Great response for the last post on this topic, so let's keep the negativity going. Here are 2 more things that are killing BI, and what we can do to stop it.
4. The BI vendors are killing BI.
From the comments from my last post:
"Vendors don't 'get it,' and customers aren't accustomed to making it easier for vendors to define adequate BI solutions. Vendors all too often focus first on how to sell their products, and only secondarily on really learning how various business-types function. A successful BI implementation can and must meet BOTH sets of needs: the customers' and the vendor's. Win/Win isn't just a goal, it's essential to ensuring adoption and sustainability."The issue is that many BI vendors focus more on sales, than thought leadership and understanding the businesses their supporting. Thus, you get a bunch of features and functions stuffed into releases, but no real "core" value that allows the businesses to better understand their data.
Shame on the vendors for not providing the important features required. However, shame on the customers for seeking more bells-and-whistles, and not addressing the core needs of the business decisions makers. We need to get together on this.
5. Lack of leveraging unstructured data is killing BI.
We clearly understand how to deal with structured information, but what about unstructured data, which is most the data out there?
Truth-be-told we don't think about leveraging unstructured data, but we're doing that at our peril. If you think about, when going into a doctor's office, the critical information is what he or she writes down in your chart, and not which boxes he or she checks that are ultimately entered into the database.
Indeed, there is much to be culled from thousands of terabytes of unstructured data existing as documents, raw text, OCR text, Web pages, etc. While not create technology that allows this information to be considered within the context of BI systems? This information is as important, or more important, than traditional structured information. We need to make use of it.
Completely agree with your point about unstructured data. I've been waving that flag for ages. How can you base your most important decisions on only a fraction of the information? I feel like an evangelist shouting in the wilderness sometimes, though. At Pervasive, we've had a solution for scraping unstructured data into data warehouses and BI data stores for years, but it's a feature that isn't well known. A huge percentage of our own customer base never considers it an option to pull in that data, and doesn't even realize that they already have the software tools right there to do it. The small percentage of customers I've spoken to who are aware of the capability, use it daily, swear by it, and can't imagine how they got along without it.
Paige Roberts
Tech Content Developer
Pervasive Software
I agree that unstructured data has a place in BI. However, simply entering all data into a system will cause data overload, and hide crucial information in a sea of data. The real trick is to make the unstructured data more structured. For example, scan notes for key words, and have those words trigger certain tags, which could then be applied to that note. This would not only allow notes to be entered, but make them searchable and useful as well.
In my sometimes not so humble opinion, the phrase 'Unstructured' information is an oxymoron. It has a structure, it's just not one that is easily recognized by application developers.
The key to corralling digital assets - in the largest sense of that class - is to understand that all information comes from the same place and is already organized before it gets into database tables and digital assets.
If you doubt the veracity or provability of that statement, examine the evidence of failed projects, botched requirements gathering and most importantly the fact that I can conduct my business without - or sometimes in spite of - a computer application.