92-Cisco’s New Server Product

Here is the story of an industry about to be disrupted by a Next Leader product. The server market when Cisco entered had attractive growth and predictable, but tough, competition. A few years later, some smaller server competitors began to modify their business models to provide unique Performance and Price benefits to the industry’s Very Large customer segment. That customer segment vaulted these smaller competitors to the industry heights. The original market share leaders in the server market continued to prosper for sound economic reasons. But their market shares have declined markedly.

Posted 3/30/09

Cisco recently announced that it was entering the server market. Details are sketchy right now, but we might take a brief look at what Cisco needs to do to be successful. We will use the Customer Buying Hierarchy as our analytical tool.

A bit of background. Cisco is entering the market for servers in order to increase the amount of the global IT purchase that it is able to address. Cisco claims that today it addresses about 10% of the total annual purchase of IT products. With the introduction of its server product, it believes that it will address 25% of that market.

But growth is not all that is pushing Cisco in the direction of servers. Hewlett Packard looms. Over the last few years, HP has developed and improved its ProCurve networking gear product line. This product line competes directly with Cisco routers and switches. Cisco may feel compelled to respond to HP’s forays into its market by counter-attacking in the server market. (See the Symptom and Implication, “Some competitors proliferate products around the heart of the market” on StrategyStreet.com.)

The Customer Buying Hierarchy holds that customers buy using four criteria: Function, Reliability, Convenience and Price. They use these criteria in that order for all purchases. Let’s use that Customer Buying Hierarchy to evaluate what we know of Cisco’s offering.

Function: Function refers to the characteristics of the product that affect the way the product is used by the customer. It includes all the features of a product. Cisco is entering the high-end of the server market, where margins are good and there seems to be some product differentiation among the competitors. We don’t know anything about the server itself, but the company’s offering integrates the server, networking components, data storage and virtualization software in one integrated package. It appears that the company is offering the hardware/software equivalent of Microsoft’s Windows/Office products. Everything works seamlessly together. These functions individually are not unique. The unique function is that they are all integrated together already in one package. This may give the customer a greater sense that the components will work well together. And it may enable some customers to avoid the cost of services they have needed previously to integrate these components in their IT locations. Cisco is offering a clear functional improvement.

Reliability: Reliability refers to the consistency with which the company delivers on promises made or implied to the customer by its product. There is good news and bad news here for Cisco. The good news is that the company has a sterling reputation for Reliability. Industry analysts claim that Cisco is rarely the Function leader in the industry, but Cisco treats their customers very well and ensures that the customer never gets in trouble using Cisco. It has a powerful reputation with customers as someone you can count on. So, what is the bad news? Cisco’s reputation for Reliability resides in network gear, not in servers. In servers, IBM and Hewlett Packard also have sterling reputations as companies capable of and willing to deliver superb customer service. IT customers trust all three of these companies. But if push comes to shove, wouldn’t the average IT customer trust IBM or HP more in the server product line? Cisco has a hurdle to overcome here. It must find a way to ensure customers that it will help them avoid all potential products from the new Cisco server. This may not be an easy task.

Convenience: This term refers to the ease with which the customer may acquire the product. Cisco looks to be in very good shape. There are several good companies who have signed on to add technology and sales to the Cisco product. Among the partners are such names as Accenture, EMC, Microsoft, VMWare, Red Hat and BMC Software. Customers will have no trouble finding the product and buying it.

Price: We don’t know what Cisco plans to charge for its new server product. One thing is certain. Its price has to be lower than a customer could purchase a high-end blade server and the services to install it. The new product line’s major benefit is that it integrates a number of components already available on the market. (See the Symptom and Implication, “Competitors are changing features of the product” on StrategyStreet.com.) Cisco cannot charge more for the integration than the customer would pay to someone else for a final integrated product. In fact, Cisco will have to demonstrate to the customer that it is clearly cheaper to purchase their product than a blade server and accompanying integration services from HP or IBM. Pricing is a big unknown here.

Since we don’t know pricing, we cannot reach a final conclusion. However, Cisco has taken on a big challenge here. This will not be easy. In order for this new initiative to produce the kind of results that Cisco hopes for, it is likely that its competitors, HP and IBM, are going to have to fail their customers in some way to enable HP to gain a major share of the market. Current IBM and HP customers are likely to grant these companies time to copy the unique benefits Cisco offers. In most mature markets, competitor failure is as important as a product benefit win in moving market share. (See the Symptom and Implication, “Share is tougher to shift” on StrategyStreet.com.) HP and IBM could fail if they do not create an equivalent benefit and then match Cisco’s prices. This is a tough challenge, but Cisco has met tough challenges before.

***

Update 2022:

Cisco’s market entry using a Function innovation failed. Many other competitors, most of them smaller than Cisco, offered more tailor-made Functions.  This is another example of the danger of competing on a unique Function. See HERE for an explanation.

In 2021, HP was the major brand leading the vendor server market with 15.7% of the market. Dell followed closely at 15.6%. Inspur Power Systems, China’s major competitor in the market, came in at 9.4% while Lenovo and IBM had seven and 5% respectively. The overall leader in the market was the original design manufacturer (ODM) group with 26.7% of the market.  Cisco left the market in 2018 while it still had a little over 4% of the total market.  The importance of the ODM group suggests that all competitors must tailor their systems to individual customers in order to succeed in the market. This takes away the main Function advantage that Cisco offered with its server introduction.

***

Update 5/26

ODMs are clear Next Leader Products. They have taken over the Very Large customer segment and are poised to grow even more in the future. The Hyperscalers ( Amazon, Google, Meta and Microsoft) make up the Very Large customer segment. ODMs offer that Segment more Function and Reliability than the traditional OEM competitors. They also have far lower costs and consequent Prices. Some of the larger OEMs could respond to this overwhelming challenge to help their own businesses.

The distinctive Value proposition of the ODMs is easy to see. First, their Function advantages. They eliminate any benefit not specifically required by the customer, making their products highly efficient.  They produce new product cycles in considerably less time than do the OEMs. Their products consume 25 to 35% less power. They require little or no support cost. The Hyperscalers run these ODM servers at nearly twice the utilization rate of OEM servers. ODMs enjoy unparalleled Reliability because they design their products in close collaboration with their Very Large customers. The product design is highly tailored to each customer, including custom boards, power, cooling, and firmware.

The ODMs complement this enormous Value proposition advantage with very low costs and Prices.  Their servers cost 30 to 60% less than OEM servers. The ODMs are cheaper because they eliminate anything not essential for their customers, saving on materials and labor costs. They produce these products at massive scale, with no redundant features common in enterprise computers sold by OEMs, no channel costs and no support burden. They sell their products in 50,000-to-250,000-unit orders, producing enormous economies of scale compared to the OEMs selling far more SKUs at much fewer average units per order.

From 2015 to 2026, the server industry evolved from an enterprise‑centric hardware business to a cloud‑first, AI‑accelerated, Hyperscale‑dominated ecosystem.  IDC data shows ODM share at 29% (the largest, Foxconn, is larger than the largest OEM in the server business), the top five OEM producers at 29% (the largest, Dell, holds only 8% of the market) and the remaining 42% spread among many smaller OEMs. However, current ODM share including hardware and systems sold through OEM channels brings that total share up to about 40%. Using that latter approach, the change in market shares summarizes the advantage of ODMs over OEMs. In 2015, ODM’s supplied about 15% of the global server units sold. By early 2026, ODMs controlled over 40% of the total market. ODM market share is growing at roughly 2 share points a year. ODMs dominate the Hyperscaler (Very Large customer segment) volume. OEMs continued to dominate the enterprise market. The ODM market is produced by a handful of massive US and Taiwanese manufacturers. The OEM market is broad and diversified.

Are the OEMs in trouble? Not at all. They are more in the services business than the hardware business. They enjoy high services revenues and returns on their broader businesses. The enterprise market is also relationship driven. It depends on intensive high quality global managed services, compliance with multiple geographic regulations, storage and networking, support contracts and software subscriptions. Relationships between OEMs and their enterprise customers change only rarely.

There may still be some opportunity for the OEMs in the Hyperscaler market segment. It is probably too late to enter that market de novo; the larger competitors are too well-established. However, the OEMs enjoy margins 4 to 5 times those of the ODMs. That might make an acquisition or partnership between a large OEM and a quality ODM possible. The discipline and business model the ODMs use might suggest improvements in the OEM business model that would apply both in servers and in other related markets. The situation is analogous to the efforts of several US manufacturers investing in Japanese manufacturers in the late 80s and 90s to understand their superior business models and adapt them to the US market.

***