Consolidations in the BI industry

The BI industry has seen a wave of acquisitions since the mid 1990s, with takeovers occurring every few months. The first wave was mainly other companies who were attracted by the higher growth rates in the BI industry and preferred to buy an existing vendor rather than to develop their own product. These changes of ownership did not produce any ‘consolidation’ because there was no net reduction in the number of BI vendors or products. There was also no reduction in competition as market shares were not concentrated in fewer and fewer hands.

Examples of such non-consolidating acquisitions include the entry of the various database vendors into the BI market, including Oracle’s purchase of the Express business from IRI Software, Informix buying STG MetaCube and Microsoft’s purchase of the Panorama technology. Microsoft’s purchase of ProClarity is a consolidation, though it will not in itself increase the market share we calculate for the ‘Microsoft ecosystem’ as ProClarity was already included in it.

But there have also been a number of cases where existing BI vendors acquired each other and these do lead to genuine consolidation. This is particularly true of today’s Hyperion Solutions, the combination of the former Arbor Software, Hyperion Software, Brio and several small private companies such as Appsource. Business Objects’ acquisition of Crystal Decisions also amounts to true consolidation. Although these acquisitions reduce the number of BI vendors, they do not necessarily reduce the number of BI products, as often both product line live on separately in the merged company.

Another big question is whether these consolidations are good, for the products, the existing users or shareholders in the acquiring companies. Clearly, in some cases, the products have done much better under new ownership, because the new owner has far more financial strength than the former owner. The most obvious examples are the two products acquired by Microsoft, which sold in vastly greater volumes than they ever could have done as products from small Israeli companies.

But in many cases, the product was less important to its new owner who accidentally or otherwise neglected it. Again, a Microsoft example illustrates this: Data Analyzer has had no further development after its first re-release in 2001, whereas Maximal Max would presumably have continued to develop had Maximal survived or been acquired by a different company. It is also likely that Express development would have continued more smoothly without the badly mishandled transition to a new generation product that has occurred under Oracle’s ownership, and Holos would almost certainly have done better had Seagate not acquired it. MetaCube, too, would probably have survived longer without Informix’s ham fisted intervention, and the two ROLAPs accidentally acquired by Computer Associates would almost certainty have not been snuffed out so ignominiously had they remained independent.

More often than not, the shareholders of a company that acquires an OLAP vendor or product have not received a good return. Some of the worst examples of squandered shareholders’ assets include Informix’s acquisition of STG, Arbor’s acquisition of Hyperion Software, Seagate’s acquisition of Holistic Systems and SPSS’s acquisition of ShowCase. Of more recent acquisitions, both Cognos’s acquisition of Adaytum and Business Objects’ of Crystal Decisions are likely not to have generated a good return for their respective shareholders.

The table below lists some of the many acquisitions since 1994. Note that some products have changed hands more than once, and one of them (Pilot) has had five owners since 1994. Where known, the purchase price is shown, though of course in some cases, this price is for a whole company with both OLAP and other products. Transactions that could be regarded as genuine consolidation are shown with a cream background.

Notes

These are additional notes about some of the transactions listed above:

IBM returns to the application software business by buying Cognos

IBM confirmed longstanding rumors when it announced on November 12, 2007 that it was to buy Cognos for $5bn cash, representing a modest nine percent premium on the already takeover-spectulation-boosted closing price on November 9. The deal closed at the end of January 2008.

Unlike other recent acquisitions, there is little product overlap between the product ranges, though IBM’s large services business with other BI vendors’ products may suffer. Also, unlike other large BI acquisitions, there are few apparent synergies between the IBM and Cognos software businesses — IBM has in recent years focused on infrastructure rather than application software. Its earlier forays into BI have all been less than successful: AS, IOC, Metaphor and DB2 OLAP Server all came to a sticky end, and it had looked like a bruised IBM had turned its back on the BI software business.

IBM and Cognos have long had a productive business partnership, with technical and services cooperation, and perhaps IBM was worried that if Cognos went the same way as Hyperion and Business Objects, it could lose out on some of its remaining BI services revenue. Certainly, this acquisition looks like a defensive knee jerk response to Oracle’s and SAP’s large BI acquisitions earlier in 2007, though IBM claims otherwise. It is also believed that IBM was an under-bidder for Hyperion and Business Objects (both of which also had active IBM partnerships). And perhaps Cognos was more worried than it claimed about the prospect of having to compete as an independent against Oracle, SAP and Microsoft?

Cognos becomes a new unit of IBM’s Information Management division focused on BI and performance management. It will continue to be led by Rob Ashe, the current Cognos CEO, reporting to Ambuj Goyal, General Manager, IBM Information Management. This appears to be a similar strategy to that being adopted by SAP with Business Objects, whereas Oracle has integrated Hyperion to a much greater extent. As IBM does not have an existing BI software business, forced integration with other parts of IBM probably would not make much sense.

After this, MicroStrategy and Actuate will be the last remaining public BI vendors, though there remain two larger private BI vendors, SAS Institute and Information Builders.

One side effect of this and other large BI acquisitions is that BI now has much less visibility on Wall Street. Not so long ago there was a thriving community of public BI vendors, which kept financial analysts focused on this sector. Now, with only two medium-sized public BI vendors left, this attention will wane.

SAP abandons its organic growth strategy to acquire Business Objects

The wave of BI acquisitions in 2007 got a dramatic boost on October 7, when SAP announced that it was acquiring Business Objects for €4.8bn ($6.8bn) cash (or $5.9bn net of cash in hand). The price of €42 ($59.35) per share is an 18 percent premium on the closing price of $50.27 on October 5. The deal completed in January 2008. This was the largest BI acquisition ever, and was almost certainly another response to Oracle’s acquisition of Hyperion earlier in the year.

The announcement came on the same day that Business Objects issued a profit warning for the third quarter of 2007. It blamed weak license fees for the shortfall. Despite including Cartesis, the quarterly license fees were expected to be only about $139m, compared to $131.6m in Q3 2006.

SAP says it will keep Business Objects as a separate business unit under its existing CEO, John Schwarz, who is expected to become a member of SAP's executive board. Bernard Liautaud, chairman and founder of Business Objects, is to be proposed for SAP board membership at the company's next shareholder meeting. In the meantime, he will serve as an adviser to SAP CEO Henning Kagermann.

Customers can read a more detailed analysis of the future of the many overlapping Business Objects and SAP products by clicking here.

CPM consolidation continues apace as Cognos buys Applix

Cognos announced on September 5, 2007 that it had reached a definitive agreement to purchase Applix for $17.87 a share, or $339m ($306m net of Applix cash on hand). The deal closed on October 26, 2007. This was the fourth CPM acquisition in 2007 (not counting Pilot).

Applix had always seemed a takeover candidate: it was much too small to remain as a public company in the Sarbanes-Oxley regulatory era, and was competing with companies far larger than itself. But it was nevertheless one of the fastest growing BI vendors, and TM1 is a well-respected OLAP server. Indeed, it is the only OLAP server to have grown its market share faster than Microsoft Analysis Services.

Cognos had at one time been a major OLAP vendor, but it has lost its way. PowerPlay has had little investment for the last few years, and Cognos 8 has not been seen as a strong OLAP client for other vendors’ OLAP servers. The older Cognos financial applications were also unsuccessful. Cognos has therefore had to resort to acquisition to re-establish its presence, with Adaytum in 2003, Frango in 2004 and now Applix in 2007. It will be interesting to see if TM1 continues to be marketed as a product in its own right in the long run, or if it is absorbed into the Cognos 8 platform.

Cognos is keeping the TM1 and Executive Viewer brands and products for some time, but will rapidly phase out the Applix name. TM1 is likely to be positioned as the Cognos tool for financial performance analytics, with less emphasis on its planning capabilities, as this role is assigned to Cognos Planning. A version of Cognos 8 that links directly to TM1 cubes is likely in the second half of 2008. The Applix products, as well as Cognos Planning (the former Adaytum) and Controller (the former Frango) will form a new Financial Performance Management division. This will be run by the former Applix CEO.

Both Applix and Cognos are likely to be affected by Microsoft PerformancePoint, which was also launched in September 2007. It is a low-priced, aggressively marketed competitor for Cognos Planning and TM1, which will get a lot of attention. The first release is not fully competitive functionally, but it may be adequate for many customers, and so is a threat to all existing CPM products. This is likely to have been a trigger for the sudden wave of CPM acquisitions in 2007.

SAP responds to Oracle’s acquisition of Hyperion by buying OutlookSoft

Continuing its PM acquisition spree, and in contrast to its oft-stated strategy of organic growth, SAP announced on May 8 2007 that it was to buy OutlookSoft Corporation. The deal closed on June 1, 2007. The price has not yet been published, but we think that OutlookSoft was worth no more than $150m to $200m; however, some reports have put the price paid by SAP as much higher, at about $375m.

It seems likely that this unlikely and probably over-priced acquisition was triggered by Oracle’s recent purchase of Hyperion, which was presented as an aggressive move against SAP. “Thousands of SAP customers rely on Hyperion as their financial consolidation, analysis and reporting system of record,” said Charles Phillips, president of Oracle, when the Hyperion takeover was announced. “Oracle’s Hyperion software will be the lens through which SAP’s most important customers view and analyze their underlying SAP ERP data.” If, indeed, SAP did significantly overpay for OutlookSoft, Oracle could be said to have already achieved its objective.

SAP already had its own SEM consolidation and planning software, based on SAP BW, but it is not as highly rated as the products from specialist vendors like Hyperion, Cartesis and OutlookSoft, all of which are now being acquired. SAP and Oracle will now be competing against each other in this area, as in so many others, but Business Objects (with Cartesis, ALG and SRC) and Cognos (with Frango and Adaytum) have also bought their way into this market. Microsoft will also be entering the same market at the end of 2007 with PerformancePoint Server 2007; unlike the other large vendors, Microsoft has taken a number of years to develop its own product, rather than acquiring one off the shelf.

OutlookSoft’s Microsoft-based technical architecture is very different to SAP’s. Although OutlookSoft recently launched an Oracle OLAP-based version, none of its production sites yet use it, and that version’s functionality and performance lag well behind the lead version based on Microsoft Analysis Services and SQL Server. In contrast, all of SAP’s products use relational rather than MOLAP databases.

Unlike some other CPM vendors, OutlookSoft was not an SAP partner and nor did it have a certified SAP connection or a NetWeaver version. Either SAP will have to significantly rewrite OutlookSoft, potentially degrading its performance and functionality, or it will have to adapt its own architecture to use Microsoft Analysis Services (which performs much better than SAP BW) when appropriate. But if it does this, its more numerous enterprise accounts using Oracle and DB2 databases may be dissatisfied, so it is more likely that SAP will rewrite OutlookSoft to use SAP BW (aka NetWeaver BI) instead of Analysis Services, even at the cost of greater complexity and much worse performance. Either way, non-SAP OutlookSoft sites are likely to reconsider their use of the product, reducing the value of OutlookSoft to SAP.

SAP’s subsequent acquisition of Business Objects (which had already acquired Cartesis, ALG and SRC) throws further doubt on the logic of the OutlookSoft acquisition.

Business Objects adds to its PM collection with Cartesis

Business Objects announced on April 23, 2007 that it had agreed to buy Cartesis for €225m (~$300m) in cash. The deal closed earlier than expected, on June 1, 2007. Business Objects and Cartesis are both originally French companies, headquartered in Paris, though in recent years, Business Objects’ focus and much of its management team has moved to North America. Cartesis was much less successful in North America, and remained as a mainly French company.

Although both Paris-based, the Cartesis and Business Objects headquarters are almost four miles drive apart whereas, strangely enough, the former ALG Paris office is literally next door to Cartesis, and overlooked by the Cartesis executive terrace. The Business Objects and Cartesis main UK offices are also less than a mile apart as the crow flies, both in Maidenhead, but in the US, they are at opposite ends of the country, in San Jose, CA and Norwalk, CT. It is no coincidence that the latter location is very close to the former head office of Hyperion Software, as the US branch of Cartesis was initially staffed mainly by ex-Hyperion people (the same is true of OutlookSoft).

Business Objects was reluctant and late to join the performance management race, and loudly derided Cognos when it acquired Adaytum. But after competitive pressure from Cognos and Hyperion, which were able to persuade many customers that they should be buying both BI and PM from the same vendor, Business Objects was forced to act. By then, many of the more obvious acquisition candidates had already gone — for example, Business Objects should probably have bought Adaytum, its then planning partner, rather than letting Cognos snap it up in 2003, and it could probably have bought Cartesis in 2004 from PwC for a fraction of the €225m it is paying Cartesis’ VC owners now.

Instead, Business Objects bought two smaller, less well known PM companies: SRC in 2005 and ALG in 2006. Both SRC and ALG had planning products, as does Cartesis, thanks to its acquisition of INEA in 2005. This puts Business Objects in the awkward position of owning three acquired, incompatible planning products, each from a different country (something that would not have happened had it purchased Adaytum in 2003 and Cartesis in 2004). At the same time, Cartesis has been beefing up its front-end analytical capabilities using an OEM version of Panorama NovaView, ironically replacing its unhappy former OEM use of Crystal Analysis Pro. It will probably now end up with a descendent of that product as its analytic front-end yet again.

Oracle buys Hyperion Solutions for $3.3bn

In by far the biggest-ever BI consolidation, Oracle has acquired Hyperion Solutions for $3.3bn cash, with the deal closing on April 19, 2007. By April 18, Oracle had purchased 91.3 percent of the outstanding shares for $52 per share. After the closing of the offer, the purchase of the remaining shares became compulsory.

Hyperion itself was based on a series of mergers and acquisitions, bringing together products that originated in at least a dozen separate companies (AppSource, Alcar, Arbor, Brio, Decisioneering, IMRS, Pillar, Razza, Sapling, Sqribe, Tableau and Upstream), and in Oracle, it has joined a BI/CPM stable that includes products from the former IRI Software, PeopleSoft, nQuire (which became Siebel Analytics) and Oracle’s own Discoverer.

There is considerable overlap between these various BI products: for example, there are several different relational reporting tools, and several OLAP servers. Clearly, not all will survive, and it is Oracle’s home-grown BI tools and applications that are the weakest. But it is likely that Oracle was most attracted to Hyperion by its market leadership in the faster growing CPM space. It puts Oracle in a better position to compete with SAP’s existing SEM CPM applications and Microsoft’s pending PerformancePoint. Previously, Oracle only had the ancient Oracle Financial Analyzer (which dated back to the 1980s) and its unsuccessful newer Enterprise Planning and Budgeting application. These are likely to be replaced by the equivalent Hyperion products, just as Essbase is likely to become Oracle’s primary OLAP server (despite Oracle’s claims that both Essbase and its unpopular own OLAP Option will both continue into the future).

Ironically, Hyperion had spent the previous 18 months starting the integration of its many tools into its new System 9 platform; now Oracle has to start this all over again, as it attempts to rationalize and merge its messy range of overlapping BI and CPM tools, and then integrate them with its other tools and applications. This could be a rough ride for both Oracle’s products and their users. But after PeopleSoft, Siebel and a series of smaller enterprise software acquisitions, Oracle certainly knows how to execute such deals.

Pilot changes hands for the sixth time

Pilot Software was acquired by SAP in February 2007. Depending on how you count, this is arguably the sixth time Pilot has changed hands.

Pilot Software had reappeared in the guise of Pilot Software Acquisition Corporation in May 2002, a new company formed by members of the Pilot management team backed by institutional investors. This reborn Pilot Software purchased Accrue’s Pilot and Hit List software product lines, including related intellectual property, customer contracts and other assets. Accrue received $1.5m in cash and retained rights to use the Pilot technology within Accrue G2. John D’Albis transferred from Accrue to resume his old role as CTO of Pilot Software, a role he continued to perform when SAP acquired the company in 2007.

Accrue had bought Pilot from Platinum Equity Holdings in August 2000 for approximately $20m in stock. Before August 1997, Pilot was owned by Cognizant Corporation. Dun & Bradstreet, Cognizant’s predecessor, had bought Pilot in November 1994, before which it had been independent since it was formed as an EIS pioneer in 1983. In 1987, in another all-paper deal, Pilot had acquired Thorn-EMI Computer Software, previously EPS Consultants, which was formed in 1972. Today’s Essbase also had its roots in EPS’s FCS product line.

Microsoft gets more aggressive with BI front-end tools by acquiring ProClarity

On April 3, 2006, Microsoft announced that it had agreed to acquire the privately held ProClarity. The price was not disclosed, but we speculate that it was close to or just below $50m, which is lower than ProClarity had been seeking in 2005. The deal closed in May 2006. The OLAP Survey 6found that ProClarity was the most widely used third party front-end for Analysis Services, which may be one of the reasons for this acquisition.

Unlike Microsoft’s three previous BI acquisitions, where Microsoft acquired just the technology from very young companies, and not the companies themselves, this time Microsoft emphasizes that it is buying the entire ProClarity organization and business. It says it intends to build on all ProClarity’s strengths, including its technology, sales and marketing, partner networks and future plans.

Unlike those previous acquisition, ProClarity has a relatively mature product line, now on its sixth major release, and some 1200 established customers. Microsoft will therefore be keeping most of ProClarity’s products unchanged in the short term. The one exception is the Java-based ProClarity Live Server, which will be withdrawn from the market, though some of its capabilities are likely to surface in future Microsoft BI products.

As part of this continuity, the ProClarity organization will become a Microsoft subsidiary and continue to be based in Boise, Idaho, and most of the management and employees are expected to be retained. The ProClarity sales force is likely to be integrated into (or form the nucleus of) Microsoft’s own growing BI sales and marketing organization. Microsoft wants to be seen as a fully-fledged BI vendor, not just a large software company with some random BI modules, and for this it needs some credible front-end tools, which ProClarity provides. Previously, Microsoft was hard-put to compete for enterprise BI business without the help of partners such as ProClarity.

Microsoft has not yet disclosed what will happen to the ProClarity name, nor the detailed road map of how ProClarity’s products will fit into the Microsoft product line or how they will be priced. But it is almost certain that they will remain as products in their own right, and not just be bundled with any edition of Office 2007. Previous experience suggests that Microsoft may reduce the prices and simplify the price structure. Microsoft will probably present the ProClarity-derived products as suitable for high-end BI, whereas Excel’s embedded BI capabilities will be aimed at mainstream users.

This move is bad news for other specialist front-end tools vendors for Analysis Services such as Panorama, Targit and Temtec, as well as for the larger BI vendors, who were previously able to shrug off Microsoft’s threat to their business because of the weakness of the then Microsoft front-end tools. Microsoft will now be a much more credible one-stop-shop for BI technologies, though it may attract less support from third party tools vendors who will see it as more of a competitor than a partner. It is probably no coincidence that two hitherto loyal Microsoft partners, Panorama and OutlookSoft, announced support for SAP BW and Oracle, respectively, on very the day of Microsoft’s announcement.

An indirect consequence could be to accelerate the much-expected, but stalled, consolidation in the BI industry. The larger independent BI vendors, now certain that they will have to compete with Microsoft for an increasing proportion of their business, may feel the need to fill any gaps in their product stacks, particularly if they had previously relied on Microsoft products to perform such roles. Equally, many of the smaller specialist vendors may feel that a long-term independent future is not a feasible option.

Business Objects switches strategy and acquires SRC

Having previously derided Cognos for diversifying into planning and consolidation applications, Business Objects reversed strategy and followed suit by acquiring SRC Software. The purchase agreement was announced on July 20, 2005, and the price is $100m cash. This is a high price to pay for a company with revenues of only about $26.5m in 2004 and 135 employees, though SRC’s 2005 run rate was probably closer to $35m. Business Objects expects to take a $5m charge on the conclusion of the deal.

This also adds complexity to Business Objects, as it has no previous experience developing, selling, marketing and supporting financial applications (unlike its main competitors, Cognos and Hyperion). This is one reason why Business Objects previously avoided acquiring or developing products in this area — it preferred to remain focused purely on BI (mainly meaning relational reporting, as its OLAP capabilities remain weak). Ironically, SRC’s products had previously been distributed in Europe by Frango, now part of Cognos, so many of the European SRC experts now work for Cognos rather than Business Objects. This will make it harder for Business Objects to boost SRC sales in Europe, as it must do to justify the high price it is paying.

With SRC being based in Portland, Oregon, Business Objects now has yet another development center, in addition to those in Paris, San Jose, Vancouver and Bangalore. It also has to embark on another round of integration, having just spent 18 months merging the former Crystal and Business Objects product lines. Of course, Cognos and Hyperion are also distracted by integration projects.

Once this deal concludes, Business Objects, Cognos and Hyperion will be competing across the board, though unlike the other two, Business Objects does not offer an OLAP server (apart from the discontinued Holos). Though they straddle the whole range, they each have different strengths and weaknesses: