Cost Efficiency And Warehouse-Scale Computers

“The Datacenter as a Computer” is a very detailed look at treating data centres as warehouse-scale computers. A massive data centre is treated as one big computer. A cool perspective for sure.

If you want a technical overview of how these massive data centres are put together, this is your weekend reading. There are also interesting points made about the economics of warehouse-scale computing.

One of the things that jumps out at me is how cheap IT infrastructure components are in the United States relative to New Zealand. Fixed costs cut both ways I suppose.

Jason Cohen Thinks Some SaaS Products Could Be Permanently Unprofitable

Via Hacker News: The Unprofitable SaaS Business Model Trap

Jason Cohen walks through how SaaS operations can get into a trap where they keep growing customer numbers but earning back the acquisition costs doesn’t happen because of cancellations and rapid cost acceleration.

I recommend reading the full piece because he breaks down how this could occur for a firm that doesn’t make it as easy as possible to join (low acquisition costs) as well as keep average contract length high. (retention through an awesome product)

If the payback period for a new customer is 24 months and you have a retention rate of 75% – each year you must acquire 25% of your starting customers from previously untapped sales channels who will be unprofitable for the first 24 months of the customer relationship.

Upselling and earning as much revenue from current customers as possible is a key strategy for shortening that payback period. The economics of SaaS demand constant fine tuning of acquisition strategy, a strong handle on cost growth and aggressive customer retention ideas.

Microsoft Takes A $1 Billion Charge On The Surface Tablet

The latest Microsoft earnings report included a US$900 million dollar charge related to the Surface RT. (H/T NBR Paywall).

Microsoft is trying to focus on enterprise IT because that’s where the money is.

Not every firm, particularly public sector organisations, can embrace the cloud.

It should be noted that in terms of Microsoft’s annual revenue – US$77.85 billion dollars – the charge is quite small.

It should also be noted that they paid a US$733 million dollar fine to the European Commission in the 4th quarter!

It is really easy when you live in a tech blog bubble to forget that most of the world’s IT spend is by large corporates, public sector organisations and other firms for whom it is actually not possible to seamlessly upgrade to each new version at a time.

There are substantial fixed costs that you incur when moving from a long term support version of a product to a new one – how much enterprise IT spend has been deferred because of cost-cutting during the global financial crisis, for example?

In terms of applying economic thinking, the cloud is great for small startups because they can defer large fixed costs like setting up an internal sever farm until they’re big enough to not worry about the cost.

Using cloud computing services, before adding a risk weight for the possibility of not having access to your app or data when you lose internet access, the marginal cost of using that capital productively is very low and getting lower because of enormous levels of competition in the space.

At the margin, cloud computing enables patterns of sustainable specialisation and trade to be found through trial and error by entrepreneurs in the tech space far faster than even 5 years ago.

 

Will Companies Drag Their Feet On The Way To The SaaS Cloud?

Software as a service is all the rage. But not all companies are rushing towards cloud computing. Some can’t even take advantage of cloud services because of legal or client confidentiality concerns. “Microsoft gets that Windows is done. But it’s betting companies will drag heels on way to the cloud.” is worth a read.

Not only is Microsoft positioned to weather a downturn in new Windows deployments, it will thrive in the long run because the company realizes that most customers will move slowly to the cloud. That’s right: enterprises aren’t quite as ready to embrace the cloud as blindly as all the froth around companies like Workday and Google and Salesforce suggests.

It’s time to stop obsessing about the growth rates of Software-as-a-Service (SaaS) companies that are still a rounding error for IBM and Oracle and Microsoft, and temper the apparently foregone conclusion that on-premises software is all but dead.IDC predicts that by 2016 $1 out of every $5 spent on software will be on a cloud model. In case that’s not clear enough: Only 20% of software revenue will be cloud based . . . three freaking years from now! There’s data behind that prediction, and most CIOs won’t find fault with it.

Sure, there’s tremendous promise, but enterprises want the flexibility to keep some applications and business processes on premises, with an eye, and an occasional shift toward the cloud. Technology providers that can promise a seamless transition and flexibility will win. Microsoft’s recent moves indicate that it gets the point.

When you add in the rapid rise of Windows Azure to a 20% share of the cloud computing market with $1 billion+ in sales, when it is up against Amazon Web Services, this is an interesting problem to think about.

While SaaS gets all of the media attention, companies doing less sexy stuff in the technology space can deliver substantial revenue growth in enterprise focused channels like servers and virtualisation. The cloud is an easier sell for PR managers at SaaS firms obsessed with growth at all costs.