Google IT: By directly rewarding efficiency in the data center, Google pioneers replicable and profitable organizational structure

In today's IX blog post "Treasure Hunting for Gigatons," EDF's David Witzel and Gwen Ruta urge us to "make money [by] addressing the first 20% or so of the climate change problem. These 4 to 5 gigatons of 'low hanging emissions' come from energy inefficiencies that actually cost firms money."

For one such example of low hanging emissions which chronically cost companies money, we pick a sweet one from the Innovations Review tree.

Our knowledge-based economy depends on data centers to store and process electronic information. Yet data centers also waste enormous amounts of energy. Shockingly (so to speak), a typical data center uses less than 5% of total power for computing operations and loses the other 95% along the way as heat in the servers, as conversion losses in power supplies, powering fans and lights and in cooling systems required to remove all that waste heat.

Not a technical problem, but technically a problem

It'd be tempting to throw up our hands and mourn all that wasted energy and money as casualties to a necessary evil. But we can't excuse ourselves so easily. This is a management error, not a technical conundrum.

In 2008, McKinsey and the Uptime Institute co-investigated the IT sector and found technical solutions readily exist to clot this chronic energy hemorrhaging. Yet, companies are typically not organized in ways designed to minimize lifetime energy costs. For instance, facilities and IT divisions are siloed, so utility bills virtually never go to IT; and they often don't reach facilities, either.

McKinsey and Uptime found most companies would institute easy, cost-effective efficiencies if they instituted two organizational corrections:

1) Incorporating the true lifecycle cost of ownership in the business case for adding capacity to data centers.

2) Formally consolidating responsibility for data center IT and facilities expenditures under one executive (an "energy Czar" tasked with cutting costs through improved efficiency).

What would Google do?

Among all technology companies, Google has the operational structure and procurement policies that most closely resemble the ideal McKinsey-Uptime model. And in-line with the philosophy behind that model, Google's data centers have become models of efficiency by way of sound organizational strategy instead of mere technical prowess.

How? Google put all data center operations (IT and facilities) under the control of one executive, SVP of Operations Urs Hölzle, and encouraged employees to consider the total lifecycle cost of purchases or total cost of ownership (TCO).

The result? Google now powers its data centers using less than half the energy as the industry average.

Executives at every company operating data centers should consider making these types of organizational changes in order to take full advantage of cost-effective energy efficiency measures.

According to Google's SVP Urs Hölzle, "We've saved many hundreds of millions of dollars by managing TCO."

Kenneth G. Brill of the Uptime Institute agrees. He estimates that Google's industry-leading data center efficiency practices have saved the company at least $500 million in capital expenditures, $33 million in annual depreciation and $17 million in annual electricity costs.

The Institute also estimates that data centers account for between 8 and 35% of overall energy consumption at non-manufacturing firms.

Why don't other companies manage their data centers like this? What sort of obstacles stand in the way? How can they overcome them?

See EDF's Q&A with Urs Hölzle

Sign up to follow the Innovation Exchange blog via email or RSS.