3 out of 4 startups will never make it. Even fewer will become breakout successes. But the moment growth turns the corner of the hockey stick, it is time to grab a hold of the bull rope and try to avoid the horns. Overnight, an apparently stable service can become a hydra as hoards of new customers probe into its weaknesses. Engineering teams begin reacting to new data and modifying the service appropriately, like the Borg Adaptive Shield Matrix (yes, I had to Google that).

Then reality sets in. Resources haven't starting coming in. There aren't enough bodies. It isn't possible to hire fast enough. It is extremely difficult for any organization to fight a land war in Asia… and to prevail on all fronts simultaneously.

Fortunately, scale growth can be hacked!

Up to this this point, most infrastructure has been a series of projects or individual solutions. Most IT deals are modeled, pitched and sold this way. But you don't need a point-in-time solution anymore, or a data center made up of a bunch of little deployments. You need a life cycle to continually deliver capacity while you are under load. Life cycles are built on standards, and standards are built on generations.

First Generation Standard

Start buying the same thing: server after server, rack after rack, datacenter after datacenter. This includes the network. It won't be terribly efficient, but it buys time to engineer improvements and limits new issues from being introduced.

Outsource production of this first generation to a trusted partner or VAR. Oversee production to ensure consistency, but rely on a partner to focus on the day-to-day manufacturing and delivery.

Stack up changes for the next round. It is tempting to want to fix problems the moment you find them, but this quickly devolves into a game of Whack-A-Mole.

Second Generation Standard

Size your app and build to it. Once you know what resources each component of the service needs you can build a scale model of the service so costs become predictable. Cost per customer will begin to drop because resources are more accurately matched to needs. Your finance and business planning teams will love you for both of these.

Third Generation Standard

You finally get to have some fun!! Now that you know what resources your service requires, you can focus on optimizing the least efficient parts.

Now that you have a delivery engine cranking out capacity, you have the time to evaluate, test and operationalize more complex and innovative changes without introducing risk by throwing it into production willy nilly.

You should even have the flexibility to insource certain components of manufacturing or delivery if they add a strategic value to your business.

Generations @ ServiceNow

Our first generation standards were focused on scaling delivery by standardizing and outsourcing. This allowed us to focus on the second generation, in which we improved performance per dollar 40% by sizing intelligently. And both of these gave use a stable platform for the next.

The third generation is where we really unlock serious value: 16X performance, less than 50% the footprint, reducing cost by 40%. We are just beginning the broad rollout but customers are ecstatic about it so far.

After writing most of this, I stumbled across a Bill Gates Reddit Ask-Me-Anything where he discusses working with TerraPower's fourth generation reactor to provide safer, cheaper, cleaner power. It's striking to note that the same generational concept can be used to safely drive innovation in areas that are even more risk averse and protective of their SLAs than we are.

Backstage @ ServiceNow

Dan McGee's #KNOW14 keynote on May 1 will dive behind the curtain and give you a broader backstage tour of ServiceNow so that you can understand the other ways we develop, monitor, manage and support customer instances. It's been a wild ride, and I am excited we get to share our perspective with you.

Not coming to Knowledge14? If this sounds right up your street then it's not too late to register for Knowledge14. It'll be informative and fun; what more could you ask for?