In the February issue of DM Review, many of us wrote about the history of business intelligence (BI), its roots and how far it had come from its early days. This month, I would like to focus on BI's future a future in which the principles of mass customization will play an important role.
To understand how BI will adopt the new paradigm of mass customization, I would like to compare BI's progress to another industry that has embraced mass customization the printing industry. Figure 1 illustrates the three phases that the printing industry went through in its progression to mass customization and the relative costs and market size associated with each phase.
Phase 1: One-of-a- Kind Artisans
The early history of printed materials began as we emerged from the dark ages. Books were created by skilled artisans who labored over parchment, scribing each word by hand. The books were rare, prone to errors and very costly. Only a handful of people or organizations (such as churches) could afford such works. Generally, each book was commissioned for a specific customer.
Figure 1: Printing Industry's Progression to Mass Customization
Early data warehouses were similar to these early books. In the industry's early days from 1985 to the early nineties BI was not mainstream, to say the least. There were only a handful of artisans (consultants) who even understood the concept, much less how to build one of these systems. They developed these early BI environments by hand, much like their earlier printing counterparts.
Like the first books, the costs for these technological wonders were quite high. Early warehouses were generally in the tens of millions of dollars in terms of their overall implementation costs, and very few companies had one.
Although cost was quite a hindrance, there were other problems with this approach:
- Correcting or changing the existing design was almost impossible.
- Losing a critical resource such as the ETL programmer or DBA became an almost insurmountable difficulty.
- Reaching and fulfilling more than a few users was very difficult, if not impossible.
Phase 2: Mass Production Craftsmen
The next big development in the printing area was the printing press. Skilled craftsmen could mass produce books on these early presses. The ability to set the type and print multiple copies was revolutionary. Not only did the cost of books and other printed materials come down, but their increased availability also opened up a much larger market. Changes could be made with relative ease; and the loss of a critical resource, such as scribe, was not as devastating. Mass production became the paradigm of efficiency and inexpensive manufacture. Henry Ford summarized mass production perfectly in his now famous expression, "You can have any color Ford that you want as long as it is black."
However, to remain a viable paradigm, mass production requires the following conditions to exist:
- To realize the lowest possible costs, the production process must be as routine or automated as possible. Reduced costs also guarantee the largest market.
- Mass production requires a large, homogeneous market.
- Efficiency of mass production requires stability in the processes used, market stability and stability in the product development and life cycles.
The doctrine of mass production could be summed up as efficiency through stability and control.1 Mass production depends on stable, steadily growing demand to fuel its fire.
In the BI arena, a similar situation occurred. The late 1980s and early 1990s marked the advent of the "data mart in a box." The idea was to create a canned data model, canned cube or star schema and a set of canned reports. You could have any flavor of BI as long as it fit into these canned, mass-produced reports. Changing anything the model, the schema or the reports was not recommended and may not have been possible.
Mass production was a short-lived venture for BI. The good news was that the costs of a BI environment were reduced just as they were for printed books. The cost of a data mart in a box (just add data...) implementation ranged around $1 million and up.
The problem with these mass-produced marts was that they simply did not fit well into most industries. They provided some generic analytical capabilities, but every business has its own quirks and oddities surrounding its BI. To be prohibited from incorporating these into the applications made them only moderately useful.
Finally, these mass-produced BI applications were still expensive compared to what they could be. The opportunity still existed to improve on fit and costs.
Phase 3: Mass Customization Practitioners
The recent phase in printing is the ability to individually customize a book. This was recently brought to my attention when my daughter received a book for her birthday from her grandmother. The book looked like an ordinary book until we began reading it. The very first paragraph started out, "Jessica Imhoff lives in Colorado with her three cats, Sailor, Abby and Alice. One day her grandmother traveled from South Carolina to visit her..." and off they all went on their adventure.
Obviously this was not a mass-produced book. The manufacturing process had been changed to establish flexible setup points in the creation of the product to permit inexpensive and effective customization. Words and phrases could be inserted into the body of the book, giving it a customized, personal appearance for a minimal cost. Today we see examples of this mass customization in many of the items we use every day catalogs specific to individuals and their needs, national newspapers that contain regional news and local ads, Web sites that make recommendations specifically for you, e-books where you choose the ending, and so on.
What about mass customization for BI applications? Well, we are just beginning to see some of these mass-customizable BI applications in the marketplace. Like their printing counterparts, the principle behind these mass-customizable BI applications is to meet a high percentage of every company's generic needs while allowing the customization of the process for the remaining percentage of specific functionality. For example, the generic application may meet 75 percent of your company's needs and be customizable for the remaining 25 percent.
The instability of our market, the constantly changing needs and wants of our users, and the technological advancements for BI make mass customization an ideal paradigm for our industry. Now, we have the ability to create a better fit for these customizable BI apps into each company's individual requirements.
Similar to the mass-production phase, the ability to create mass-customizable components has again reduced the costs associated with the ownership of these technological marvels. In the case of BI applications, the implementation costs have dropped to hundreds of thousands of dollars and up another order of magnitude decrease in total cost of ownership. This has also opened a huge market for BI.
Mass customization, like any new paradigm, requires different thinking. This new paradigm is one of creating variety and customization through flexibility and quick responsiveness. In his book, Mass Customization The New Frontier in Business Competition, B. Joseph Pine states, "Creating high levels of variety in productions cannot be accomplished through specialized mass-production techniques. Creating variety requires flexibility in the manufacturing process."
Vendors of BI products are just beginning to adopt these mass- customization techniques. They are building applications such as key performance indicators, campaign analysis applications, product and customer profitability analyses that contain significant functionality functionality that you will find in almost 90 percent of businesses today. They are also building flexibility and variety into the implementation process to allow significant customization. A company may be able to purchase up to 80 percent of its BI needs through these new BI applications and then tailor them easily to attain the last 20 percent of specialized analyses.
Likewise, ETL vendors are beginning to create mass-customizable extractions against ERP installations such as SAP, PeopleSoft and Great Plains, again giving the users a high percentage of the programming and meta data needed to create their data warehouses and marts. However, they have inserted flexible process points within their products or have modularized their extraction processes so that you can modify the ETL process to fit all of your specific needs with ease and without losing the value of the ETL meta data.
Mass customization is coming to our world of BI. I see it as a very innovative, cost-effective and efficient means of achieving the majority of the functionality of most BI implementations.
1. B. Joseph Pine. Mass Customization The New Frontier in Business Competition. Harvard Business School Press, 1993.
The author wishes to thank Randall Porter and Brian Burnett of Appsmart for their contributions to this article.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access