Using Guns or Butter Economics in Data Centers

John Regan, VP of Data Center Division, Carter Validus
565
918
163

John Regan, VP of Data Center Division, Carter Validus

While some readers may well be aware of the g u ns -ve r su s -but t e r paradigm of economics, a fresh look at how this applies in the data center space may be beneficial. If your core business is analogous to making and selling guns then your resources are best directed in driving business efficiencies and effectiveness to those ends. Furthermore the paradigm implies that company making guns, if the need arises, should buy or trade butter from someone that can do it more efficientlyand effectively (their core competency).

More often than not, determining whether to own data centers, lease them, or outsource them partially or in entirety is no simple task. Information and ready access to it have driven needs for absolute control, security, protection of intellectual capital, and adherence to customer and regulatory requirements. These are just some of the variables that can influence choices about the data center and they need to be looked at closely in any decision making process. At the same time the opportunity costs of addressing these variables should not be taken lightly lest a company potentially invests capital and resources in areas that are less accretive to their core business.

After having visited and assessed numerous (est. 150) data centers throughout the U.S. my observations include seeing the gamut of extreme inefficiency, waste, and ready oppor t u n it ie s to let someone else make the butter, through the other end of the spectrum with exemplary efficiency, superior operations, and expert use of resources. Typical performance falls closer to the median as would be expected in a bell curve. This means that there are various degrees of opportunities for performance gains in how resources are utilized.

The biggest gains can be made in the use of capital as alignment of capital expenditures with fluctuating use over time has been a notable deficiency. Many enterprise data centers have overbuilt and/or overpaid for their facility and supporting infrastructure and CIOs are necessarily answering to CFOs calls for rationalizing ROI. As traditional obstacles in strategic sourcing of data center functions are addressed, it becomes increasingly difficult to justify spending significantly more than the butter makers to build a data center when your core competency isn’t building or operating data centers. A stick to your guns breed of CIO is replacing the old guard by embracing a focus of investments better aligning with support of their business processes.

These CIOs are correcting issues like overbuild and overspend on data center infrastructure with strategic sourcing, using managed services and hybrid cloud (sticking with their guns). They are looking at where they can possibly recoup a significant portion of capital through sale leaseback transactions and repurpose the proceeds in alignment with more strategic objectives of the business. The latter effectively frees trapped capital allowing it to be put to work more efficiently generating profits.

To illustrate this point understand the opportunity cost of a grocery chain paying a $100m for a data center vs. opening five new stores at $20M each. Opening five new stores will help generate more revenue and align with the core competency of selling groceries.

Furthermore if your cost of paying rent is 6-8% of the monetized value of the data center facility and your net profit margins are 20, 30, 40%, or more, sale leaseback can be an attractive option as it continues to be used by many large successful businesses. While an argument of “my cost of debt is lower than the lease rate”; short and long term debt has limitations to their utility and comprise only a part of a business ability to leverage. The CIO should not overlook the ability to include sale leaseback or leasing within the available capital stack. There are many variables to consider with how and when companies’ capital resources should be used and evaluating the opportunity cost should be quite high on the list of variables.

Another common challenge I have seen CIOs and businesses faced with attempting to monetize a data center real estate asset through a sale leaseback transaction is overcoming the hurdle of what they paid to begin with. When the butter makers are building critical infrastructure at $6m-$10m per Megawatt of resilient tier III configurations there is a hurdle to overcome when trying to sell an asset that cost $15-20m or more per Megawatt. With an accommodating deal structure this can be done but it usually translates to one of two things; writing off book value and/or paying more in rents.

“Many CIOs are looking at sale leasebacks where they can possibly recoup a significant portion of capital that can ultimately be repurposed in alignment with more strategic objectives of the business”

Optimal data center performance typically comes with the higher rates of utilization. I have seen under-utilized facilities operating at a power utilization efficiency (PUE) of up to 4.0 which can translate into paying three times the power costs that you would experience in a well-run facility at a 1.33 PUE. The biggest culprit is poor forecasting of technology’s use of the data center resources. A focus on understanding what you are really getting for the capital you are putting to work and how that translates to expected facility performance and cost of operations can really pay off while also helping to determine if you would be better served sticking to your guns and taking advantage of what the butter makers have to offer.

Since data centers are mission critical assets supporting performance of companies’ business processes, it would seem to me very realistic that they be treated accordingly. I have seen data centers that are dirty, used like a closet or warehouse for storage, and operated in a state of disrepair and apparent neglect. From an operating standpoint, there should be no second guessing how each and every device is connected to and supported by the facilities infrastructure. This means maintaining as-built documentation and labeling of all infrastructure components is a must have discipline. All equipment should be properly maintained in accordance with manufacturer specifications to ensure the operating stability and reliability of the plant. If these elements are not being addressed it may be an opportunity to give the task to the butter makers.

Lastly, while there are many questions to ponder and challenges to address in managing a continually evolving environment we should be frequently asking who can do it better, in what frame of time and at what tangible and intangible costs and at what risk(s)? Is what you have been doing in the data center space a core business competency? Is your strategy flexible enough to accommodate your businesses strategic imperatives? Can you effectively address potential risk, regulatory and compliance issues? Do you have the necessary utility within your budgets to respond to the blurriness of your crystal ball as you are looking at three, five or ten years in the future?One thing is certain, in the end the market will hold you accountable for your decisions on whether to make or buy your guns and butter.

Read Also

Data Center Evolution

Paul Schultz, VP of Data Center, Cloud, and Infrastructure BU, KGP Companies

What Will Matter Most Tomorrow is the Data Center Location You Choose Today?

Jeff Rossate, Executive Director, Office of Business Development, Minnesota Department of Employment and Economic Development

Innovation: Not Just Shiny Widgets

Michael Silla, SVP, Skanska USA Mission Critical Center of Excellence

Data Center of the Future

Ari Bose, CIO, Brocade