≡ Menu

Pretty pictures are not enough


I saw this article today on Real-Time Analytics: The Importance of Strategic Alignment and Ease of Use, and it made me think (again) about the danger of pretty pictures. While the article makes some good points about analytics in manufacturing, it seems to me that the author is sucked in to thinking that pretty pictures are enough. Despite talking about the complexity inherent in large manufacturing operations where one might have thousands of products and customers and complex interactions between them, he still assumes that a person is going to look at the data. This seems flawed to me. There is too much data and it changes too fast for someone to always look at it. As more and more manufacturing is outsourced or done in another part of the world or run 24×7, there may not even be someone awake to look at the data.
What is needed is decision automation – the ability to make systems smart enough to take useful, appropriate, analytically sound decisions as and when they need to. This means that those who understand the business need to write rules about when and how to respond (including rules about when to get a person to look at a picture) and build models that optimize or predict aspects of the operation. Modern organizations will not become analytically driven, and certainly won’t become real-time without this kind of automation.
The article does make a good point about the need for a common metric around which an organization can align. Richard’s metric is manufacturing-centric, however, and I think I prefer the more general “always measure shareholder value” approach of my friends over at Provisdom.
So do think about data quality and integration, do come to an agreement about common metrics. Don’t think that pretty pictures are all you need.


Comments on this entry are closed.

  • Dave Dixon November 8, 2007, 11:27 am

    Hi James. Thanks for link to Provisdom’s blog. We posted some further discussion of shareholder value in the context of the DMReview article. Here’s the link: http://blog.provisdom.com/?p=26