You are hereBlogs / enherent's blog / Google Prediction API and some Thoughts
Google Prediction API and some Thoughts
Well this is interesting: Google Prediction API
So now Google is adding analytic capabilities to their storage service, which gives users more information about their data. Their BigQuery developer tool uses Google’s search capability to analyze large data sets. Much more intriguing is their Prediction API, which exposes Google’s advanced machine learning algorithms as a web service. Google claims that by using this web service, a user can make real-time decisions such as recommending products, assessing user sentiment from blogs and tweets, routing messages or assessing suspicious activities.
So this business analytics space is getting interesting. Its one thing to leverage a service to monitor social media, but leveraging a predictive analytics service is a different story. This underscores one of the key issues that I believe exists right now with business analytics; the space is fragmented, as is the way companies think about it. Within large companies, different Departments pursue their needs in isolation. Now I realize this is not knew news, but I see this as being very different than the ERP days when you found multiple products supporting various business units. That was about automation - with business analytics it’s about outcomes. It’s about a comprehensive view of data that enables prediction and delivers those outcomes. For example, how can one speak of Voice of the Customer when that voice is heard through different channels and there is no comprehensive view?
Is history repeating itself yet again? Is it best-of-breed or integrated application? IBM calls what is needed here an Information Agenda. I’ll buy into that, but would stress that any comprehensive view starts with the desired business outcomes. From there, a strategy for delivering those outcomes should be developed. Only then can one deal with the issues of business analytics.
What I’m seeing is just the opposite. Company "A" has a need for semantic search capability. They look for a solution to that problem. Next, they move to a sentiment analysis challenge and they look for a solution to that problem. And so on. Each challenge may require its own ROI to justify it, and inevitably you end up with mix of software products and SaaS-based offerings that now need to be maintained. What if you started the other way? What if the collective ROI were pulled together? What if a comprehensive solution for all these needs is affordable on the strength of that collective ROI? That same solution – easier to maintain and less expensive in the aggregate – was not affordable when you were simply looking at a semantic search need.
Call it an Information Agenda - or call it a comprehensive business analytics strategy – either way outcomes depend upon it.
Post new comment