this and economies of scale)

You have to manage it carefully because people will grumble about you eating your ecosystem but each time you do you provide new components which grow it.  Not only do you get all five effects but all five increase as your ecosystem grows which increases super linearly with the physical size of your company i.e. under a well run model, as you get bigger, you get more innovative, more efficient, more customer focused, more stable and more wealth generation all at the same time.

Have people really not twigged why Amazon isn't just winning the space it's accelerating further away from its competitors? Do you think it's random that they sometimes gobble up their ecosystem? 

Yes, you can grumble about Linux or MS-DOS but these were product plays. You couldn't play an ILC game effectively in a product world only in a utility because the consumption of APIs is what gives you the mechanism to do this. Oh, and there is absolutely no point in trying old product based feature differentiation strategies against it. 

You might as well pick up a copy of Porter's Competitive Advantage and go "how shall we beat Amazon? Do we do innovation, customer focus or efficiency?" because guess what, they're doing all three of these plus stable revenue plus maximising wealth generation simultaneously against you and growing all of this with the size of the ecosystem. 

So, sure, if you want to go and try a differentiate on APIs route then you better magic up a larger ecosystem rapidly which means you need to drop billions upon billions upon billions in the next six months and even then its doubtful. A much more sensible route is co-opting that ecosystem as old tales about product battles ain't going to help you one jot.

Finally, it's not like this game hasn't been known about for the best part of a decade. Wake up, smell the roses and realise you're not fighting in a product world. I'm sure you can find some analyst or flat earther to tell you that "it's just like Proprietary product vs Open product" or that "Ecosystems aren't important" but realise the game is different by simply opening your eyes and looking.

Blindly following a product mentality of feature differentiation in a utility world against a well managed ecosystem is not going to work out well for you.

--- Update 17th July

Just to comment, Porter's book was fantastic for its day and I still recommend it but the game has changed since then. Still it's a good read just the game is different.