Support and Maintenance: No Longer the Software Industry's "Best Kept Secret"?

For the longest time, the traditional pricing strategies of software vendors have included a reliance on support and maintenance (S&M) contracts, which is where most vendors could be assured of meeting, if not surpassing, their margins. The sizable profit made from S&M contracts (with customers paying more over time) has been one of the software industry's best kept secrets. Furthermore, user enterprises have not had the option of turning—that is, not until recently. To learn more, please see part one of this series Will User Enterprises Ever Get onto an Easy (Support and Maintenance) Street?

To be fair, some vendors, such as Oracle, do not require customers that are purchasing license agreements to purchase software support as well. But, for any customer that is implementing a new software system, software support is similar to an insurance policy. Software support ensures that customers will have access to technical support, regular updates, and new releases of the vendor's purchased applications.

In the past, software buyers have been mostly focused on bargaining over the up front, one-time-only license fees, and have not been paying enough attention to how much they will be paying in software maintenance fees thereafter. This has prompted many vendors to offer ludicrous discounts during the final negotiations, if not even give away free software, knowing full well they will recoup any losses from these "freebies" through support and maintenance agreements.

Sure, a software support contract is usually charged as a percentage of the software's net price, meaning that if a vendor gives a 50 percent discount on the up front license fee, the customer gets an ongoing 50 percent discount on the price of the maintenance. Still, over the entire life cycle of the system, most customers ironically pay far more in maintenance fees than they ever pay in up front license fees.

Lately, however, feelings of "being had" (taken advantage of) are slowly but surely sinking in among user enterprises. Vendors are beginning to see that corporate buyers of enterprise software are taking a much closer look at what they are paying for their applications' S&M. In other words, while information technology (IT) managers have traditionally only moaned about the S&M tab, lately more and more IT managers are intent on finding ways around ever-increasing S&M costs.

Vendors will quickly cite that the bulk of software support fees is related to upgrade rights. Customers receiving S&M instantly acquire an appreciating asset, as they receive access to all future innovations at no additional costs. Furthermore, vendors will point out that this practice is common in other industries.

For example, Best Buy, the consumer goods retailer, offers two types of maintenance contracts: a service repair contract and a service replacement contract. The service repair contract promises only to fix the product the customer bought. The service replacement contract means that, for the life of contract, the consumer can bring in the product and have it replaced with a brand new one—no questions asked. This latter policy is logically much more expensive, and it costs about 1020 percent of the original price.

But this is where the similarity with enterprise applications ends. Everyone realizes that trading in an outdated enterprise application, or replacing one that is malfunctioning, is far too difficult and inconvenient (if even possible) for the customer than doing so with an appliance or a car.

In reference to the previous analogy of shopping for a new TV, appliance, or car (see Will User Enterprises Ever Get onto an Easy (Support and Maintenance) Street?), some vendors point out that a customer does not get a new and improved TV set, appliance, or car for free every year or two from the vendor that the customer purchased support programs from. If one could purchase a car at 3050 percent off the list price, and then pay 20 percent of that per year to get the latest, newest version of the car every year, then that would actually be a pretty good deal for the customer. Application technology is advancing at a far greater rate than automobile technology is. Therefore, upgrades should be seen as even more valuable, albeit only to the minority of customers that really need the "latest-and-greatest" technology, complete with all the "bells and whistles" (with every feature possible).

Vendors also cite that their research and development (R&D) costs are constantly going up as software becomes more complicated, and that any profit and loss (P&L) view of their support businesses has to include all costs for R&D, and should amortize the cost of the original sales effort. But again, it is the free market that should determine the fair price, as is the case in other industries (given that retail has been mentioned for comparison).

To be fair, many software customers radically alter the products they receive, making support for those products infinitely harder. Staying with the retail goods example in Will User Enterprises Ever Get onto an Easy (Support and Maintenance) Street?, when a consumer buys a TV set, the purchase agreement generally states that any tampering with the product itself (even opening the back panel for a peek inside and consequently breaking a security seal) negates any warranties, and automatically cancels any service contracts.

The question is, why then have software vendors not been that strict in enforcing such clauses? Perhaps the answer is that vendors are tacitly conceding that customers cannot wait for lengthy upgrade cycles, whereby often the needed improvements for some customers will not come in the next few product releases. Also, vendors realize that an enterprise system is a mission-critical infrastructure or platform for an entire business, as opposed to a mere appliance purchased for individual use.

In any case, vendors' "best kept secret" status (see Will User Enterprises Ever Get onto an Easy (Support and Maintenance) Street?) is no longer tenable, given many recent quantitative findings and analyses. One such finding from International Data Corporation (IDC) shows that, in US dollars, businesses will have paid nearly $100 billion in software maintenance fees in 2006, whereas maintenance fees accounted for $86 billion (or 41 percent) of the $210 billion in revenue collected by software vendors in 2005. These figures for software maintenance fees are expected to grow 9.6 percent per year, and reach $137 billion (USD) by 2010, accounting for nearly half of software vendors' revenues.

One should note, however, that the growth in software maintenance fees is due to a combination of license revenue growth (which in many cases includes software maintenance) and inflationary adjustment increases in support revenue fees. The expected increase in revenue of 9.6 percent annually is not totally due to increases in support fees. It is useful to note that as more software is sold each year, the installed base becomes larger, resulting in more customers receiving more software for free through upgrades. For instance, more Oracle customers will receive Siebel 8 for free (no additional charge beyond their support fees) than any prior release of Siebel Systems' software (as well as the new releases of PeopleSoft and Oracle E-Business Suite [EBS]).

This growth also reflects a consolidation of vendors and the widespread adoption of enterprise software among businesses. With more customers continuing to work with their existing vendors (that is, the installed base revenue taking a larger share of the pie), the revenue stream will shift from new revenues to recurring revenues. This is natural and does not necessarily reflect diminished value either for customers or vendors.

A bigger installed base, like Oracle's or Infor's, should also help vendors (and arguably, customers) by spreading development costs across a larger customer pool, and lead to quicker identification of best practices for many customers. As vendors support larger and larger installed infrastructures, their recurring bills will naturally increase, while at the same time, the relative spending on "new stuff" will decrease.

Certainly, while this state of affairs is great for vendors, it is definitely not so great for customers. To that end, a Business Technographics survey of 436 IT decision-makers in 2005 found that 70 percent in North America and 60 percent in Europe cited high software maintenance costs as their biggest challenge related to software licensing models.

Further, according to a recent Forrester Research survey, the current number one IT initiative is to reduce software costs in any way possible. S&M, which consumes one-third of all software costs, becomes a logical line item to consider when reducing costs. IT entitlement costs, such as enterprise software maintenance, have (at some companies) soared to a record average of 70 percent of total annual IT budgets.

Logically, as the maintenance costs become a larger percentage of spending, users want to look at how to reduce those costs. But many chief information officers (CIO) also know that the software bill he or she receives can be minor in comparison to the costs of running an internal IT department. This is particularly true within larger companies, where both items need to be thoroughly reviewed to make a real impact on the bottom line.

Oracle cites its attempts to take on a bigger role in helping to reduce overall costs for user enterprises. As an example, prior to the Siebel acquisition, customers needed to build and support customized integrations between Siebel and other Oracle applications. Now, as part of the Oracle Fusion integration strategy, the vendor is building engineered links between Siebel and key modules of Oracle EBS. This means that the vendor will assume the development and support duties for these integrations (thereby freeing the customer from investing in this effort) without increasing fees to the customer. While highly commendable, this effort on Oracle's part is certainly not a universally useful development for many other customers with many other applications in place that will not benefit from these ready-made links.

Nonetheless, it is still surprising to see most vendors continue to (oblivious to the general users' feelings or not) increase maintenance revenue in many ways, starting with increasing maintenance fees directly. A decade ago or so, an "upper teen" percentage of the nominal (price list-based) software license fee was a typical benchmark for maintenance fees, but this percentage has slowly but surely crept up into the "lower twenties," or even higher.

For those not yet dismayed by these figures, maybe the following notion will make some of them wince: in less than five years, user enterprises pay the amount of their entire license fees once again in maintenance. In other words, software customers effectively "re-buy" their applications every four to five years through maintenance fees, while enterprise software vendors continue to raise annual S&M fees year over year and well beyond the inflation or consumer price index (CPI) adjustments (which could be justified, at least).

Oracle, for instance, recently reported its highest revenues ever for software license update and product support, with corresponding profit margins of a whopping 90 percent. The estimate is that about 14,000 licensees of Siebel, PeopleSoft, and JD Edwards software combined are paying Oracle more than $1.5 billion (USD) in annual support and maintenance fees. With support margins running as high as 90 percent, what does this mean for the rest of Oracle's business?

To Oracle's credit, the greatest value of the vendor's annual software support fee (over 70 percent) is associated with the product development that leads to new releases of the customer's licensed software. The costs for R&D and "bug" fixes are not included within the above-mentioned 90-percent margins. Also, it is not too difficult to believe that the combined maintenance fees for all Siebel, PeopleSoft, and JD Edwards customers are $1.5 billion (USD), given that under these previously individual entities, that amount was even higher.

Based on the above numbers, the average customer is then paying $107,000 (USD) annually, and one should note that many large companies could not hire their own staff to maintain, research, and develop regular updates; fix errors (24x7 and in multiple languages); and continue to provide new features and enhancements to the software for $100,000 (USD) annually. But what about the smaller companies that have stable, mature systems in place and that do not need the latest features that their larger brethren might require?

Another interesting (yet more general) question here is this: what would happen if vendors did not make their biggest margins on maintenance? At best, vendors average margins of 1015 percent on the entire business even after the larger margins made on maintenance are taken into account. The vendors' stockholders demand 1015 percent profit, so if vendors' revenues from maintenance were to decrease, where would that leave these vendors?

The original price of the software (license fee) is really quite arbitrary, since vendors charge only what they can, given customers' tolerance and what the competition charges. Maintenance dollars do pay for product development, so one could imagine the consequences should profits in this area decrease.

On the other hand, sales and marketing expenses (aimed at getting more revenue) may actually go up if the expected maintenance dollars do not come in. The logical "victim," besides product development, would then be support services, since fewer available dollars would mean fewer offered services. But given the current high margins for maintenance, vendors can still afford to provide services with fewer dollars. So in summary, vendors need the S&M money. But what if the customers rebel en masse and refuse to pay it?

Second, vendors have been tightening enforcement of existing agreements lately. In the past, vendors were quite lax about pursuing customers that exceeded agreed-upon, concurrent, or named user counts, which usually form the basis for software pricing. But lately, it is "no more Mr. Nice Guy," as vendors have been enforcing their contractual rights to audit customer usage of software. Vendors are now charging customers license fees for additional seats, plus the maintenance fees on those seats.

Furthermore, vendors' support policies keep customers from removing unused user or product licenses (see Application Erosion: Eating Away at Your Hard Earned Value) from maintenance contracts. Specifically, in the case of users that want to remove unused, functional modules or users, some vendors then simply and tacitly (if not sometimes furtively) reprice any remaining licenses at higher rates to ensure total maintenance fees remain virtually intact. As a positive example of this practice, Oracle has always had a license management services group that works collaboratively with customers to ensure proper licensing and intellectual property protection. Most customers reportedly want to be compliant and welcome this kind of assistance.

Still, the "concurrent user audit" is a big source of revenue for many consolidating vendors, and users' "horror stories" are beginning to mount. This ties in to the vendors' business models. How does any vendor with multiple products and those products' releases plan to make money if not by "soaking" (charging higher and higher prices) the users?

A look at what drives the cost of S&M up may be interesting. The top reason for S&M's high cost is the need to support multiple backdated versions. It is quite common for any vendor to be supporting several (even umpteen) old versions (major release numbers, dot and service pack releases, translations, etc.) at the same time, and this costs a lot of money. This is despite the contention of some vendors that, owing to their acquisition strategies, they will have consolidated the best practices of all acquired vendors, and will have rationalized prices across all product lines. Every vendor is still occasionally tempted to cut off support for old releases, but few vendors ever do for fear of customers' defections.

This is part two of the series Will User Enterprises Ever Get onto an Easy (Support and Maintenance) Street? In part three, the value proposition of support and maintenance will be looked at, and maintenance agreements examined.

comments powered by Disqus