A talk given to the Canadian Information Processing Society and the Data Processing Management Association, Vancouver, B.C.
APRIL 10, 1979
Why is it important to review this segment of the computer communications industry at this time?
For years the excitement in the computer industry has centered around new hardware announcements or, in the last decade, new software advances. Now the computer services industry, which draws these two together, is in many ways the trend setter.
The computer services industry is the fastest growing segment of the computer field. According to the work done by the Computer Communications Secretariat it could outdistance both the hardware and communications parts of the industry in terms of revenue by 1985.
The computer services industry represents in microcosm the trends we can see in the industry as a whole, and by its nature must be at the leading edge.
We can start to meaningfully analyse the industry as it now has a ten year track record in its modern form. We can date the modern era from about 1968.
Finally, it is worth noting that although I will discuss the commercial service bureau industry, many of you run what are in fact internal service bureaus. The trends seen in the commercial sector apply just as readily to the in-house service bureaus.
THE INDUSTRY AS IT IS
In some ways this is an odd time for self-examination. The industry is just coming off its best year ever
• industry revenues exceeded $600,000,000.
• the industry provides employment for over 11,000 Canadians.
• the industry has enjoyed a growth rate of over 20% in revenue and has finally attained a level of profitability sufficient for it to reinvest adequate funds in long range R&D.
However, the industry could be in the same position as many happenings that looked their best just before a decline.
Are we the dinosaur of several million years ago, the stock market in the mid 1920’s, or the nuclear aircraft carrier in the early 1970’s?
The problem with many of these seemingly successful situations was the inability to foresee and adapt to a changing environment. Even with perfect foresight, size or momentum can make it difficult to change.
Is the computer services industry in Canada in this situation?
A DYNAMIC DECADE
We can get some perspective on where the industry is going by being very honest about what really happened to get the industry where it is today.
During the late 60’s many books and articles were written about the benefits of the Computer Utility industry, the coming Information Age, the Wired City, etc. But really there was a particular set of circumstances that allowed a number of entrepreneurs to capitalize (literally and figuratively) on several important trends:
a. Large universal computer systems became available in the mid-60’s with operating systems that allowed multiprogramming.
b. Data communications facilities developed to the point where they were usable on a large scale.
c. There was a buoyant stock market.
The economy of scale of the new/360 Series, 1108’s, and similar large machines was a significant advance over their predecessors. More important, the generalization and power of the operating systems meant that the application programmer saw himself as interfacing with the operating systems rather than the computer. Physical nearness to the computer was no longer necessary. The powerful systems also allowed meaningful sharing of the computer resource.
Data communication had been a laboratory curiosity before the early 60’s, but was finally recognized by the common carriers as having real revenue potential. This ensured that attention was paid to it.
Finally, there was at that time a real financial incentive to undertake new ventures (for the entrepreneur B.C. refers to Before Capital-Gains-Tax).
The industry was an opportunistic response to its times.
It was amazingly successful, considering it was developed around the available technology rather than in response to specific user needs. It has always been an ‘alternative industry’ with other feasible approaches for the user readily available.
Two reasons for its success were
1. It demonstrated considerable dexterity in adapting technology to meet the needs it had convinced people they had.
2. It realized that with good marketing, you can sell anything.
To provide a preliminary answer then to the question posed at the outset, I believe that the computer services industry can avoid going the way of the dinosaur if it concentrates on imaginatively adapting new technology to client needs and on good marketing. Come to think of it, this applies to most industries.
TRENDS – PERCEIVED AND REAL
Before we examine where the industry is going, we should quickly look at where it is in terms of what we perceive we sell, and the user perceives he is buying.
1. WE SELL CYCLES
The industry believes that it sells a sophisticated form of computer cycle which provides a viable alternative to what a user could do on an in-house installation.
These cycles are usually sold to data processing people.
They are usually bought because these data processing people:
• need access from time to time to a very large machine, but would prefer to pay for the privilege by the pound (or the kilo!).
• they need assistance with peak loads.
• they would like a buffer against the constant change in hardware and software
• that is expensive and time consuming to implement.
• they need access to specialized software.
• they are trying to improve their programmer productivity which is their most expensive resource by ensuring the cycles he or she needs for testing will be available when needed.
In a word, users buy these sophisticated cycles because of the improved cost effectiveness and advanced function.
2. WE SELL RESULTS
The industry believes it is adding value to the client’s data processing needs by providing application oriented systems. It is the approach of solving a problem for the customer, rather than simply doing what he is already doing in a more cost effective manner.
Such services are usually sold to non-data processing people, i.e. end users.
These end users usually buy these services because:
• they are looking for a form of expertise that does not exist within their own organization.
• they do not want to build up staff in-house for what may be a temporary or specialized need.
• they likely do not have the expertise to manage internal staff of this type even if
• they could acquire it.
• they believe they can get faster implementation by contracting outside.
• they may be able to get a less expensive result because of prior experience or prepackaged approaches or programs developed by the supplier.
In a word, these people are buying something they believe they cannot get as effectively internally.
THE CONVENTIONAL WISDOM
The conventional wisdom in the industry says that the latter approach is the way of the future.
WHY? The argument is that:
The cost of cycles is declining. Therefore, it is getting cheaper to install in-house hardware. The economy of scale is not what it used to be. Now the mini is making decentralized processing not only possible, but practical.
The complexity of large operating systems that allow multi-programming is getting in the way of economy of scale. Too much of the time of a large central processor is used up with unprofitable overhead.
While the cost of cycles is declining, the cost of data transmission is not declining nearly as rapidly.
Finally the user always really wanted his own machine and only the temporary cost advantage of the 70’s caused him to go outside.
Therefore, the need for large central cycle selling is doomed. The only solution is selling value-added services.
FACTS VERSUS THEORY
I was one of those who not only believed the conventional wisdom, but tried to implement it. After several years of trying this at SDL I have had to admit that cycle selling still pays the overheads. I have been unable to grow the value-added business to more than about 25% of our total revenues. Even this was done largely through a series of acquisitions.
Further, the machine replacement business which is really cycle selling taken to its logical conclusion, is doing just fine. This approach is the ultimate in Facilities Management, where the facility itself does not even exist on the customer’s premises.
This business which everyone predicted would decline with the falling cycle costs, is so healthy that at Datacrown we now have more equipment installed than all but three other organizations in Canada (all three happen to be banks). The business has not only grown to be large – about $50 million in revenue last year for Datacrown and SDL together – but profitable. For 1978 this rate of revenue produced a profit of over $5 million pre-tax.
What is really happening?
I believe that what we have done is lead the computer industry in a fundamentally new direction.
WE HAVE CHANGED PEOPLE’S FOCUS FROM MANAGING HARDWARE TO MANAGING INFORMATION.
Having made this breakthrough, the sophisticated user is not likely to revert to the time consuming diversion of managing the process rather than the data. The computer services industry has a fundamental reason for existing.
However, there are other reasons why the cycle selling business is booming:
• while cycle costs are declining, people costs certainly are not. It still costs a great deal to staff computers for multiple shifts, pay scarce systems programmers, and provide the management time for a major installation.
• software is becoming more costly as more unbundling occurs. A service bureau can provide one central copy of a program, rather than each individual having to pay the full cost.
• not all the costs of computing are declining at the same rate. For example, peripherals such as printers are not getting substantially cheaper. Economy of scale still exists.
• communications costs are high if the facilities are not fully utilized. A large computer services supplier can provide a major shared network, thus reducing costs to all users.
• the user today is sophisticated and is now quite willing to accept that external facilities are at least as secure as internal computer centres.
In a word, the user will buy a well managed pool of cycles because of its overall cost effectiveness. It also relieves the user of the endless hassle of management supervision of a constantly changing environment.
WHAT ABOUT THE NEXT DECADE
(a) Specialization – Sophisticated cycle filled networks will be provided by a few very large companies. This need for large capability, is one of the reasons for recent merger of SDL and Datacrown. I predict there will be a number of other mergers in the coming months.
Co-existing with these large national companies, will be a number of local suppliers of custom programming. Such work by a local supplier may be economic because it will not have to bear the overheads that would be applied by a major network supplier. The economy of scale does not necessarily apply to custom programming.
There will be some national organizations who will provide both value-added and network services, but I believe these will only be successful if they organize their value-added product lines as separate specialized divisions. These divisions will operate like small companies with their own sales force marketing to particular end users.
I do not believe that single supplier responsibility is as important as it once was in the computer field. The growing use of mainframes, peripheral equipment, software and common carrier services, all from different suppliers has led the sophisticated user to be fully prepared to buy cycles from one supplier and custom programming from another.
The result, in Canada, may be two or three very large national network suppliers with perhaps two or three hundred smaller suppliers of software.
(b) Systems Designed Around Data – The widespread use of improved Data Base Management Systems will encourage users to concentrate on the management of data rather than the network. I believe this to be a far more fundamental change than the widely discussed Distributed Processing concept.
Distributed Processing will play a role, but will become only a cost/effectiveness consideration. It is not a fundamental breakthrough.
Although a certain amount of intelligence will certainly be distributed, a large central machine somewhere in the network will still be required for:
• the processing of very large files.
• the consolidation of data from many network sources.
• integrated corporate modelling, and other projection techniques.
• availability of centralized information banks that would not be worthwhile keeping up to date on a distributed basis.
• availability of specialized programs where it would be uneconomic to keep these on a decentralized basis.
Users will depend on computer services network professionals to optimize the delivery system which will include a consideration of the best balance between centralized and distributed processing. The user will concentrate on the data and the application.
(c) Growth Of The Stand-Alone Mini Installations – The growth of the one machine/one application approach will be important, but primarily for small businesses. This will have little effect on the computer services industry as there is not much concentration in this area in any case.
IBM recognized that the small business market is not a good service bureau market when it sold the Service Bureau Corporation some years ago. I believe that if IBM re-enters the computer services field it will be through the Satellite Business Systems approach of concentrating on the Fortune 500 companies. The non-IBM computer services industry can approach a broader customer base, but still will not likely be able to market effectively to very small businesses or the home market. Small businesses whose major data processing problem is essentially in one area, e.g. distribution services, may well be served by a small dedicated in-house mini, perhaps with a link to a central site for some needs.
The implementation of Electronic Funds Transfer Services and some of the Wired City approaches are best left to the banks, the cable companies, or the Bell. Computer service companies could, however, become ‘nformation suppliers’ to any Wired City approach.
(d) Increased Data Transmission Capability – The pending competition between AT&T and IBM will lead to rapidly improving capability in the data communications field. The cost structure of data transmission should also improve with the availability of satellite facilities. There appears to be an over capacity at the present time which should make pricing more attractive. This will tend to encourage the use of networks by the few very large companies.
(e) Declining Cycle Costs – This downward trend will continue for both CPU and storage costs. However, the cost of leasing unbundled software and the cost of operating staff will continue to rise, offsetting this cycle cost advantage.
(f) Minimal International Computer Services Business – I have often expressed concern about the pending growth of international barriers to data transmission. I do not believe that computer service firms in Canada will be threatened by remotely located computer cycle suppliers and do not feel they need artificial protection.
The world wide growth of nationalism is very clear. I believe therefore that ultimately Trans-border data flows will be allowed only in certain areas such as access to generalized information banks the use of international credit cards, and certain types of information for Multi National Corporations. Access to many other forms of data will be restricted despite the attractiveness of lower satellite transmission costs.
At Datacrown we have seeded the expansion into the United States from a Canadian centre, but are planning to locate centres in other countries as required. What we will do in the long run is export the know-how, not transmit the cycles.
In general, I project a very vigorous decade for the computer services industry.
• The need for improved productivity in this country is immense. This is the service we provide to other organizations.
• Qualified computer-professionals will be in short supply and organizations are going to have to use all the talent they have for application development rather than cycle delivery.
• The technology gives a great range of implementation possibilities to the computer services supplier. All we need to do is use our imagination.
The approach for computer service suppliers will be to provide specialized knowledge and facilities that are sufficiently advanced that it will not be worthwhile for the generalists to compete. This specialization may be in cycle selling, or application programming, or in certain instances, both.
The success will go to the firms who choose well what they want to do, and then do it well.