Search and if you are lucky you will find

This morning I was following the tweet stream from the ongoing Gartner Master Data Management (MDM) conference here in London, when another tweet caught my eyes:

This reminded me about that (error tolerant) search is The Overlooked MDM Feature.

Good search functionality is essential for making the most out of your well managed master data.

Search functionality may be implemented in these main scenarios:

Inside Search

You should be able to quickly find what is inside your master data hub.

The business benefits from having fast error tolerant search as a capacity inside your master data management solution are plenty, including:

  • Better data quality by upstream prevention against duplicate entries as explained in this post.
  • More efficiency by bringing down the time users spends on searching for information about entities in the master data hub.
  • Higher employee satisfaction by eliminating a lot of frustration else coming from not finding what you know must be inside the hub already.

MDM inside search capabilities applies to multiple domains: Party, product and location master data.

Search the outside

You should be able to quickly find what you need to bring inside your master data hub.

Data entry may improve a lot by having fast error tolerant search that explores the cloud for relevant data related to the entry being done. Doing that has two main purposes:

  • Data entry becomes more effective with less cumbersome investigation and fewer keystrokes.
  • Data quality is safeguarded by better real world alignment.

Preferably the inside and the outside search should be the same mash-up.

Searching the outside is applies especially to location and party master data.

Search from the outside

Website search applies especially to product master data and in some cases also to related location master data as described in the post Product Placement.

Your website users should be able to quickly find what you publish from your master data hub be that description of physical products, services or research documents as in the case of Gartner, which is an analyst firm.

As said in the tweet on the top of this post, (good) search makes the life of your coming and current customers much easier. Do I need to emphasize the importance of good customer experience?

Bookmark and Share

12 thoughts on “Search and if you are lucky you will find

    • Henrik Liliendahl Sørensen 9th February 2012 / 14:28

      Thanks Peter. Maybe Gartner will take own medicine one day 🙂

  1. Dylan @ Data Quality 9th February 2012 / 14:55

    I think search is one of the biggest causes of data quality in the average organisation.

    I remember profiling some equipment data and found that the users had entered 250+ variations of just one equipment type!

    The problem was the product type search had no wildcards and it was a really complex product name, as a result they couldn’t find the master version so just kept adding new equipment types.

    It always strikes me as odd that companies will invest heavily in downstream cleansing tools that do – matching, standardisation etc…

    Precisely the capabilities that should be deployed in search functions!

    • Henrik Liliendahl Sørensen 9th February 2012 / 15:13

      Right on the money, Dylan. Fuzzy matching techniques and other tricks from the matching trade may exactly be implemented in search capabilities.

    • Henrik Liliendahl Sørensen 9th February 2012 / 17:21

      Thanks for joining Scott. Well, there is this old saying: “The shoemaker’s children go barefoot”.

  2. Jean-Michel Franco 10th February 2012 / 17:23

    Fully agree on your post. In fact we have an experience with my company where a search engine has been embedded in the CRM system not only to provide a customer 360° view (including potentially what you call search from the outside e.g. to include last personal e-mail sent to this customer) but also to resolve data quality issues due to duplicate entries (eg. Customer appears with the name of the mother company or subsidiary…).

    • Henrik Liliendahl Sørensen 11th February 2012 / 09:32

      Jean-Michel, thanks for adding in.

  3. Prashanta C 11th February 2012 / 22:55

    Good post Henrik,

    One of the most important feature of MDM. I have worked at many customer and product MDM implementations and I can’t remember an instance where we didn’t spend considerable amount of time enhancing the search functionality.

    The different search combinations, wild card vs exact matching etc can be overwhelming. A careful consideration of these options will not only bring in correct data for the interested parties, but also allow MDM perform efficiently. For example, we wouldn’t want (At least not very often) users use criteria like A* which would bring in all customer with name starting A.


    • Henrik Liliendahl Sørensen 12th February 2012 / 10:26

      Thanks Prashant.

      I have for many years not found wildcards optimal for search for names and addresses. There are performance issues and they are very hard for users to actually use.

      Fuzzy search based on the very same technology we use in fuzzy matching is much better. For example if you search for “Liliendahl Limited” you will find “Liliendahl Ltd”, “Liliendal Limited”, “Liljendahl Limited” and heaps of other alternate ways of spelling.

  4. Matt Gagan 17th February 2012 / 04:15

    I would like to point out that TIBCO MDM embeds TIBCO Patterns (acquired 2010 as Netrics). The technology is used in TIBCO MDM in two ways: Matching engine (for data de-duplication); Fuzzy search engine exposed in UI and as a Web Service. The technology is entirely math based and requires NO rules in order to deliver incredible results in searches. (You will want to use rules to control exactly how data de-duplication takes place, of course.)

    Please try out the technology at

    • Henrik Liliendahl Sørensen 17th February 2012 / 11:04

      Matt, I remember the Netrics product before it was rebranded by TIBCO. I think it’s a good move to have the fuzzy search in there.

      With customizing rules in search I have been involved in setting up specific industry related synonyms. Probabilistic learning is another approach to that.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s