by [TC]²
A monthly column of technology rambling, rumination and reality
By: Jud Early, Corporate Vice President, Research, [TC]²
November 2005
Hello, faithful readers,
About five or six years ago, we were contacted by a well-known brand owner with global operations asking for help with supply chain difficulties. A colleague and I traveled to the offices of the requesting company and met with several “C level” and senior managers in an effort to determine the cause of their problems. With manufacturing operations both here in the US and abroad, they had a collection of owned as well as controlled factories that were producing product. The difficulty was that when orders were filled, in many cases the order was shipped short, due to unavailability, and later, when orders were less demanding, there was abundant product, with distribution centers full of product that had arrived too late to be shipped. Does this sound like a familiar story? The company will remain unnamed, but the story is quite real. As we interviewed the senior managers in the company, the picture became clear.
The company has a number of brands under one corporate brand, and product families of numerous styles within each brand. Fourteen different people had responsibility for forecasting the needs of a single brand, and each produced a forecast that drove a number of production schedules which dictated sourcing. In addition to each person forecasting, each one was at liberty to pad the production needs based on anticipated orders, and would do so in varying amounts depending upon whether the product was to be produced in domestic facilities or sourced from abroad.
Incredibly, each of the forecasters operated independently, and the production schedulers were driven by these forecasts. The production schedulers also had latitude in where the product was produced, seeking first the lowest cost, then acceptable delivery times and of course, fabric availability. The result was chaos, a function of rapid growth and success in the market place. As production orders were placed, in quantities often not justified by real demand, the bullwhip effect adding to the harmonic in production and delivery, things were out of control and a major overhaul was needed. Fortunately, the company had recognized the need for help, and had, in fact, begun an internal project to fix the problem. Although we were prepared to submit a proposal for developing a solution, the internal project was in progress, and it appeared that what the prospect wanted was a reality check by a company outside who could assess the nature of the problem and validate the internally devised solution. The good news is that the company survived its prosperity and is today a strong brand.
I was reminded of this company and its travails as I attended the Apparel Magazine Technology Conference last week. Many vendors had tabletop displays and presented their products in tightly scheduled seminars. ERP, PLM, PDM and FLM were all presented as solutions to bringing a product from Concept to market or beyond. BI, or business intelligence is the area in which I was reminded of the company whose story is told above. BI has been in the enterprise applications marketplace for several years. Some that we have looked at offer a real-time dashboard that can be customized to display all critical information to the user that he or she needs to perform their job responsibilities. One demonstration that I remember framed the BI application as just for the high level executive that is always busy but needs a snapshot of how subordinates are performing. Others position the application as suited for middle managers whose operations responsibility can be enhanced by having up-to-date information. I believe that BI can be the key to operational excellence if properly used at all levels.
Before I discuss the business intelligence application that I found interesting, I must state that until last week, I had no knowledge of the vendor, and I am in no way endorsing or promoting the product. No compensation or gratuity of any kind have been offered, solicited, nor would be accepted. The purpose of this mention is to bring to light a simple and what appears to be effective way to provide an easy to use management tool to people at all levels in an enterprise.
Implementing a business intelligence application can mean great cost to a company, with integration costs often exceeding the cost of the application’s license. The ability to connect to legacy systems, spreadsheets, databases, financial, and other company information, which often exist as islands of data, has required programming to interface with API’s (application programming interfaces), which may, or may not be present in the existing available applications. One provider prided itself in utilizing “screen-scrapers” that intercepted information that was to be displayed on a monitor, parsing out the data that was to appear in data fields when the data could not be obtained otherwise, and sending it to a huge data warehouse, where it was stored for retrieval when a request was made. Spreadsheets, which often are not designed to link to a database, or which have been developed by users, may not be of consistent design. This is not an exhaustive list of difficulties facing the integrator, but is representative of challenges faced in bringing together the disparate bits and pieces of information that can help the user make informed decisions.
Data warehouses can also be an item of significant cost, even with disk storage costs plummeting, the cost of maintaining the huge amounts of data, arranged in neat logical tables, and updated in real-time, present a high cost of ownership for one who wishes to be up-to-the-minute in data accuracy.
One of the beauties of the Blue Fox Porini Business Intelligence application is its simplicity. It derives the simplicity through key differences in approach. First, it is not, and is not intended to be a real-time application. It may be updated daily, or twice daily, or hourly to suit the needs of the company, but more frequent updates are often not necessary. The second area that is of interest is a data layer that is structured for fast queries and which is populated from uploads from the legacy, and other resident systems within the company. Spreadsheets, which can be the bane of integrators, may need to be restructured to provide consistent cell locations where changing data will be held, and where the data layer will look for its frequent uploads. In similar fashion, the other applications, from which the data layer will pull its information, will provide the information from mapped fields or database tables, allowing the information to be gathered for the many user screens that will be possible. The third element is the user interface. Offering a wide assortment of gauges, dials, charts, and numerical and text fields, the “dashboard” that is created is tailored to the needs of the corporate user, and may be different for each person. A click and drag query generator allows the configuration of what is to be displayed in the various display elements, and creates the Boolean logic to perform calculations on the fly from data that is in the data layer. The speed is quite fast, with each result displayed in seconds.
You might wonder about the wisdom of providing any and all company information to all users. Using a permission based access allows top level executives to view all facets of the enterprise, while middle and lower level managers and supervisors may see only information that is critical to their needs, and cannot access inappropriate data, such as payroll, salaries, sales, or other areas deemed off-limits by higher level executives.
As the user needs new information displayed, it is a simple matter to create the query, establish the proper way in which to display the result, and have a new calculated field or chart populated with the resultant data in a short time.
Example of a user screen that can be configured differently for each user.
Layer where data is stored for queries – calculated result sent to dashboard for display in graphic, numeric or text fields
By now the benefits of such a system should be obvious. Analysis of emerging trends, response in a short time, exceptions highlighted, and action plans to exploit the availability of previously non-available information provided by structured and ad-hoc queries. Training time is short, due to interface conventions used, and implementation can be two weeks, not months.
For those seeking an alternative to the data warehouse, this application should be on your short list.
In reflecting about how such an application might have avoided the supply chain conundrum of the company mentioned in the beginning of this story, I’m sure the visibility of all facets of the business would have helped. Pulling together all the islands of information would have been a good start to managing the problem. BI is not an ERP system, nor is it a supply chain or sourcing product, but is a tool to provide better, more visible information for decision making.
Graphics courtesy of Blue Fox Porini.
Gratitude
Of the 193 countries in the world today, we have readers in 138. Of the 193 countries in the world, eight celebrate thanksgiving or a harvest festival. We here in the United States are celebrating Thanksgiving Day on Thursday of this week. I would like to wish a happy thanksgiving to those who will celebrate this week, and to those who have or who will, in coming weeks or months, celebrate in your country, my best hope for a happy holiday and prosperity for which you can be grateful.
Until next month, be safe in life, and in computing.
Jud