Report Writer Cost Benefit Analysis


This document is intended as a cost benefits analysis of report writers as pertaining to a company which is considering internet enablement. Two report writers are judged as being the leading contenders, Actuate and Seagate Info. Actuate is an object-based report writer and Seagate Info is a VB-like product.

As a side-issue, it is hoped that this document will help to make companies more aware of the enormous revenue potential extensive investment in internet technology and service provision development.


How can a Report Writer be Used ?

A report writer can be used for both internally (behind the firewall) and externally, specifically on the internet. The following assumptions are made.

  1. The company is planning on becoming a fully internet-enabled company.
  2. It is the intent of management to make as much use of the internet as possible in order to take advantage of potential new markets and thus ultimately increase revenues substantially.
  3. The company may in future effectively be an internet service providor of information, therebye automating provision of information such as is presently provided by customer service, to whatever degree possible.
  4. The company will provide internet based transactions, ECommerce or otherwise. This could potentially automate much of operational requirements, once again to whatever degree possible or desired.

The Long Term Objective

How Far Ahead is your Company Looking ?

Requirements for a report writer, as presented here, are projected many years into the future. Due to the cost of development it is expected that software should have a useful life of atleast over ten years. Even 20 years is attainable. An advantage of using leading-edge development tools makes it more likely that a company will be able to find programmers to maintain this software ten years from now.

Use of bleeding-edge software is dangerous. In general, leading-edge can sometimes equate to the latest versions of SDKs or database engines. Companies like Oracle would have gone bankrupt many years ago if they had consistently released unusable software. Specialised end-user software products with low market share tend to have a lot more problems. They are also subject to less risk than products like Oracle since there are few specialised products available. Specialised low market share software products are probably more bleeding-edge than any other types of software products. Software products like Oracle and Java have very high market share. Also the larger the size of companies using software is usually a good indicator since there are a lot more people and money available to make more carefully considered decisions. Microsoft products are however a contradiction. However, unstable Microsoft products tend to be aimed at small-scale, non-mission critical tasks. One might ask how Microsoft NT survives on US-Navy Aegis Class Cruisers ?

The objective with report writer selection is a long-term rather than a short-term perspective. The more carefully software is designed and implemented, the better it is constructed and thus the longer it will remain useful. Modern software development tools tend to be much more effective at cleaner and more reusable software components as compared to SDKs such as Clipper and Dbase.

The Limitations of Shrink-Wrapped Packaged Software

This document is not intended as a comparison between in-house software construction and purchase of shrink-wrapped software packages. The only point that should be put forward is that some shrink-wrapped packages probably have self-contained report writers. However, self-contained implies that reports can only be constructed within the bounds of the shrink-wrapped package. Note that it would be difficult and very time consuming to construct reports from multiple shrink-wrapped software packages.

The Costs of In-House Development

In-house development is costly. However, with modern software development tools this development time is getting shorter and more easily achievable.

What Type of Software Should a Company Use ?

It is a suggestion that the most restrictive software, methodologically, should be used as much as possible. This means that code can only be written in a specific manner and strict rules of design and implementation have to be adhered to. One good example of this is Java. Java is restrictive intentionally. It is a purist type of programming language. Java tends to force programmers to be more consistent. Visual Basic on the other hand is over-flexible. The result is that Visual Basic applications are frequently rewritten since different coders can write code any way they choose. Complaints from programmers about the non-maintainability of Visual Basic code are frequent.

One gets more longevity out of software when more care is taken in its construction. Thus the better the quality of the coding of that software the more easily that software can be passed between different programmers. Therefore the obvious conslusion is that the more precise the programming language used the better the software produced.

Object-like development tools allow for extensive reuse of previously written components. Therefore tools like Java and Actuate could be used to develop many of the reporting requirements across the whole spectrum of a company's operations. Money can be saved in the future !

This leads us to the choice of report writers.


Which Report Writer to Select based on Cost and Capability ?

Actuate is an object-based report writer. Java is an object-based programming language. Therefore Actuate will provide the best long-term solution. Also, Actuate is the cheapest option which has the scalability for internet usage.

$ Cost
Report Writer SDK per Machine Named User Power Unit (www) Special Initial Cost Internet Scalability Scheduling Technology Security
Actuate 2,000 0 35,000 Cost is 35,000 per CPU. 70,000 Highly scalable. Excellent. Object-based. Extensive reusability. Purpose-built for Internet. Oracle usage improving. Excellent.
Seagate Info 400 (1 free) 300 (49 free) 70,000 Freebies misleading. Cost is for 25 concurrent users. 141,000 Good. Heavily tied-in with Microsoft. Good. Crystal Report based. Very dubious. Poor usage of Oracle. SLOW !!! Good.
Oracle Reports 2,000 See Special 5 Developer & App. Server reqd. & no MTS Expensive Not investigated Not investigated. Antiquated and primitive because query-based. Part of Developer Suite. Many complaints. Security is either Oracle Application Server (costs extra) or database-based (not appropriate for the internet).
Crystal Web Server 0 0 0 Already purchased. 0 Very Limited. Suitable for internal non-internet use only. None. Incapable of concurrent processing. This means that if more than one person runs a report at once there will be a significant wait for a report. No security whatsoever. Totally inappropriate for internet usage.

Expected Future Costs

At this point Actuate and Seagate Info appear to be the best selections for a report writer. Actuate appears to be the better product both technically and cost-wise.

Now we can attempt to extrapolate into the future in order to assess numbers of potential users, both customers and internal to a company considering web-enablement. Ultimately, if we look far enough into the future, the cost of internal usage may become negligible in relation to that of internet usage. This is assuming that software is contructed in the most reusable fashion possible such that development work becomes less and less time consuming.

Expected Internet Usage Increases

It is difficult to assess potential internet usage into the future. The internet is new so there is not much in the way of publicly available information. The only information available is on the internet itself and is sketchy at best. There are various companies which sell comprehensive projection information of this type. The quality of this information is unknown. The future population of internet usage is important to to any company with respect to the choice of report writers. Particularly with concern to the pricing structure of Seagate Info. Seagate Info is priced by number of users accessing the reports server, Actuate does not care about number of users. Seagate Info can potentially get astronomically more expensive than Actuate based on internet usage.

The inclusion of Internet usage statistics may seem unneccessary at this point and possibly irrelevant to the choice of report writers. The intention is to attempt to demonstrate a number of points.

  1. Internet usage is growing rapidly, both in the US and worldwide
  2. Even though Internet usage increase in the US is slowing the non-US area is still growing at an extremely rapid rate
  3. US ECommerce is expected to keep growing past the point of US Internet usage saturation
  4. Note that all the values shown here are statistical estimates. In the past these estimates have generally been low. It is probably something to do with human nature that the pessimists are most often heard. Probably something to do with the media.
  5. The internet is growing rapidly. The potential for the number of users and the quantity of data was probably unimaginable to even the most visionary minds in the past. That potential is probably still underestimated. Therefore the potential for doing-business on the internet for a company considering web-enablement is substantial. Investing in the most flexible and reliable report writer is expedient. That report writer is probably Actuate.
  6. The potential number of internet users and the size of that market could be tapped. How should this affect the types of software that a company invests in ? Are shrink-wrapped packages able to cope with the sheer volume of data on the internet that ECommerce companies could be dealing with in the very near future, next year for instance. If a company were to become fully internet enabled within two years what would be the scalability of previously purchased software packages ? Would it be better for a company to build software in-house ? Would in-house implementation allow for more flexibility in future growth ? Is your company planning to replace shrink-wrapped packages in the near future when the capabilities of those shrink-wrapped packages are exceeded ? Most GUI written shrink-wrapped packages were developed years ago and may very well have problems with ECommerce data volumes. If a company's ECommerce business is to expand consistently with the rate of Internet usage expansion, surely investment into in-house development could take advantage of potential internet expansion much more effectively.

Check out these internet population links. These are some example sites of how and where internet usage statistics were gathered. There are many more.

The following two diagrams are two histograms. These histograms represent current and potential internet usage numbers based on past growth of internet usage. The years shown are 1998 and 2002. Growth does not appear to be exponential in these diagrams but does appear to at least double in every country shown.

Following are two sets of internet usage statistics. Note that all grayed, italic figures are guestimates based on statistics that have been gathered online. Online future projection figures tend to vary. In addition the general assumption is that the US market will be saturated by the end of the year 2002. It has also been assumed that the non-US market will be saturated by the end of the year 2005. The ratio of US to non-US internet usage has been estimated based on the worldwide installed base of PC's. The worldwide installed base of PC's is supposedly 28% in the US. This figure could be incorrect since it probably does not show a true reflection of commercial to non-commercial PC use in the US in relation to the rest of the world. The fact that US GDP is based on 90% internal trade and 10% foreign trade (I think) could possibly alter this 28% figure drastically. Obviously at 10% foreign trade is not all that important to the US economy.

US Internet Users in Millions

 

Year US Internet Users in Millions Increase % Increase Change % Notes
1995

22

 

 

 

1996

38

73

 

 

1997

58

53

-20

 

1998

88

52

-1

 

1999

110

25

-27

 

2000

133

21

-4

 

2001

144

13

-8

US market may become saturated in 2002.

2002

163

7

-6

 

2003

175

3

-4

 

2004

180

1

-2

 

2005

184

0

-1

 



The US Internet usage market is assumed to have become saturated by the end of the year 2002.


Worldwide Internet Users in Millions


Year

US Internet Users in Millions

Non US Internet Users in Millions

Worldwide Internet Users in Millions

Increase %

Increase Change %

Notes

1995

22

10

32

 

 

 

1996

38

21

59

84

 

 

1997

58

42

100

69

-15

 

1998

88

86

174

74

+5

 

1999

110

152

262

51

-23

 

2000

133

194

327

25

-26

 

2001

144

298

442

35

+10

.

2002

163

392

555

35

0

 

2003

175

446

621

20

-15

 

2004

180

464

644

7

-13

 

2005

184

473

657

3

-4

Non US market saturated when US is 28% of worldwide installed PC base.

The non-US Internet usage market has in the past trailed behind the US Internet usage market. Thus since the PC installed base ratio of US to non-US markets are assumed to eventually be 28% the assumption is that this installed base ratio will become a reality by the end of the year 2005.

The pictures shown below (US ECOMMERCE 1998 - 2003) could be indicative of the possibility that the US Internet usage market will by no means be saturated by the end of the year 2002, in fact it could still be climbing rapidly. This diagram is actually completely contradictory to the assumption that US Internet usage will become saturated by the end of the year 2002. In fact this diagram assumes that US ECommerce is still expanding in 2003, therefore Internet usage could still be expanding at a similar rate.

US ECommerce 1998-2003 - Business to Business US ECommerce 1998-2003 - Business to Consumer

Expected Cost Analysis

This section is based on possible number of internet users at a fictitious cpmpany within the next 5 years and 10 years. Probably 10 years or greater would be a more realistic target for software survivability. However, the expected number of potential web site internet-based report writer users is enormous in scale compared any current usage. This is to be expected with the internet. Note that these estimates of numbers of expected users could be be optimistic. The better a company is prepared for internet usage then the more marketable that company will.

Let us take the example of a company which has 125 customers for instance. The number of customers have increased over the last 5 years @ 20% per year on average. It could also be assumed that the number of customer's clients also increase by 20% per year. Extrapolated values can be calculated using the compound interest formula shown below. The calculation is performed on an annual and not a daily rate simple interest form. All numbers have been rounded in the interests of simplicity.

Compound Interest Formula : total = principal (1 + rate)number of years

Customers

Customer's Clients

Let us say that on average, over the 10 year period, that 80% of customer's clients use and will use the internet. Thus the following adjustments can be made.

How do we assess the percentage of customer's clients who do or will in future trade on the web ? This is presently not an answerable question. Consumer purchasing rates may be subject to economic fluctuations. Economically lean periods could affect consumer purchasing power. Internet usage numbers are probably high now. These percentages will increase in the future as the internet gains better general acceptance.

How Much Time do Customer's Clients Spend Online ?

 

Customers

Customer Internet Users

Total Customer's Clients

Customer's Client Internet Users

Total Internet Users

Actuate Costs (35K/CPU)

Seagate Info Costs (25 concurrent users)

Today

125

50

70,000

1

51

$140,000

$210,000

5 years

300

120

500,000

5

125

$140,000

$350,000

10 years

800

310

3,200,000

32

342

$140,000

$980,000

The $140,000 cost for Actuate assumes an NT-Server with 4 CPUs (current maximum for NT). Seagate Info does not care how many CPUs there are. If volume gets too large for a single machine Actuate is scalable to Unix and Seagate Info is scalable into NT clusters. The price for Seagate Info scaled upto clustered NT machines would not increase. If Actuate was scaled upto Unix then the number of processors on the Unix machine would determine the cost of Actuate. Actuate would have to be placed on a Unix machine with 6 processers today, 10 processors in 5 years and 30 processors in 10 years to be the equivalent in cost to that of Seagate Info. Unix machines use multiple CPUs in powers of 2, ie. 1, 2, 4, 8, 16, 64. A multiple-processor Unix machine will probably be much, much faster than clustered NT machines and also a lot more expensive. The power of an NT-Server cluster is in some ways negligible compared to that of a multiple processor Unix box. You get what you pay for. At this point in time it is probable that even in 10 years the Actuate installation can be servicable on a dedicated multiple CPU NT-Server.

Why is Actuate more Cost Effective than Seagate Info

The pricing structure of Actuate is completely different to that of Seagate Info. Actuate is based on power-units and Seagate Info is based on a number of concurrent users. When those concurrent users increase to large numbers the cost of Seagate Info in relation to that of Actuate is disproportionately larger.

Actuate is initially two-thirds the price of Seagate Info and in the long-term potentially as low as 15% of the price of Seagate Info depending on future hardware requirements. Hardware requirements are dependant on internet usage numbers. NT machines are probably a lot cheaper than Unix machines but Seagate Info may need a lot of NT-Servers. In general, Seagate Info has an inappropriate pricing structure in order to remain competetive with Actuate.


Technical Comparison of Actuate vs Seagate Info

Actuate Seagate Info
Initial learning curve Initally Steep. Approximately 3-4 months for an average programmer to gain a complete understanding without the help of training courses. Simple.
Maintainability Exceptional because object-based. Becomes rapidly better with time. Deteriorates to complete non-maintainability over time to the point of requiring rewriting.
Reuse of code Excellent. A substantial cost-saver after period of time. Completely non-existent. Everything new has to be recreated. Development becomes much more difficult over time.
Database oriented Flexible and getting better. Highly restrictive. Very tied-in with Microsoft. Very poor usage of underlying database processing capabilities. Potentially very SLOW !!!
Purpose-built Purpose-built for internet reporting. Scalable to the nth factor but not by clustering, YET. So what ! Clustering capability appears to be over-emphasised. Possible that it may cause serious speed problems and be completely useless in extremely intensive environments.
Complexity of coding Initially requires more preparation and careful thought. The more code that is written the more code that is available for reuse. Pre-built components may become available in the near future. Generally becomes more capable and more resusable over time. Initially simple extending to potentially impossible to maintain for anything other than the most simple reports. Becomes convoluted, very difficult and expensive to maintain over time. Not capable of reuse at all.
Complexity of implementation Initially complex since uses a new methodology. Initially simplistic requiring frequent rewriting when too much complexity introduced.
Ability to Implement Complex Reporting Excellent. Extremely Poor. Originally intended for simplistic group-type relational database reporting.
Skills @ ADS 2 out of 5 developers have object experience. None with Actuate. 3 out of 5 developers, 2 are very experienced.

Scalability, Clustering and Failover

Technically Actuate is the better product in terms of the future use of a report writers. It is ultimately less expensive. Seagate Info seems to place a great deal of emphasis on their clustering scalability on the Seagate web site. This is a little suspicious in that Seagate Info clustering capabilities appear to be over-sold. It appears that Seagate Info processing can be scaled extensively. Probably too extensively ! Why does Seagate Info require such extensive clustering capability ? Actuate runs a report server process on a single machine and supposedly has some very large installations. Traditionally Seagate Info extends from Crystal Reports. Crystal Reports was initially a PC-based single-user database report writer. Actuate appears to originate from the Unix-world. The Unix-world has always been the far more robust and scalable in relation to NT. Unix is also much more reliable than NT.

OLAP and Data Warehousing

Actuate does not have this capability at present. Seagate Info uses Seagate Analysis which it is assumed is used to reanalyse reports produced. Analysis of previously produced reports is known as OLAP (Online Analytical Processing) of cubes of data. A cube of data retains references between data items within a report instance, ie. after the report has been executed.

OLAP is an extension of data warehousing to a certain extent. Data warehousing is in some respects the result of the need to denormalise the data in a relational database. Data Warehousing effectively degrades the structure and efficiency of a relational database. OLAP and Data Warehousing was created because a relational database falls short in the area of data analysis to a certain extent in terms of speed requirements. Generally, a relational database structure is incapable of intense cross-entity structural analysis, as required by OLAP and Data Warehousing. Use of Seagate Analysis as an OLAP tool is therefore suspect.

The Lack of Clustering and OLAP in Actuate

Simply put, if there is a need for either of these capabilities it is quite possible that Actuate will cater for it in the future.