Will Hybrid Search get you better mileage?

The recent news of the acquisition of Endeca by Oracle has triggered a number of research notes by analysts. In particular Sue Feldman of IDC talked of the rise of a Hybrid Search architecture. What is it? Is it good for you? Should you have one? And where does MuseGlobal stand?

Sue defined Hybrid Search: “Search vendors perceived this logical progression in information access a number of years ago, and several were at the forefront of creating new, hybrid architectures to enable access to both structured and unstructured information from a single access point.”

She went on to point out that the new hybrid architecture was more comprehensive; “The new hybrid architectures incorporate the speed and immediacy of search with the analysis and reporting features of BI.” and to find a justification for it – “the enterprise of the future will be information centered, and will require an agile,adaptable infrastructure to monitor and mine information as it flows into the company.”

Nick Patience and Brenon Daly of 451 Research went on to define The hybrid architecture’s capabilities in a bit more detail for Endeca’s version: “Endeca’s underlying technology is called MDEX, which is a hybrid search and analytics database used for the exploration, search and analysis of data. The MDEXengine is designed to handle unstructured data – e.g., text, semi-structured content involving data sources that have metadata such as XML files, and structured data – in a database or application.”

These definitions acknowledge the growing importance of information from everywhere, in unstructured as well as structured form, and the need to be able to access and analyze it in the modern enterprise. Information can, and does, come from anywhere – internal CRM systems, company independent blogs and forums, totally differentiated social media such as blogs and tweets, competitor websites, news services, and even raw data repositories. And it comes in the form of database records, blogs, emails, tweets, images and more. In the modern enterprise the need is to be able to analyze and use all this information immediately and easily.

Mining information from these disparate sources is not something that business analysts or product managers should be spending their time on. They need a reliable supply of the information where the semantics can be trusted, the information is up-to-date, and where the analyses can be set up easily. This is where the “plumbing” comes in. Two stages are involved: gathering the information, analyzing the information, then the user can take action on the intelligence provided. Companies like MuseGlobal take care of the first stage, and repository and BI companies take care of the second.

Some companies, like Endeca, take care of both stages, but then you are locked into both products from a single vendor, and it is not usual that they are both “best of breed”. So MuseGlobal concentrates on what it does best – gathering, normalizing, mining and performing simple analytics on data - and seamlessly passes the information on to your choice of Data Warehouse, Repository, BI, analytics engine – whatever best suit
e
s the company’s needs.

What this means is that your organization sets up a Muse harvesting and/or Federated Search system once, pointing to the desired Sources of data, configures authentication where needed, and determines how the results are to be delivered to the analysis engine, specifying a choice of standards based or proprietary protocols and formats. Adding new Sources (or removing unwanted ones) is a point and click operation, and the Muse Automatic Source Update mechanism (and our programmers and analysts) ensures the connections remain working even when the sources change their characteristics – or even their address! About as close to “set and forget” as you can get in this changing world.

On schedule, or when requested by users, the data pours out of Muse in a consistent standardized format, with normalized semantics and even added enrichments and extracted “entities” or “facets” (Endeca’s terminology) and heads to the next stage of the information stack. This raw and analysed data input means the BI system (or whatever is in use) can now deliver more comprehensive analyses so staff can now concentrate on the information they have in front of them, not on seeking it in bits and pieces from all over the place.

And the information is not only from many sources, it is in varied formats. Forrester have just released a report which asks the questions “Have you noticed how search engine results pages are now filled with YouTube videos, images, and rich media links? Every day, the search experience is becoming more and more display-like, meaning marketers must align their search and display marketing strategies and tactics.” So the need to handle a complete range of media types and convoluted structures is becoming paramount or the received data will be just the small amount of text left over from the rich feast of the retrieved results. This is a topic for another blog, but suffice it to say that Muse can deliver the videos as well as the text.