Blogs

Best practices in records management teach us that lifecycle retention and disposition policies apply to content based on context, independent of the content medium. But I can’t help but have nostalgic affection for records media when it comes to music. My collection runs the gamut of vinyl, cassette tape, CD, and now digital formats.  Changes in media over time coincide with my discovery of new music.  But I’ve many friends and family members who haven’t embraced new music nor new media types, who find themselves stranded in known tunage without a clue about new stuff they might enjoy if they only heard it.

Some stranded listeners might be disinterested, but for this curious type, I celebrated the arrival of Pandora.  Based on the Music Genome Project, Pandora leverages almost 400 attributes to describe songs and a complex algorithm to organize them. Each song is attributed numerous “genes” based on musical genre and is analyzed by actual humans and backed by a technician’s review for gene sequencing standards and reliability. Listeners can then create music channels music by providing basic druthers (a song title, an artist, an album) that Pandora uses as attributes to extrapolate and serve up music with similar traits that the user might like.  Users tune their channels with “thumbs up / thumbs down” valuations and other feedback to manage signal vs. noise.

Pandora’s rich metadata serves to guide users to music they like in an intuitive fashion that can seem almost psychic at times. But it’s not witchcraft. It’s a decision tool that leverages what’s already there:

• the metadata about harmony, melody, and rhythm;
• the structure of rhythm, form, and composition;
• and the context of lyrics.

If stranded music fans used Pandora, I believe they’d discover new worlds of similar,  relevant musical content to enrich their human experience.  One of Randall Munroe’s comics last week on xkcd.com illustrates the Pandora principle perfectly, and describes it in the “hover over text”: 

“The most exciting new frontier is charting what’s already here.” 

Like music, content is wealthy with attributes ripe for charting and business benefit.

Information volume overwhelms knowledge workers to where they’re stranded on content islands, unable to find the information that matters for trusted decisions in a timely fashion. And blind to what’s available to them in the form of rich, valued knowledge assets. Like Pandora does for music, ECM delivers decision tools for analysis of the metadata, structure, and context of information. Metadata rules, automated classification, and content analytics can be tuned (by humans!) to explore enterprise content genes to assess what information matters; to filter signal from noise; to examine content relationships; and to derive meaning.

With these ECM decision tools, humans verify classification results for accuracy, adjust automation rules for identification and collection, and can investigate contextual analysis results to determine significance and make sense of content. These tools empower workers with confident decisions based on content they can know and trust as accurate and authentic; enable information governance stakeholders to consistently meet requirements, and yield new business insights from the meaningful discovery of collective corporate knowledge.

Business users may not always “like” their content the way they like their music, but they need to know the content that matters. And trust it. My colleague Craig Rhinehart recently did some crystal ball gazing on ECM in 2020 and what the future may bring. I’m excited about those possibilities. And I’m fascinated by the new frontier at hand right now and what ECM decision tools can do to help organizations chart what’s already here.

Please note:  Druthers expressed here are mine and not those of IBM or AIIM.