Oracle’s perennial confusion about analytic technology
Oracle is badly confused about analytic technology, and indeed long has been. It would be tough for me to coherently explain why without being, well, confusing. So I’ll just list a series of data points, which hopefully should suffice to illustrate the point.
- Classic BI tools have at various times fallen under the purview of the app dev tools group and the app server group.
- Data mining, stemming from a Thinking Machines, Inc. acquisition, is under a whole other group on the East Coast.
- That group is now collocated and somewhat integrated with the group that oversees the MOLAP database capability, which came in via a different Boston-area acquisition (IRI/Express).
- While Oracle brags of its integrated BI stack, enterprise reporting is an exception.
- Discoverer 1.0 (Oracle’s original BI tool), was one of the most impressive new products I ever saw. But then BI technology at Oracle almost stagnated. The reason seems to have been largely a series of platform ports – client/server, Java client, thin client, etc. Other BI vendors faced the same problems, however, and they now have products generally agreed to be ahead of Oracle’s.
- Oracle didn’t seem to have a coherent analytic apps strategy even before the Peoplesoft acquisition, which obviously just confused things further. (Of course, neither does SAP, really, Dennis Moore’s passionate insistence to the contrary notwithstanding.)
- ETL/data integration is of course a historical Oracle sore spot.
That’s even before getting to Oracle’s problems in data warehousing itself, where it can’t beat Teradata and DB2/mainframe at the very high end, and low-cost options like Netezza are a looming threat as well.
What’s particularly ironic is that some of Oracle’s core marketing pitches have a lot to do with analytics. The whole integrated stack story? Doesn’t make much sense when you’re only talking OLTP; only with analytics in the picture is it coherent. The whole scalability story? A few huge websites and the like aside, that’s mainly about data warehousing now.
Obviously, Oracle has the potential to be a titan in analytics. But it doesn’t have its act at all together yet.
| Categories: Analytic technologies, Business intelligence, DBMS vendors and technologies, Data mining, Oracle | 2 Comments |
Why Oracle doesn’t “get it” about apps
Since the mid-1980s, Oracle has put huge investment and market strength behind its apps efforts. Given those resources, success has been extremely limited. Obviously, there are many reasons for this run of (relative) failure, including a lot of internal management/cultural/political issues. But much of the problem can be summed up in one short phrase: Oracle doesn’t fully understand the importance of business process.
The majority of an application can be created by the following three-step process:
1. Design the right database.
2. Automatically generate the add/change/delete/query/report functionality.
3. Add other BI/analytic functionality as needed.
And Part 3 is a relatively recent addition in most cases.
But that’s not the whole story. What’s left over can be described as “business process,” which is where SAP shines. And Oracle underrates business process. To see what I mean, go to SAP’s web site, search on “business process,” and look through the first few pages of results. Then try the same exercise on Oracle’s. There’s a dramatic difference. Siebel’s and Saleforce.com’s sites also talk much more about “business process” than Oracle’s does.
So far as I can tell, Oracle has always believed that if you design the right database — and create the obvious interfaces to it — you’ll have a great application. I remember Larry Ellison seeming to believe that in very early days, before Jeff Walker led the first high-investment Oracle apps effort. The whole CASE-model-based vertical apps strategy of the early 1990s clearly depended on that premise, and probably failed because of the premise’s flaws. And finally, I remember a bizarre conference call in connection with the release of some generation of Oracle’s app dev tools in the late 1990s, which Larry touted as one of the most important events in the entire history of software. I can’t imagine him saying that unless he thought that these tools would automatically generate your apps for you. But they couldn’t actually do that – and the extent to which they couldn’t was almost exactly the extent to which they didn’t capture the business process aspects of the app.
So will Oracle overcome its longstanding business process blind spot, now that it’s made such a huge bet on apps? It well might. But until it does, Oracle’s prospects in the app business aren’t good at all.
Is Oracle headed for hard times?
A lot of you probably remember me writing about Sybase in 1994. (And yes, Tony Percy did admit at least that Gartner sped up its research drastically after I published, so as to follow me into print a month or so later.) Well, I’m not at that level of certainty yet this time. Indeed, I haven’t even sold all my stock in the company I’m worrying about. But I do increasingly find myself wondering: Is Oracle headed for hard times?
My nonobvious reasons for concern fall mainly into three large areas (follow the links for more detail on each):
- Oracle may be losing its edge in DBMS. Or at least (to brutally mix my metaphors), there are some cracks in the colossus.
- Oracle has never “gotten it” in applications.
- Oracle is perennially confused in analytic technology, which is becoming ever more important.
Obvious reasons for concern include the difficulties of integrating large acquisitions, general slow growth and price pressure in the technology sector, and a rolling management transition at the top of the company.
I’m far from ready to call the turn with assurance, however, because Oracle also has some formidable strengths. It indeed holds an excellent position in its core DBMS business, and by the numbers is very strong overall. Charles Phillips was once the best software stock analyst ever, which may make him the single person with the greatest understanding of software industry strategic dynamics. And Larry Ellison, detached as he may seem (and actually be) at times, has an amazing track record of making good decisions before it’s too late.
So I’m not yet predicting that Oracle will fall; I’m just pointing out some of the key issues it has to address if it is to remain prosperous.
| Categories: DBMS vendors and technologies | 1 Comment |
Autonomy/Verity merger
As posted in the Text Technologies blog, I’m skeptical yet slightly hopeful about the combined Autonomy/Verity company. Each of those companies sells overall technology that’s less than the sum of its parts. Maybe the merged company will be big enough to wake up, add what’s missing, and grow the enterprise text search market beyond its current level of:
1. Collection of niche markets plus
2. Unimportant universal technology add-on
| Categories: Analytic technologies | 1 Comment |