Free Republic
Browse · Search
News/Activism
Topics · Post Article

Skip to comments.

Why we need a high-tech shakeout
McKinsey Quarterly ^ | March 9, 2003, 6:00 AM PT | Special to CNET News.com

Posted on 03/09/2003 10:10:12 AM PST by glorgau

The technology industry badly needs a shakeout: A consolidation of the myriad providers that sprang up in the 1990s would benefit both the industry and its customers. Yet because of the interlocking interests of executives and board members, this catharsis probably won't come from within. Interlopers from the edge of the industry or beyond will probably drive the change.

During the 1990s, the boom in high-technology spending spelled prosperity for new and established companies alike. The popularity of enterprise-resource-planning (ERP) and electronic customer relationship management (e-CRM) systems, preparations for the millennium bug," personal-computer upgrades to take advantage of the Internet and expanding corporate networks, demand for cell phones and personal digital assistants, and the telecommunications companies' rush to build new networks--all of these factors created about $1 trillion more in demand than trends would have suggested in 1989.

Now this demand is gone, and we know there was less benefit from the technology purchases than the new-economy prophets would have had us believe. The overhang of capacity is significant. In software, for example, the number of companies increased by 15 percent during the 1990s while the market grew by 12 percent. For these companies to meet the projected consensus earnings reported by Zacks Investment Research, average net margins would have to increase by 94 percent, to a record 18 percent.

At the same time, software investment rose to 11 percent of worldwide capital spending during the 1990s. Although software may gain a slightly larger share over the coming years, the sheer scale of the industry will probably limit growth in demand to a maximum of 8 percent to 12 percent.

Yet the companies formed are still with us. Some are substantial businesses with good long-term prospects. They will invest through this downturn and emerge as leaders; think of Intel riding out the perilous financial times of the mid-1980s as the initial PC bubble burst. Many companies, though, are investing for a future they cannot reach. They need to be restructured--downsized, merged, acquired or liquidated.

Many companies, though, are investing for a future they cannot reach. They need to be restructured--downsized, merged, acquired or liquidated.

But the restructuring hasn't happened. Managers and boards of high-tech companies share an interest in maintaining their positions. Because high-technology companies favor executives in the industry as board members (for their expertise), these informal coalitions cross company boundaries. The investment bankers and private-equity firms that might propose a restructuring fear disrupting their relationships with this community and reducing the odds of winning future business from it.

In fairness, these executives and directors are also cautious because of the conventional wisdom that hostile deals do not work in high technology and that most mergers have not been successful. This exaggerated belief is grounded in a misreading of history. When the last extended downturn hit in the 1980s, a much smaller high-tech industry was populated by a fraction of the number of companies that compete in it today. More important, the installed base of enterprise technology was smaller, so predictable revenue streams for support and maintenance were neither as large nor as stable as they are today.

Moreover, the potential benefits of consolidation are quite substantial. Most enterprise-software companies, for example, spend 25 percent to 35 percent of their revenue on selling, general and administrative expenses. If companies selling to the same customers were consolidated, the new company could slash these outlays. Customers would benefit as well: The plethora of vendors has left most IT organizations with confusing and expensive choices.

Further, many companies are investing in research and development for product lines that have a dimmer future than management would like to believe. These outlays can run to an additional 15 percent or so of revenue. Coolheaded managers can find savings here as well.

Last, a number of companies are trading at or below the value of the cash on their balance sheets. There is economic value to harvest either by finding tax-efficient ways to return that cash to investors or by merging companies to channel it to businesses with a brighter future.

In the larger context, restructuring could spur the industry to do a better job for customers. Customer spending will probably pick up when enterprises become convinced that they can get top- and bottom-line benefits from their technology investments. Companies need tailored solutions and help to make their organizations, business practices and processes more productive. Many small technology companies cannot accomplish this.

What will it take to open the gates to restructuring?

Most likely the industry will require outside intervention. Much as T. Boone Pickens and Drexel Burnham Lambert shook up the energy and other industries in the 1980s, people and companies outside the circle that formed during the prosperity of the 1990s, may show the way.

Most likely the industry will require outside intervention.

Many leaders of the restructuring of the 1980s are reviled for the excesses of that period, but their actions ultimately made U.S. industry more competitive, laying the foundation for the burst of productivity and growth that blessed the U.S. economy for much of the 1990s.

This is a task worth undertaking. As the industry works through the current slump, conditions are ripe for a return to average annual growth rates of 8 percent to 10 percent. Top companies will outperform those levels. The sooner the restructuring gets under way, the sooner there will be a chance for better times

For more insight, go to the McKinsey Quarterly Web site.

Copyright © 1992-2003 McKinsey & Company, Inc.


TOPICS: Business/Economy; News/Current Events; Technical
KEYWORDS: technolgy
We still have a way to go...
1 posted on 03/09/2003 10:10:12 AM PST by glorgau
[ Post Reply | Private Reply | View Replies]

To: glorgau
Most likely the industry will require outside intervention.

Is McKinsey implying that the industry will need government imtervention?

2 posted on 03/09/2003 11:15:17 AM PST by Slyfox
[ Post Reply | Private Reply | To 1 | View Replies]

To: glorgau; Dominic Harr; Bush2000; chance33_98; HAL9000
I tend to agree.

There are way to many vendors out there for certain business areas. I looked into data mining tools a couple of years ago and may have found at least half a dozen vendors outdide of SAS and SPSS.

Similarly the ETL tools market is way overcrowded. It seems that Informatica is now becoming somewhat of a standard, but there are all sort of other tools out there which do similar tasks, in addition to the offerings from the major database vendors. And of course, in IDE's and UML modeling tools and source code repositories, there seems to be a plethora of options.

The problem with this variety is that you can never find anybody trained on these obscure tools. And nobody wants to specialize in a tool that has a 10 per cent market share.

My biotech clients are the worst, because they have created specialized environment for data mining of DNA and protein sequences using whatever tools came to mind, and now they are looking for someone with bizarre skill combinations which don't exist (known in the trade as "purple squirrel" positions) instead of looking for solid developers who know all the basic techniques. Since they have no prior experience with information technology, and since they tend to be run by propeller-head type academics, they can make the worst clients.

On the other hand, I am opposed in principle to monolithic applications of any type for large, Fortune 1000 clients. Sure, they have already sunk their teeth into SAP, PeopleSoft and the like, but the idea that Oracle or anyone else can produce a suite of applications, once size fits all, to run and entire multinational company seems ludicrous to me personally. I know it happens, but that is my position.

For smaller clients whose IT budgets are smaller, I can understand using a commercial off the shelf application. However, I have one client who is a large regional bank with extremely limited client/server experience (and a massive mainframe back end) who is so afraid to develop anything internally that they try to do foolish things like turning channel applications for limited purposes into enterpise data stores.

These applications were never meant for such purposes, need to be customized beyond recognition (and beyond decent tech support from the vendor) and don't have the robustness that is needed for a mission critical enterprise application. However, that is the situation I have to deal with until the client realizes the foolishness of its ways.

Personally, I belive that the wave of the future lies in componentized applications running on top of application servers such as Web Sphere, BEA, Sun One, Apache/Jboss and even .Net for smaller clients. In such situations, applications share a common transaction bus, can be removed on somewhat of a plug and plag basis, and can be integrated as legacy applications using Web Services, which seems to be the latest, hottest trend.

I would be interested in your thoughts on these issues.

3 posted on 03/09/2003 11:19:49 AM PST by Fractal Trader
[ Post Reply | Private Reply | To 1 | View Replies]

To: Fractal Trader; Bush2000; Dominic Harr
I would be interested in your thoughts on these issues.

I have no thoughts on this yet, seeing as how I am eagerly awating Chelsea Clinton's opion on this... See Chelsea Clinton Gets Six-Figure Job and seeing how she will be working at McKinsey ;)

4 posted on 03/09/2003 11:29:14 AM PST by bwteim (IMHO)
[ Post Reply | Private Reply | To 3 | View Replies]

To: Fractal Trader
OSI for Databases ;)
5 posted on 03/09/2003 11:33:15 AM PST by chance33_98 (Government cannot give freedom, it can only take it.)
[ Post Reply | Private Reply | To 3 | View Replies]

To: bwteim
>>> I would be interested in your thoughts on these issues

hahahahahahahahahahah

Don't you just love these investment house types who have all the answers, and seem to miss the boat every time. Maybe they should consult with Chelsea.

snooker
6 posted on 03/09/2003 11:39:49 AM PST by snooker
[ Post Reply | Private Reply | To 4 | View Replies]

To: Slyfox
Is McKinsey implying that the industry will need government imtervention?

Nope, they are saying that somebody from outside the industry, be it a Warren Buffet or whomever, is going to come in and "restructure".

7 posted on 03/09/2003 11:43:31 AM PST by glorgau
[ Post Reply | Private Reply | To 2 | View Replies]

To: glorgau
BTTTTTTTTTTTTT
8 posted on 03/09/2003 11:53:33 AM PST by dennisw ( http://www.littlegreenfootballs.com/weblog/weblog.php)
[ Post Reply | Private Reply | To 1 | View Replies]

To: Fractal Trader
Yep, Web Services is a comer. It just makes sense. Though the bandwidth is high for these sort of apps. XML is like they used to say about Cobol - "It puts more more words into fewer ideas than any other language".
9 posted on 03/09/2003 12:15:35 PM PST by glorgau
[ Post Reply | Private Reply | To 3 | View Replies]

To: glorgau; Bush2000; Dominic Harr
Yeah, it always seems to go the way of "easier to code, harder on the CPU."

I attended recent .Net seminar and came away with the realization that the only decent way to communicate with Java applications is through Web Services, or else you are stuck with the dreaded COM interface and .dll hell. What than means is that if you have a .Net application, you will have extremely high overhead if you try to communicate with a Java or other non-supported application.

If you try to code to another .Net program, you can communicate using a simple, fast interface running on top of the CLR. I don't understand why, from a technical or legal perspective, they can't support Java in the CLR. Reverse engineering is always legal, except possibly under the legal sledgehammer of the Digital Millenium Copyright Act.

What Microsot has done with .Net violates component architecture and will probably get them in trouble with large enterprises which have Java and Cobol in the back end (and worse).

In terms of Web Services, I am writing a forthcoming book for Manning on the subject. From what I can tell, it will put pressure on CPUs and, more importantly, I/O interfaces. Lots of internal fiber optics supporting clustered apps on server blades seems to be in store for us.
10 posted on 03/09/2003 12:26:33 PM PST by Fractal Trader (Broadand Rocks! Hope Saddam doesn't have a cable modem and access to Free Republic!)
[ Post Reply | Private Reply | To 9 | View Replies]

To: Fractal Trader
On the other hand, I am opposed in principle to monolithic applications of any type for large, Fortune 1000 clients. Sure, they have already sunk their teeth into SAP, PeopleSoft and the like, but the idea that Oracle or anyone else can produce a suite of applications, once size fits all, to run and entire multinational company seems ludicrous to me personally. I know it happens, but that is my position.

I take a similar view. I like MSSQL and use both it and Oracle 8i. While using third party tools can be efficient I think that overall a method is more important.

Control, documentation, and information sharing is lax just about anywhere I have gone. Companies lack a coherent vision and plan, and a way to implement it.

Writing your own data mining programs, or even your own structures in something like PERL, is not the problem (and can often lead to better solutions, similar to what some research labs are doing with TCP as a google.com/unclesam will reveal). The problem, to me anyway, comes in when there is no control and documentation of such things. People failing to work together isolates good solutions and creates a myriad of problems. I may not need an oracle database for the project I am working on, but if I want others to be able to communicate with my data structures and mine them for information, or add to it, I need to make a general method for interfacing with it that is easy to use and understand (which it appears from your comments you have seen many who do not).

More simply stated, if I create my own structure but make it accessible through something like crystal or toad (and supply odbci drivers or whatever is needed) then having something out of the mainstream can work nicely. Clear documentation of back end for those who may need to administer it should be readily available (in case I get canned).

This problem is evident though in mainstream apps as well. MsSql data bases may have scripts, triggers, stored procedures, and so forth that while easily accessible can be difficult to put together for a broader view when someone needs to do work on them. Standards in documentation and procedure could make work flow much easier for all involved, no matter which product (off the shelf or custom) a company chooses to use.

IT dept's have, in cases I have seen, become less integrated in some areas from the rest of the company - integration and problem solving being examples. At a previous company I worked for most people knew how to use access and we wound up with over 600 access data bases strewn about. Most of them shared common information which related to our customers and their needs. Had these people tapped our dept we could have provided a much better soltution which would have saved them time, headaches, and money. But people figured since they could just throw things easily into a spreadsheet or access they did not need our help (actually, my help, as I was the whole IT dept at that site). I tried getting managers to pull the info and needs together for their people so we could consolidate but they always had other priorities (and none of the ones I found were documented).

The solution to me then is strong leadership and standards within a company.

11 posted on 03/09/2003 1:10:56 PM PST by chance33_98 (Government cannot give freedom, it can only take it.)
[ Post Reply | Private Reply | To 3 | View Replies]

To: Fractal Trader
What Microsot has done with .Net violates component architecture and will probably get them in trouble with large enterprises which have Java and Cobol in the back end (and worse).

Not if you're using J# ... http://msdn.microsoft.com/vjsharp/default.asp.
12 posted on 03/09/2003 2:12:12 PM PST by Bush2000
[ Post Reply | Private Reply | To 10 | View Replies]

To: Fractal Trader
In terms of Web Services, I am writing a forthcoming book for Manning on the subject.

Excellent!

From what I can tell, it will put pressure on CPUs and, more importantly, I/O interfaces. Lots of internal fiber optics supporting clustered apps on server blades seems to be in store for us.

Of course -- but that's part and parcel of adding any new layer of abstraction. Previously, people were using custom RPCs (DCOM, Corba, etc) -- binary crapola that could be packed and unpacked very efficiently. Using XML as your data storage container means a lot more IO (parsing, etc). But don't forget that processors, memory, and storage subsystems keep getting larger and faster all the time. It wasn't all that long ago that people questioned wither a PC was fast enough to run a GUI... ;-p
13 posted on 03/09/2003 2:16:17 PM PST by Bush2000
[ Post Reply | Private Reply | To 10 | View Replies]

To: Bush2000
It wasn't all that long ago that people questioned wither a PC was fast enough to run a GUI... ;-p

In some regards, I am beginning to believe that Web Services are to back end application interfaces what the GUI was to the Macintosh and Windows.

Sigh, maybe I should just spend my time learning to write and interface with fibre optic drivers.

14 posted on 03/09/2003 2:22:10 PM PST by Fractal Trader (Broadand Rocks! Hope Saddam doesn't have a cable modem and access to Free Republic!)
[ Post Reply | Private Reply | To 13 | View Replies]

Comment #15 Removed by Moderator

To: chance33_98
Control, documentation, and information sharing is lax just about anywhere I have gone. Companies lack a coherent vision and plan, and a way to implement it.

Actually, there is a school of thought that says "If you document it, you can be replaced"

16 posted on 03/09/2003 10:12:29 PM PST by glorgau
[ Post Reply | Private Reply | To 11 | View Replies]

To: glorgau
Actually, there is a school of thought that says "If you document it, you can be replaced"

Which is why I didn't at my last job ;) I was literally the last employee out of the building and even then they kept me on to consult.

17 posted on 03/09/2003 10:14:31 PM PST by chance33_98 (God gave man freedom, government took it away)
[ Post Reply | Private Reply | To 16 | View Replies]

To: Fractal Trader
If you try to code to another .Net program, you can communicate using a simple, fast interface running on top of the CLR. I don't understand why, from a technical or legal perspective, they can't support Java in the CLR. Reverse engineering is always legal, except possibly under the legal sledgehammer of the Digital Millenium Copyright Act.

That is because .Net is a Microsoft product. Web Services don't have to be done with .Net. Lately, anything that MS does is done with a nod towards standards but somehow they never seem to interoperate nicely with other company's standards based products. Or if they do now, they won't in the next release :-)
As an example, I've just been going through the low-level LDAP that Active Directory puts out. Stuff that works fine with other servers is subtley different coming out of Windows 2000 Advanced Server. And the fun little things they do with the schema!
The MS API that they put out for interfacing with Active Directory works wonderfully if you stay inside their environment, but performing operations with LDAP is much more problematic and not nearly as well documented.

18 posted on 03/09/2003 10:27:09 PM PST by glorgau
[ Post Reply | Private Reply | To 10 | View Replies]

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
News/Activism
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson