November 10, 2010 / by Enrico Montana

Socializing your enterprise analytics boils down to a single word—integration.  Even though we can agree that this is a core need as Social Intelligence (SI) permeates across the organization, it’s not a straightforward operation that can be managed in a one-time project.  Like the BI of your existing systems and processes, SI is an evolution.  As you learn more and ask deeper and broader questions, while reacting to changes in the marketplace, your needs and your vision (and therefore your requirements) will evolve.   However, if you wait until there is a crystal-clear understanding, well let’s be honest—you’ll be waiting a long time.


Before you jump to integration, your first step is an understanding of goals.  Where are you going?  Figuring that out will help drive two major components for integration—what the data points are in your existing BI ecosystem that you’re going to need, and how you’re going to align your SI taxonomies to appropriately merge with that data.


Once you have a concept of your integration goals, you’ll need to set some standards.  From an analytics perspective, this usually shows up as one of definitions.  The values or “buckets” your data needs to be segmented by are driven by your BI taxonomies, but the definitions in SI are driven by the search representation.  Standardization will help you in a few ways.  The first is one of extension.  As social intelligence grows across the enterprise, you’ll have a single source of truth for the data underneath a given value.  The second is one of agility.  As you learn and refine what a value means, you can change the search representation of that value, while keeping the taxonomies constant.


Depth comes in three levels, driven by the goals you’ve stated in step one.

Level 1: The Mashup

Your dashboard will need an awareness of how to ask the same question from two different systems, receive the answer to that question, and transform it into a common user experience.  Nearly all of your reporting needs should fit into here.

Level 2:  Information sharing

Your systems will need to share and store information in two different places, requiring them to know where they need to put it (taxonomy) and how (API and Data format) they’re going to place that information.

Level 3:  Programmatic management

Not only is data being pushed in both directions, but the taxonomies and structures are also programmatically managed, such that when new standards are developed, they show up in all systems appropriately.

As you travel down the layers of depth, complexity and cost will grow.  Going back to step 1, understanding your goals will help you assess whether that investment is really required to get the answers that you need.

The Practical Aspects

As you work through the three D’s of data integration, the difficulty lies in the definitions.  In many ways, the depth is defined by the destination.  How you define your world and the taxonomies of measurement is one that includes a fuzzy arena where you move from managed (explicit tracking of concepts using searches) to organic (discovering and identifying topics using unstructured analytics).   You’ll need to be prepared for a number of challenges, and your platform will need to be capable of dealing with those for you.

Definitions change.  The taxonomy in your company may not change often, but how people talk about your products will.  This ongoing evolution will drive changes in the underlying definitions.  Since you can’t predict what you don’t know, don’t be afraid to start.  Just make sure your platform can allow for changes in definitions, and ease of back checking historical baselines as the searches evolve.

Taxonomies are static, but interests are not.  Know where you want to transition from a managed search definition to an organic discovery of what people are talking about.  For each subject you’re studying, think about how important sub-topics are from a managed versus organic viewpoint.  Managed viewpoints have the advantage of standardization at the cost of work.  Organic discovery has the advantage of surfacing what’s important (without the need for asking that question directly), but comes with the cost of not being a managed, and therefore consistent, point of interest.

There’s a lot of data out there, and always more ways to describe the same thing.  For each definition, you’ll be balancing recall (getting every single post that pertains to your topic) against precision (getting the right posts that pertain to your topic).   Every change results in alterations in measured volumes, sentiment distributions, and other metrics.  In a world where there aren’t right answers, be ready to commit to at least one answer.  And from my perspective, precision wins over recall.  I can address gaps in the comprehensiveness of my definitions over time, but as I move forward, make decisions, and discover new things, I like the assurance that the analytics represents true signal.

What has your experience been with socializing your analytics?

Tags : social mediaanalytics

Most Recent Posts

Cision Blogs Topics

  • Communications Best Practices

    Get the latest updates on PR, communications and marketing best practices.

  • Cision Product News

    Keep up with everything Cision. Check here for the most current product news.

  • Executive Insights

    Thought leadership and communications strategy for the C-suite written by the C-suite.

  • Media Blog

    A blog for and about the media featuring trends, tips, tools, media moves and more.