New ways to measure TCO for networks

Forrester Research says current methods of measuring total cost of ownership for networks are broken and suggests a new approach.

Total cost of ownership (TCO), one of the mainstays of modern IT marketing-speak, can be a valuable tool if used correctly -- but it rarely is, according to Forrester Research.

James Staten, principal analyst with Forrester, noticed this disturbing trend while working with clients not long ago.

"Nobody wanted to have multiple measures and metrics when figuring out their financials in business, and so they gravitated to TCO," he said. "It's a hammer, and everything else is a nail."  

Indeed, as Staten rattles off potential cost metrics – total cost of acquisition, total economic impact, relative cost of operations – it is easy to understand why networking professionals and vendors settled on TCO as a simple, single standard. However, companies are giving up a lot of flexibility and adding a lot of paperwork, Staten said, if they do not consider alternatives.

"To do a fully documented TCO is hard work, and … a lot of companies … don't have the bandwidth or data to do it properly," he said. Instead, many companies rely on various vendors' criteria, which may or may not fit their own business cases and may or may not include soft variables such as improved worker productivity. Staten said these soft variables should not be included.

Vendors love to tell companies that they make employees more productive, he said. But if that productivity is going to make a financial impact, that will mean fewer employees after deploying the product, and products rarely have that type of impact.

"When we talk to customers about how they justify purchases, they're all over the board," he said. "Some of them are really awful."

In the recent research note "TCO is Overrated," Forrester suggested that non-strategic purchasing decisions be made using relative cost of ownership (RCO). Instead of comparing every variable of a scenario, as in TCO, RCO analysis compares only variables that will change with a potential new investment: Additional hardware costs, staff training costs, and licensing fees are all common criteria, as are the greater or lesser support costs for switching hardware.

By comparing options in columns only where that data differs, more than half of a typical TCO's calculations can be left out, Staten said.

That does not mean that TCO is completely out of the picture.

"We don't recommend anybody try this unless they have a good account of TCO some time in the past," Staten said. "You absolutely want to do big analysis on big decisions … but if you don't use a degree of that on smaller decisions, then an accumulation of poor small decisions makes you inefficient."

Charles King, principal analyst for Pund-IT Research, was quick to defend TCO calculations generally, while underlining some of their faults.

"I think the TCO can be a valuable metric," King said. "But the problem is like the annual mileage rating for vehicles: Your mileage may vary, is the rule of thumb."

King said it was important to flesh out what he called the "line drawing" that TCO provided, particularly in considering factors that may be unique to a given company.

"What you can do is get enough subjective information from a variety of sources, and you can actually build an objective analysis from that," he said, suggesting that IT decision makers reach out to peers, third-party reviewers and media sources to build a composite picture that can be factored into a final decision.

Switching from one metric to another can also be helpful, King said, but if the underlying data is not good, the result may be little more than a shuffling of one wrong set of numbers for another. He suggested paying close attention to the source of data and being sure to compare any data provided by a third-party with internal usage numbers.

"I think the mistake many companies make is to create a process and view it as sacrosanct after that point," King said. "Things change over time. Prices of products can change pretty radically over time. And depending on the solutions you're looking for, there could be better deals out there that could affect the long-term health of not only your IT infrastructure but also your budget. Asking questions is good."

Read more on Data centre networking