IBM buys Cognos

I have been saying to anyone who asks, for
the last two years, that IBM would not buy Cognos. This was largely based on
discussions I had with senior IBM executives in 2005 about who IBM might buy in
the BI space and what they would be looking for. IBM’s view (with which I
concurred) was that if it was going to acquire a BI vendor then what it wanted
was a next generation capability. And both IBM and I felt that Cognos, Business
Objects and the like were previous generation technologies that would find it
hard to transition to an environment that would support any query, at any time,
against any data, without IT department involvement. And that remains the case.

Moreover, although IBM does not regard
Cognos as an application vendor, its offerings in the performance management
space clearly are applications. What else would you call planning, budgeting or
financial consolidation? So, although IBM will no doubt attempt to split hairs
about the definition of what constitutes an application it is, in my opinion,
going back on its long-standing commitment not to (re-)enter the application
space. And I am sure that both SAP and Oracle will think so too, and trumpet
the fact whenever they feel that that is to their competitive advantage. And,
of course, the big question is: is this just the thin end of the wedge? Time
will tell but that’s another discussion.

So, anyway, what has changed? Well, as
Harold Macmillan famously said, “events,
dear boy, events”. Said events being Oracle’s clear intention of becoming a
major BI vendor and SAP’s recent acquisition of Business Objects. Now, Steve
Mills of IBM has clearly said that IBM neither makes defensive acquisitions nor
goes shopping simply because other people do. However, he would say that
wouldn’t he? Now, you could argue that there is no way that IBM could pull off
a $5 billion acquisition within a month and that’s not an unreasonable position
to take. However, Oracle signalled its intentions a lot longer than a month
ago. Moreover, since it bought Sunopsis a year ago it very much had the same
overlaps with Business Objects that IBM did: but not Cognos. So, if Oracle was
going to acquire another BI vendor it was likely to be Cognos and that was too
big a risk to face. In my view IBM saw itself between a rock and a hard place:
it could gamble that Cognos would stay independent or get bought by HP but it
couldn’t afford the possibility that they might be acquired by Oracle. So, I
find it hard to believe that that threat did not play at least some part in the
decision to acquire Cognos.

Actually, there’s a second point we need to
consider. If IBM was looking two years ago at acquiring a second generation BI
technology, why hasn’t it done so? Why didn’t it go out and buy QlikTech or
Information Edge or somebody else doing innovative things a year or more ago?
In which case it wouldn’t have been forced into this bind. I don’t know the
answer to that. Concerns over scalability, maturity, market momentum, lack of
clarity about the right direction to go in? Your guess is as good as mine.

Leaving all that aside, who’s going to be laughing
all the way to the bank and who’s going to be crying into their coffee once
this is all done and dusted? Well, in the former category you can certainly put
Cognos shareholders and as for the rest it rather depends on your view of
independent vendors. Mine is that they enrich the market, frequently offer more
advanced capabilities than the major players, and that there is a significant
market for such independent suppliers, which IBM is now cutting Cognos out off.
If you agree with me then all the remaining pure play BI vendors, especially
SAS but not to forget MicroStrategy, Information Builders, QlikTech et al,
should be very happy with this news, as will independent performance management
vendors such as Infor. Of course, if you’re at the sort of company that wants
to buy all of its food at Tesco or Wal-Mart and never wants to try a Harrods
hamper or a farm shop, then you’ll probably disagree with me. SAP/Business
Objects could be a big gainer too, providing Business Objects remains (in
effect) independent of SAP as Cognos will certainly not be remaining
independent in any sense.

Crying into their coffee will be
Informatica and Composite Software both of which are major partners of Cognos,
whose areas of expertise will no doubt be attacked by IBM salespeople who want
to sell IBM’s own data integration and federated query capabilities. However,
this will be complicated by the fact that Cognos owned part of Composite
Software, which means that now IBM does. Further, Informatica embeds
Composite’s technology along with PowerCenter, which means that IBM will be
supplying (in part) software to Informatica. So maybe it will only be latte
that they are crying into and they won’t need the really strong expresso (and
in any case, these only represent minority portions of their respective
businesses). As a corollary, this is also potentially bad news for Cognos users
of these products as these relationships are likely to fray over time, though of
course IBM supports heterogeneous environments. Cognos users of that company’s
ETL product may not be too happy either. SymphonyRPM and its customers may also
be upset because the company plays in the performance management space and is
therefore now a competitor to IBM despite the fact that IBM’s AlphaBlox is one
of the building blocks that underpins SymphonyRPM.

For most users, IBM will no doubt do a good
job of continuing to enhance the Cognos products though it will be a tough task
to convert its product line into a next generation one. However, I do think
that this will make it more difficult to sell new Cognos implementations (both
BI and applications) outside of the existing user base and outside of dedicated
IBM shops.

The acquisition also poses some interesting
questions. For example, it might make sense for IBM to reinvigorate its data
mining offering: it is only relatively recently that it stopped marketing this
as an independent product (instead bundling it with the DB2 warehouse) and it
could make sense to reverse that decision. Then there’s the question of
AlphaBlox: currently it is in the data warehouse group but it would make sense to
push it in with Cognos (and the same is true of data mining). But then there is
the issue of whether IBM needs a separate data warehousing group at all if it
has a database group on the one hand and a BI group on the other. We will have
to await answers on all of these.

To sum up: I can see why IBM felt that it
had to make this acquisition and I am sure that it will do the best that it can
by both the Cognos products and its newly acquired Cognos customers. What I am
disappointed about is the fact that it became necessary.