A core strategy of many hardware and software vendors is introducing the concept of planned obsolescence to their market segments. The technology customer base has become inured to this by years of beneficial and rapid advances in information techonlogies, but the state of the art has advanced past a point where each advance is globally applicable to each customer. Nonetheless, most follow the established pattern and without excessive examination of costs or benefits, move to upgrade their platforms shortly after newer versions become available.
It is worth re-evaluating this standard procurement approach. For most businesses, recent advances in hardware and software will not bring the margin of advantages that past updates have generated.
The days have past when computer hardware was bulky, balky, and prone to sudden failure. The primarily solid-state electronics of computers constructed after 1995 to standard reference designs are for all practical purposes infallible. With no moving parts and adequate ventilation to defend against heating cycles, there is nothing short of actual physical damage that can impede the proper function of these devices. In most computers, the only items prone to failure are those with moving parts--hard drives, CD-ROM drives, and cooling fans. Each of these items is cheap enough to replace individually if an issue occurs, and despite the fact that they are more prone to failure than the rest of the computer, in truth most components that function through the first month of operation will function properly for many years. Wear and tear on most office systems, which have little exposure to extremes of environment or to physical disturbance, is not an adequate reason to upgrade.
Performance has traditionally been the other major reason to upgrade hardware, and it remains the primary reason today. Faster processors, system busses, peripherals, and memory are all released regularly to market. Moore's Law remains in full effect and will for the foreseeable future. There is no question that an upgraded computer will be faster than an older one. However, there is often some question as to whether or not technically faster will equate to operationally faster. Past a certain point, the speed with which computers execute certain functions is completely un-noticeable to the user. Systems which have already reached this plateau are generally not good candidates for technology upgrades, even if several years old. If performance to the end-user is adequate, it will remain adequate without replacing the hardware. System speed does not degrade over time--the installation of unnecessary programs can cause this appearance, but recovering the "lost" speed is as simple as uninstalling the flashy screensavers, backgrounds, and toolbar knick-knacks... all of which should typically be done for security and stability reasons in any event.
The case for software obsolescence revolves around two factors: interoperability, and features.
Features are the most commonly cited reason presented by manufacturers for performing software upgrades. New features, in fact, are generally all that differentiate new versions from old from the perspective of the user. However, a feature is only as good as it is useful. Although there is no foreseeable endpoint to advances in software capability, it is true that most users do not make use of most features in most software packages that they use. Rather, a small subset of available features are used, and most newly introduced features are not added to that subset in subsequent upgrades. There are only so many requirements that the general user has of a typical spreadsheet or word processing program, and past those, excess features are fluff. The latest versions of many line of business programs do not functionally improve the user's job performance; in those cases, there is no feature-driven reason to upgrade to those newer versions.
Interoperability, though less often cited, remains the one major reason for upgrading line of business software. Many businesses, an increasing number, are finding that a major part of the value in their electronic documents is the ability to exchange them freely with other companies. The de facto standard has become Microsoft Excel and Microsoft Word, and although versions past the 97 suite are generally forward and backward compatible, enough issues exist with document version interoperability to make upgrades desirable in many cases, if not essential.
The key to avoiding interoperability issues is the adoption of solid and wide-spread open standards in document format. Adopting standard formats that are widely readable will also tend to ensure that they will continue to be useable between different software versions with minimal effort. Examples include HTML (Hyper Text Markup Language) and RTF (Rich Text Format) standards for text documents--both have remained completely readable across nearly all platforms and largely consistent in implimentation for years now, even as multiple and incompatible versions of the Word Document standard have been released and abandoned. An RTF document written in Word 2003 will be completely readable in Word 6.0; or in a variety of non-Microsoft applications as well. Not being established and maintained by a single manufacturer, these standards are not as susceptible to change.
No similar standard exists for spreadsheets or databases, although the new XML (eXtended Markup Language) may be developing into a promising alternative for those applications.
A special sub-case of both the above issues exists for operating system software. The interoperability issue is more serious, as failure to integrate between versions can result in network issues and an inability for computers to communicate with one another or the Internet. The feature issue is also more pronounced, as operating systems are more complicated and have not advanced at the same pace as business applications, and are therefore not as completely fleshed out with basic features. However, at some point, and in some operating system versions, the advances in a later version may not be particularly pronounced over the previous one, and therefore may not be worth an upgrade. Here, too, the adoption of open and general standards can prevent interoperability issues even over several generations of software.
Another issue is the availability of product support. Generally, manufacturers will cease to provide product support services on a product after a certain number of years. The exact duration of support is variable, usually depending not on any pre-arranged or contracted agreement, but on the relative popularity and use of the software. The same is true of many free software packages, which tend to develop an ad-hoc support community that corresponds to the depth of usage in the market. Generally, as long as the package remains in common use, support will be available, despite manufacturer's attempts to phase it out. There is also a phenomena with most software where after a number of years, it will reach a certain stage of maturity where new bugs or kinks will no longer appear, having been previously discovered and dealt with. Software in this phase of life generally does not require active support--once patched and implemented, it's unlikely to experience issues that local users or technicians are unable to deal with.
Knee-jerk upgrades are no longer an acceptable course of action for cost-conscious businesses. Careful evaluations as to the benefits of version upgrades need to be made, and both management and technical staff should recognize that intact working systems should continue to work almost indefinitely as long as the business system they support remains constant. Effort should be made to consciously break the expected cycle of release and upgrade that technology customers have become accustomed to.