Commentary: The Hyperion merger and aftermath

You can contact Nigel Pendse, the author of this section, by e-mail on NigelP@olapreport.com if you have any comments, observations or user experiences to add. Last updated on July 23, 2003.

Hyperion Solutions finally began to perform as a market-leading vendor in 2002, with good financial results and resumed license fee growth. This followed four years of post-merger traumas that were an object lesson to any other OLAP vendors misguided enough to contemplate a merger or large acquisition. The company was formed from two of the most successful companies in the OLAP industry, and although described as a merger, it was in reality a reverse takeover by the smaller Arbor Software of the larger, more established Hyperion Software. Now, five years later, Hyperion is risking another relatively large acquisition, of Brio Software.

Background

Arbor Software had developed and marketed the popular Essbase OLAP server. Arbor’s focused marketing had put OLAP on the map. Indeed, through commissioning E. F. Codd and Associates to produce a white paper, the OLAP term itself was invented. Arbor had not invented the concept — indeed, OLAP servers had been around for some 20 years before Arbor was formed — but it was the first company to understand the need for good marketing. In contrast, Hyperion Software — then called IMRS — was originally a true pioneer when it proved in the early 1980s that the PC was a viable platform for financial consolidation systems. Its Micro Control product competed with a variety of expensive mainframe products, and it beat them not through being cheaper but because it was a better product. The Windows-based successor to Micro Control was originally called Hyperion, and it was so successful that, in 1995, IMRS renamed itself to Hyperion Software, and the product to Hyperion Enterprise. Hyperion’s attempts to develop other applications all failed, but it succeeded with its acquired Pillar product. Though both Enterprise and Pillar were overtaken technically, Hyperion’s focus on this area allowed it to dominate the analytical financial applications market.

Arbor had had a good relationship with Microsoft until the mid 1990s, but when (in early 1996) the latter decided to develop its own OLAP technology, rather than acquiring Arbor as had been expected by many people, it put pressure on Arbor and its premium-priced Essbase OLAP server. Arbor partnered with IBM (which has frequently been suggested as a likely acquirer), but ultimately decided that it needed to move into the applications space, to sidestep the inevitable commoditization of OLAP servers that would occur once Microsoft shipped its low cost OLAP server (which finally occurred in late 1998). The logical solution might have been to buy a number of smaller, specialist companies (such as Comshare, some of whose products already integrated with Essbase), but instead it decided to acquire Hyperion Software, the largest analytical applications company and its then second-most important competitor (after Oracle).

On paper, this combination of the leading OLAP technology company with the leading OLAP applications company made perfect sense, both to the two boards and the executive management teams. In the interests of diplomacy, the Hyperion name was preserved and the move was described as a ‘merger’ when it was announced in May 1998. However, astute observers were not fooled. For example, the Wall Street Journal commented on May 26, 1998 that, “Arbor appears to be the acquirer in all but name”. Unfortunately, many of the employees were less astute and wasted time trying to operate the new entity as a merger of equals.

Like Hyperion, the new company was headquartered in Sunnyvale and had a senior executive team from Arbor. Hyperion shareholders even suffered the indignity of receiving ARSW stock in exchange for their HYSW holdings, before it was renamed to HYSL. But the majority of the board and employees had come from Hyperion and the culture clashes soon emerged.

• Arbor was a brash, young, Silicon Valley company with a culture and ethos modeled after Oracle. Sales people dominated the company, which focused on fast growth through license fees, with most services and many complementary products provided by a network of partners.
• Hyperion was a more traditional east coast company, with an IBM-style conservative culture. It had a lower growth rate than Arbor, and was much more focused on services revenue. It had few successful partnerships, but many loyal customers. Hyperion had become successful through its focus on solutions and support, not technology.

Immediately after the merger in October 1998, the company realized that it had to decide whether customers were ‘owned’ by sales or professional services. The former won, and Craig Schiff became the second casualty from the executive team in January 1999; Kirk Cruikshank, Arbor’s brilliant head of marketing had already departed, less than a month after the merger went through, perhaps indicating the problems to come. In April 1999, the CEO and head of worldwide sales were dismissed, and the carnage among previously very successful sales and marketing people has continued ever since then. This has been accompanied by quarterly results ranging from the outright disappointing to the barely acceptable (especially compared to either of the illustrious predecessor companies).

After a five month search, in October 1999 Hyperion surprisingly appointed an existing member of its own board, Jeffrey Rodek, as the new chairman and chief executive officer. Stephen Imbler, who had been interim president and chief executive officer since May 3 1999, when John Dillon (president and ceo) and Bill Binch (svp of worldwide sales) were dismissed, became president and chief operating officer. Rodek had been president and worldwide chief operating officer of Ingram Micro since January 1995 and had been a member of Arbor’s and then Hyperion Solution’ boards of directors since January 1998 (throughout the troubled merger period). Jim Perakis immediately stepped down as chairman, but stayed on the board until July 2000. Michael Sternad, also from Ingram Micro, was appointed the new chief financial officer (Imbler had been cfo until May 1999), but remained in the role for less than a year. He was moved to become chief strategy officer in August 2000, and then reduced to being a strategic advisor to then company president, Stephen Imbler. His short stay in the Hyperion Solutions management team was over.

In October 2001, Godfrey Sullivan joined Hyperion as president and COO, replacing Stephen Imbler who stayed on for a few months in a part-time advisory role, but no longer part of the executive team. Sullivan’s responsibilities are wider than Imbler’s had been, and include product development and marketing. One of his first actions was to unify all engineering under Robert Gersten, which is reducing the embarrassing fragmentation in Hyperion’s product line. However, Hyperion still has a curious split in its technical responsibilities, with the CTO, John Kopcke, reporting directly to the CEO, Jeff Rodek, and not responsible for product development. One key Sullivan recruit was the new CMO, Nazhin Zarghamee. On her appointment in February 2002 she became Hyperion Solutions’ fifth head of marketing in the three and half years since the company was formed. She has adopted the “Business Performance Management” slogan as the umbrella description of Hyperion’s various tools and applications, and this seems to be working well for Hyperion.

IBM

Hyperion (and Arbor before it) has had a long relationship with IBM, which has been both a user and reseller of Essbase for many years. The first reseller agreement dates back as far as 1993, though the more strategic relationship that led to DB2 OLAP Server dates from early 1997. As a result, IBM has often been suggested as an acquirer of Arbor or Hyperion, but there is no evidence that any formal negotiations occurred. This may be because IBM’s sales of Essbase have been disappointing, and IBM has had much less success with OLAP than Oracle or Microsoft.

Perhaps reflecting the IBM influence, Essbase was also the first OLAP server to ship on Linux. It was released in October 2000, but its sales have been minimal. Again reflecting the IBM link, DB2 OLAP Server is now also available for IBM’s S/390 mainframes and IBM provides a free but very restricted Essbase ‘starter kit’ with DB2 7. Essbase is thus only OLAP server available on mainframes, AS/400 and Linux, though none of these versions are selling well. But with these versions, Hyperion has unequalled platform coverage among OLAP servers.

Revenues

Despite fluctuations, it does seem that that the applications business is growing and the Essbase (technology) business shrinking. In none of the last ten quarters, the technology licenses have shrunk year-on-year.

Starting in the December 1999 quarter, Hyperion started to show modest signs of recovery, and the March 2000 quarter was better, though the growth rates were still lower than the predecessor companies or the competition. June 2000 was better still, and for the first time since the merger, the license fees just exceeded the pre-merger peak. But the recovery was not to last, as the disappointing September 2000 figures demonstrate. There was a small improvement in the December 2000 quarter, but the company was back in the red by March 2001.

We calculate that the company’s revenue is now running at least a third lower than might have reasonably have been expected had the merger not occurred and the companies continued on their well-directed former trajectories. The two companies had a market capitalization of $1.3bn just before the merger was announced in May 1998. It had already fallen to $909m when the deal closed in August and it ended 1998 with a market capitalization of just $541m. It was down to below $300m in early April 1999 (with a modest recovery following the user conference and the dismissal of top executives) but it has recovered since, with a good run up at the end of 1999 and early 2000, all of which was lost in the October sell-off, when the market capitalization fell below $400m. Even after the January 2001 recovery, it stayed below $800m, and it remained there in early 2002. Even after a well received June 2002 quarter, it was below $600m, but much of the fall was because of the general market meltdown, and nothing to with Hyperion. The last year’s stock price performance can be seen in this link.

We were always skeptical about this ‘merger’, a stance which has not won us friends in the company (though many of those who were most unhappy with our comments are no longer with the company). We thought that a technology company like Arbor would struggle to manage a solutions and applications business like Hyperion’s, and the fact that both old and new applications collectively now account for less than half the total license fees bears this out; before the merger, Hyperion Software’s license fees were about twice Arbor’s. Not only are the older ‘classic’ Hyperion Software applications (Enterprise and Pillar) fading, but their replacements were not competitive until the v2.09 releases in late 2001, and were not selling well until the mid 2002. It is clear that, even today, Hyperion Solutions is still more of a technology company, like Arbor, than an applications and solutions company like Hyperion Software. Consequently, despite a large investment in applications, Hyperion has lost market share in the financial applications market — though it remains the clear market share leader — and its new applications in other areas have all failed.

Unusually, Hyperion released a number of product and regional revenue breakdowns in its press release from the December 2000 results. These allow the latest performance to be compared with earlier periods for the ‘Hyperion Software’ (both old and new financial management products and associated third-party and allocated tools), ‘Arbor Software’ ( Essbase, Analyzer, EIS, Hyperion Application Builder and associated third-party and allocated tools) and the post-merger new applications (eCRM/PRM analysis, performance management and associated third-party and allocated tools) parts of the business. The latter grouping was no longer reported on after mid 2001, so we have dropped it from out table and included all applications in a single grouping.

In order to eliminate the worst of the quarterly fluctuations, we have looked at license fees for eight six-month periods: that immediately preceding the merger and the seven from July 1999. The last column shows the growth in the July-December period from 2001 to 2002. Remarkably, this was the first six-month period since the merger to show positive license fee growths for both the applications and technology products.

The slump in the old financial management applications in 1999 and 2000 is very evident, though in the latter half of 2001, Essbase actually declined faster than the applications. In fact, in the last quarter of 2001, technology licenses dropped by a startling 36 percent. This is worse than the figures from almost any other BI vendor. The expensively-developed CRM applications were also abandoned.

But the financial applications did well in the summer of 2002, and grew for the first time in six quarters. The technology products, in turn, had their first growth quarter, after seven quarters of decline, at the end of 2002.

The Microsoft effect

It is ironic that this merger was triggered by Microsoft’s entry into the OLAP server market, because although it was designed as an escape route for Arbor, the self-inflicted damage from the merger did far more harm to Arbor’s shareholders, management, employees, partners and customers than anything Microsoft has done. However, despite repeated denials from Hyperion, Microsoft OLAP is clearly hurting Essbase sales and prices, so this was an extremely bad time for Hyperion to have such severe self-imposed organizational and product development problems that distract it from combating the dangerous external threats. The 36 percent decline in Essbase licenses in the last quarter of 2001, and an unbroken period of seven quarters of decline are real indicators of the pressure that it is under (though there was a good quarter at the end of 2002).

Nevertheless, Hyperion has also shown that it can win some high-profile competitions against Microsoft, which usually means that Microsoft also loses the SQL Server side of the deal (often to Oracle or IBM). Hyperion has also gained from Oracle’s long OLAP sabbatical. Crystal’s Holos sales have dwindled to nothing, which may have benefited Hyperion. However, it is telling that Essbase has been in almost continuous decline since the much enhanced Microsoft Analysis Services replaced the original OLAP Services in late 2000.

The future

The entire executive team, and much of the management, from the time of the merger had been replaced by the end of 2001. The new team is, with a few notable exceptions, less experienced with OLAP, but it also brings a fresh and more diverse perspective from other companies. It is also less defensive about the past, and is more prepared to make the necessary changes that are needed to make the best of the situation the company is in. The new team also brought more stability in 2002, with the elimination of the management and strategy turmoil of recent years.

This means focusing on fewer applications than were launched with great enthusiasm soon after the merger, and concentrating on renewing Essbase and financial applications development, which suffered after the merger. The signs are that there is a restored pride in Essbase, and higher quality new releases are promised after the long delays and poor quality of the 6.0 series of releases. One welcome sign is that the Essbase brand is being promoted more strongly, and it is being used again for some of the optional modules. Essbase XTD is also being used as the name for bundled packages of modules.

Hyperion also has impressive new Planning and Financial Management applications, whose version 3.0 releases are finally functional enough to replace their tired, dated predecessors for most applications. This is sorely needed, as the old Pillar product and user base is very vulnerable (though Pillar continues to be sold for certain departmental budgeting applications).

However, and very disappointingly, the two new financial applications are not integrated and share nothing, though at least from the 3.0 versions, they finally have a very similar look and feel. The internal architectures, technologies, Excel add-ins, data entry forms, metadata, etc are all different. Apart from Hyperion Planning, none of Hyperion’s applications use Essbase as the engine. Most of Hyperion’s competitors have worked hard to provide the full range of functionality in one more-or-less integrated application, but Hyperion actually developed all-new applications that were even less integrated than their predecessors which had started life in different companies.

Hyperion still offers a much more complete range of financially-oriented OLAP tools, applications and skills than any of its competitors. Oracle’s offerings were never as complete, and have retreated as part of its bungled technology upgrade, while Microsoft does not itself offer analytical applications like Hyperion’s. Some Microsoft partners, such as Comshare, OutlookSoft and CIP do offer applications that compete with Hyperion’s, but they have had very little impact so far. Cognos is the most direct competitor, but particularly after its acquisition of Adaytum, its offerings are even less integrated than Hyperion’s, and its Cognos Finance application is less sophisticated. So far, Business Objects has chosen not to compete directly with any of Hyperion’s products, while Crystal, which did compete to a modest extent, was enlisted as an ally, though that relationship has now abandoned in favor of Hyperion’s acquisition of Brio.

Hyperion also has to compete with ERP vendors, such as SAP, JD Edwards and PeopleSoft. In these cases, Hyperion offers much stronger products, but sometimes finds that companies are so loyal to their ERP provider that they will not even consider outside alternatives. Generally, if business users are involved in the decision, Hyperion (and other BI specialists) have a strong chance, but in purely IT-driven decisions, it will often lose out, despite all the evidence that the ERP vendors’ products do not deliver good financial analytics solutions.

Thus, although Essbase remains under competitive pressure from Microsoft today (and Oracle in the future), Hyperion still offers a stronger and more complete set of financial OLAP capabilities that are not equaled by those from any single competitor. It has a lot that it still needs to improve, but it now has a management team with no legacies in the predecessor companies, and for the first time since the merger, unified sales, marketing and engineering engineering under a common management. It also has a large blue-chip customer base, which apart from anything else, provides plenty of cross-selling opportunities not open to smaller competitors, and the Hyperion brand, despite all the problems, remains strong. And Hyperion continues to have a very strong balance sheet, so even if it were draining cash (which it is not), survival would not be a problem, unlike some of the battered smaller and medium sized BI vendors, such as Applix, Brio, Comshare, MicroStrategy and Sagent. Of course, Sagent has been acquired by Group 1, Comshare by Geac and Hyperion itself has acquired Brio.

There never was any danger of Hyperion imploding, but it still has to prove that it can consistently perform as well as its major BI competitors — something that its predecessor companies consistently did before the merger but which Hyperion Solutions consistently failed to do in its first four years until mid 2002, when its performance improved markedly. If this positive trend continues, it really will look like the merger traumas are finally over, and that Hyperion is performing like the market leader it is. With the Brio acquisition, Hyperion has revenues of about $600m, comparable to Cognos, but significantly smaller than the merged Business Objects + Crystal.

This page is part of the free content of The OLAP Report, but which represents less than a tenth of the information available to subscribers. You can register to access a free preview of a small sample of the large volume of subscriber-only information.

All information copyright ©2003, Business Application Research Center, all rights reserved.