In the U.S. server market, every piece of hardware is sold with some form of warranty. For example, in the product specification of Hewlett-Packard (HP) ProLiant ML 150 servers on HP’s website, along with processor, memory, and hard drive is a warranty that “includes 1 year parts, 1 year labor, and 1 year on-site support.” Thus, a server’s warranty is an essential product attribute in this market and plays an important role in buyers’ and sellers’ decision making. There are two types of hardware warranties: base warranties and extended warranties. The key difference between these two is that base warranties are bundled with the product and cannot be purchased separately, and extended warranties are optional and can be purchased at an additional cost.

For example, under the “recommended supplies and accessories” for HP ProLiant ML 150 servers are “hardware service, support and upgrades” that include many extended warranty options such as “4 years, 4 hours, 13×5, hardware support [add $434.00],” and “4 years, 4 hours, 24×7, hardware support [add $690.00].”

There are two types of hardware warranties: base warranties and extended warranties. In this article we focus only on hardware base warranties. Base warranties in the server market cover the costs of labor and parts (repair and replacement) when a server breaks down during the warranty period and services are provided on site, on the next business day during normal business hours.

Warranties in general have four economic roles—insurance, sorting, signaling, and incentive: The definition is as followed.

o The key assumption of insurance theory is that consumers are risk averse.
o The key assumption of sorting theory is consumer heterogeneity in some unobserved (to sellers) attributes such as income, quality valuation, and risk aversion.
o The key assumption of signaling theory is information asymmetry in the sense that sellers have better information about product quality than buyers do.
o The key assumption of incentive theory is risk endogeneity, in the sense that both buyers’ and sellers’ actions can affect the probability of product failure.

The insurance role of warranties will always exist as long as (1) buyers are risk averse and (2) there is a nonzero probability of product failure. Both conditions hold in the U.S. server market; therefore, we would expect warranties to play an insurance role in this market. Further, buyers of servers are various institutions that differ substantially in size, revenue, usage, technical capability, quality valuation, and risk aversion. Manufacturers offer a menu of warranty durations for their servers. Therefore, we would expect warranties to play a sorting role among heterogeneous buyers. In other words, the economic roles are applicable to the U.S. server market.

Because different manufacturers offer similar warranty durations for their server products, the quality information content of warranties is likely to be limited. For example, Dell, IBM, and HP all offer one year or three-year base warranties for their products and these warranties have similar coverage. Thus it is difficult for buyers to infer the quality of servers from warranty duration. The incentive role of warranties is closely related to the signaling role. Major manufacturers do not change their server warranty policies over time as their product quality improves, so the incentive role of warranties in the server market should be limited. Further, although not reported here, we also find that perceived server quality is not correlated with warranty duration, indicating that warranties do not appear to contain information on product quality).

In sum, there are two main rationales for warranty provision in the server market. One is to provide insurance to customers, and the other is to sort among heterogeneous customers. This implies that our empirical model needs to incorporate (1) consumer risk aversion, a feature required by the insurance rationale, and (2) consumer heterogeneity, an assumption required by the sorting rationale.

There are three features of base warranties in the U.S. server market: (1) all servers are sold with a warranty of nonzero duration, (2) all major manufacturers offer non-uniform warranties for their products, and (3) a given server model is usually associated with a single warranty duration. A hedonic regression with warranty duration as one regressor that can tell how price changes with warranty duration. However, to quantify the value of warranties to firms, we also need to know the impact of warranties on demand and how warranty costs change with warranty duration.

What does this all mean? It’s easy. Purchasing a warranty can be beneficial for much reason for an organization. However, a measurement of when to purchase extended warranties, how much to purchase, and when there are liabilities to the organization if a warranty expires becomes a job of a specialist to ensure that your technical assets are protected. Make sure you are covered where you need to be covered? If you have more questions call us and our experts will help you figure out if your warranties cover your organization.