Oracle Endeca Business Intelligence Rules

August 21, 2014

Rules are good. The problem is getting people to do what the rule maker wants. Oracle wants Endeca to be a business intelligence system at the same time Oracle wants Endeca to be an ecommerce system. You can find the five rules in the white paper “The Five Rules of the Road for Enterprise Data Discovery.”

What are these rules?

I don’t want to spoil your fun. I want to encourage you to dig into Endeca’s rules and to work through the white paper to see if you are doing enterprise data discovery the Oracle way. What is “enterprise data discovery”? Beats me. I think it is 1998 style search based on Endeca’s 1998 technology disclosed in those early Endeca patents.

First, you want to get results without risk. That sounds great. How does one discover information when one does not know exactly what information will be presented? If that information is out of date or statistically flawed, how does Endeca ameliorate risk? Big job.

Second, Endeca wants you to blend data so you get deeper insights. What if the data are not normalized, consistent, or accurate? Those insights may not be deeper; they may be misleading.

Third, Endeca wants everything integrated. How does one figure out what is important in a syst3m that gives the user a search box, links to follow, and analytics? Is this discovery or just plain old 1998 style Endeca search? Where’s the discovery thing? Blind clicking?

Fourth, Endeca wants you to “have a dialog with your data”. I find this interesting but fuzzy. Does Endeca now support voice input to its ageing technology?

Finally, Endeca wants those data indexed and updated. The goal is “keep on discovering.” I wonder what the latency in Endeca’s system is for most users? I suppose the cure for latency and Endeca’s indexing method can be resolved with Oracle servers. How much does the properly configured fully resourced Endeca system cost? My hunch. More than a couple of Pebble Beach show winners.

The white paper is interesting because it contains an example of the Endeca interface and the most amazing leap from five rules to customer support. Oracle also owns RightNow and InQuira. Where do these systems fit into the five rules?

Confused? I am.

Stephen E Arnold, August 21, 2014


3 Responses to “Oracle Endeca Business Intelligence Rules”

  1. Richard Tomlinson on August 25th, 2014 10:16 pm

    I think you’ve misinterpreted all the rules. Let me help….

    1. Is govern self-service for results without risk. This is the opposite of let your business users download Qlik/Tableau outside of IT’s control. EID is IT managed/governed but has self-service capabilities
    2. Blend = blend structured (rows and columns) with unstructured data (text) for deeper insights. Most data discovery tools can only handle data of the structured variety
    3. To explore structured and unstructured data effectively you need search + navigation + analytics. All 3. Like 3 legs of a stool – without anyone of these 3 components the stool falls over.
    4. Dialog with the data means you don’t need to understand your precise question upfront. As you interact with the data you will change your direction based on what the data tells you. It’s a back and forth vs. a one time query then results paradigm
    5. Keep on discovering means you can kick off enrichments AFTER the data is loaded into EID to constantly improve your data model. EID3.1 allows end-users to extract terms, bucket attributes based on white lists and much more.

    Hopefully that helps explain it a little better. The white paper spells it out pretty clearly.

  2. Anil Dhadwal on August 29th, 2014 1:20 am

    Nice articulation and clarification Richard. The 3 legs of the stool needs to be highlighted and awareness established for market to consume the real benefits..
    Thanks..Anil Dhadwal

  3. sell gold coins on September 20th, 2014 10:24 pm

    Yes! Finally someone writes about how to get an ira.

    Feel free to visit my web-site – sell gold coins

  • Archives

  • Recent Posts

  • Meta