Capacity Management in the Second Machine Age
Hold on to your hats, lil’ ol’ Joe is about to get all technical on you …
Demand for IT is unpredictable in what is now a digital, consumption-led economy. Put simply, there’s no such thing as a demand year, no such thing as seasons, and to use the oft-touted consultant’s cliché – change is the only constant.
The familiar, intuitive economics of demand and supply are broken – with the marginal cost of adding new customers and revenue trending to zero. And the easier things are to consume, the more they will be consumed, driving up demand without driving up costs.
Thus the traditional method of IT capacity management – “forecasting demand as a way to plan supply” – needs to change. Because while the economics are different, the business need for IT capacity is as important as ever. It’s a conflicting and confusing situation, so how are you supposed to “do” capacity management in what has been termed the Second Machine Age?
Frictionless Demand and Supply Break Traditional Capacity Management
We’ve been talking about the Digital Economy since the writings of Tapscott and Negroponte in 1995. Twenty years later, after the bubbles and crashes, the digital economy is in your front room, in your wallet, and all over your desktop. Organizations need to embrace the digital economy, and embark on platform strategies in order to stay relevant in a time when more companies will die at an early age than ever before.
These disruptive changes are not only technological, they are also organizational – with IT service management (ITSM) impacted in many ways, one being the ITIL*-espoused process/activity of capacity management.
The Issue with Capacity Management
Firstly, human beings are pretty bad at forecasting pretty much anything even when the calculations are easy. Whether it’s the economy, the size of markets, or how much consumer X requires from supply Y, the financial market still says: past results are not a predictor of future performance.
The real issue though lies in the assumptions underlying any IT capacity planning model, and it’s these assumptions that are disrupted when your business enters the digital economy:
- The demand model is different. In this consumer and consumption-led digital economy, we open up our business to unpredictability – as the business is now operating as, or offering, a digital platform that’s not just for traditional customers (who are often unpredictable) but also to other businesses that can build value on top of our value for their customers (unpredictable to the power of n).
- The supply model is different. Technology per se gets better and better, faster and faster, and easier and easier; and we probably won’t know what supply looks like even six months out in terms of cost, performance, and features.
So how is capacity management done in this Second Machine Age? The answer is, of course, “by machines” (but maybe not in the way you might think). So it’s all about machines, but first you have to change your thinking.
Expect to Fail, and Plan for Getting the Capacity Question Wrong
Imagine this simple but exaggerated example, where you have one server and ten customers. Based on past results and forecast demand, you plan to satisfy that demand with the same server. But what if you get it wrong? What if you get another ninety customers? Do you:
- Refuse to handle more customers (an actual tactic in place with some businesses today, turn away demand because supply can’t manage, and good luck presenting this to the company board), or
- Ramp up supply (more servers) to cope with demand.
The answer is B, right? Well, I don’t think that’ll cut it. It takes you a month to add another server (from requirements, through sourcing, testing, building, to deploying). Those ninety customers won’t wait a month. So what do you do? Perhaps your crystal ball foretold in advance that you would get another ninety customers, so you pre-provisioned another nine servers. But then they didn’t turn up, so you overspent.
The real answer is C – you don’t know what the future looks like, you’re not a soothsayer, and if you value your career you won’t make “punts” on capacity that cost the business either in lost customers (affecting both brand and revenue) or an unnecessary overspend on budget. Therefore you must only plan for uncertainty (my answer C).
Let the Machines Take the Strain
In the Second Machine Age, we don’t have much choice – capacity planning must be like this (planning for uncertainty) in order to make the business happy by increasing revenue at an efficient cost (the desired outcome is from capacity planning).
So what do you need to do?
- Don’t put any restrictions on demand; in fact actively seek more demand through more channels.
- Implement a platform strategy (open up the platform, let others build value on you) while being fully aware that this will increase the uncertainty in demand.
- Build a flexible supply architecture by backing off to cloud providers using a consumption-led procurement strategy.
- Invest in a data culture and experiment with new ways of getting insights from your ITSM data.
I’m not saying that it’s going to be easy but it will ultimately be a necessity once your business steps into the digital economy.
* The ITSM best practice framework
Posted by Joe the IT Guy