Product

Data analytics model’s life cycle

What does a data analytics model’s life cycle look like? How often do they get deployed in a company? And what happens to the ones that aren’t deployed? Sunil Thomas, CEO for CleverTap recently answered this question on his Quora channel.

Just like software development methodologies have progressed over time to become more agile and iterative (as compared to the traditional waterfall SDLC of yore!), data analytics models also have started becoming iterative. Companies who adopt these modern agile techniques to their data analytics process and roadmap are definitely in better shape as compared to companies who don’t.

Technologies today be underlying infrastructures like Hadoop, or even modern BI and visualization tools like those offered by Arcadia Data or Tableau are very adaptable to an iterative approach to data analytics where you start collating in the correct data over a period of time but every week or two there are some pieces of the puzzle being visualized and available to the people who consume this data.

On the digital front, whether for your websites and/or your mobile apps we at CleverTap have seen a massive shift from aggregated and session/page views based data analytics to much more people oriented analytics.

  • There is a unique user profile created for every new visitor on the site/app and over a period of time, these profiles get very rich because you record their behavior (online, social and offline) and also profile attributes for these people.
  • Very powerful user segmentation then lets you break down these millions of user profiles into smaller, manageable user segments so you understand your users much better that you used to be able to.

In short, even for data analytics use modern Agile methods – Scrum is very popular – so you can get key people involved all through the process and iterate incrementally so that everyone reaches their set goals with data analytics.

,

4000

55 billion

1 billion

10 billion