The Business of IT

Don’t become a software business, become a learning organization

Watts S. Humphrey famously said, “every business is a software business.” Unfortunately, this statement has been misconstrued by well-meaning IT practitioners to demand primacy in the organization’s strategy, budgeting, and focus. But nothing could be more dysfunctional than letting technology lead your business.

A business exists to serve customers, and nothing else matters. Customers don’t care about anything “behind the curtain,” only the elements of your business with which they interact.

Here it is laid out in a convenient table:

Customers don’t care about:Customers do care about:
The tools you use while creating itThe tools they use to consume it
The clothing you wear while building itThe clothing you wear while interacting with them
The corporate mission statementThe company’s impact on their values
How easy it is to create itHow easy it is to consume it
Your share priceHow well you will serve them in the future

Do customers care what technology you use? No. The technology you use is just a tool to create the product/service.

Stop letting your technology organizations lead with technology. Instead, transform your entire organization—including the technology functions—into a learning organization. A learning organization rapidly learns about:

  • the customer
  • the problem domain
  • the tools they must wield to fill the customer’s need.

What are the elements of a successful learning organization? Watch this space for my next article.

The Business of IT

Three Critical Pitfalls to the Success of Cloud Offerings

Whether you operate IaaS, PaaS, or a SaaS service, watch out for these three pitfalls to the success of cloud offerings:

  1. Encouraging the wrong thing. Measuring the success of the offering by looking only at utilization metrics damages sales. One cloud service provider measured the success of its IaaS service – and compensated its sales force – based on virtual machine usage. Little surprise that, for all the years that this practice continued, the provider could not land strategically important accounts. Like in any mature sales operation, your prospective cloud service clients should be evaluated based on several criteria, including:
    • strategic fit
    • urgency
    • reputation
    • visibility
    • cost of sale
    • short-term sales potential
    • long-term sales potential.

    Make sure you evaluate all these elements, and that you prioritize your sales and marketing efforts toward prospects with the right combinations of all these factors.

  2. Ignoring the relationship. On-demand and pay-as-you-go are not simply usage models; they are components of a new type of customer relationship based on transparency and immediacy. Make sure your customer relationships allow for:
    • Seeing usage and  billing easily, and up-to-the-minute.
    • Setting custom notifications and enforceable consumption limits based on both usage and cost.
    • Getting quick notice of service degradations and their ongoing status.
    • Being reassured after service disruptions that you have identified the cause and taken steps to prevent a reoccurrence.

    Your service may have other ways to build a customer relationship based on transparency and immediacy. Identify and exploit them.

  3. Focusing on infrastructure reliability. Selling a cloud IaaS service that differentiates with higher level of service reliability is not sustainable; instead high-SLA workloads must be built using cloud-appropriate designs. Any IaaS provider who tries to compete against the commodity non-reliable infrastructure clouds by offering a higher SLA engages in an unwinnable battle: the cost structure will make it impossible to compete against commodity offerings. As computing services become more commoditized, previously established patterns of infrastructure usage (aka “best practices”) that called for highly reliable hardware and “N+1” architectures are architecturally moot. Following those patterns in the cloud will not increase overall workload reliability. Instead, workloads that use cloud infrastructure need to be designed for scale-out and rapid recovery from infrastructure failure, providing overall workload reliability via software. The overall cost of reengineering workloads for cloud-appropriate architecture will, in the end, be less than the cost of building and operating a cloud that delivers a “classic” hardware-based HA.

Don’t make these mistakes. They could be killing your cloud business.