Michael Blechar

Those of you who may have been watching the comments I have been receiving regarding metadata management will notice that most of them have been about technology (metadata repositories) – and despite my efforts to make the discussion more about business issues such as value, justification and governance we always seem to get back to technology in terms of comments. So, let me take this opportunity to look at non-technical issues affecting the success of metadata management using two examples of human behavior (neither of which involve metadata) which impacted my views on technology and business in significant ways.

The first was back around 1972. I was working as a programmer for the IT arm (SIAC) of the NY Stock Exchange rewriting the error correction subsystem for stock trading. Written in Assembler language and patched a thousand times no one wanted to touch it, but the lot fell my way to do so. Without going into detail on the complexity of the applications, suffice it to say that testing the rewrite was perhaps the greatest challenge. And, there finally came the first of three major system integration tests involving the business users. During the test, there was a problem which at one point spewed “garbage” across the ticker. I immediately ran to the code and debugged what turned out to be a very minor programming problem and the rest of the test went swimmingly. But, I was later shocked to find that the business users had gone to management to tell them what a disaster the test was since to them “garbage on the ticker” was catastrophic.

In other words, I had failed to manage the business users expectations and see what I was doing in terms of their eyes – a mistake I would not make again. What are your business users expectations for, and perspectives of, the value of metadata management? Do they help you measure “success” in their eyes as opposed to your own?

The second involved a Fortune 500 pharmaceutical company for which I worked as the head of Information Management in the 1977 timeframe. There had been terrible quality problems in the picking of products to ship to customers based on what they ordered. Management had tried everything it could (threats, bonuses, firings, etc) to boost quality in picking products for shipment to no real avail. They finally hired one of the most renowned figures in the area of materials management to come in as a consultant to figure out how to fix the problem. He and the CEO spent a number of hours each week in the order fulfillment part of the shop floor and finally decided that the problem was that the product SKU numbers were too similar and the physical layout of the inventory with products having similar SKUs also created problems. They proceeded to fix these issues and initially quality went up significantly. But within a few months quality had dropped back to previous unsatisfactory levels.

Interestingly, when they went back to the floor to see why things had reverted to previous bad levels of quality there once again was a positive improvement despite making no changes. Finally, they came to see that it was the mere fact that the CEO and consultant were visiting the shop floor which reinforced to those picking the products that this was a really important part of the business to which they needed to pay attention. Afterward, the CEO would make periodic visits and, as a result, quality remained high.

In other words, change requires those involved believing that the change is important and needed. And, some “implementations” require ongoing re-selling and justification of value to be successful. Effective metadata management needs to have a multi-year implementation strategy followed by ongoing support and championing even after it becomes the “normal mode of operation” for the organization.

And, so, I would urge readers of this blog involved with metadata management to reflect on these two examples. What is the perspective of the business users on the subject of metadata management? Most will view it as an IT problem as long as they get the data need to do their job. Or, to put it in perspective how many of us when we buy a product care what the production process was to create it or how it got distributed to the store in which we bought it? There are also people issues. What if next time you went to buy a car the salesperson told you that it is inefficient to have every person own their own car and that the one you are about to buy needs to be shared with 5 other families? And even of you agreed to share, what would the rules of governance be across all potential users of the car?

Successful metadata management requires effective use of technology, but in many ways tools are the least important aspect of the problem. People, process, culture and politics – and managing everyone’s expectations of what “success is” – tend to be the determinant factor of the success or failure of metadata management efforts than technological issues.

Gartner clients wanting to read more about how to deal with these non-technology issues regarding metadata management are advised to use the following link on how to manage master data (since metadata management faces the same issues).

Mastering Master Data Management

Submit a Comment »

Category: Uncategorized     Tags:

Share