Netflix recommends your favorite shows and movies with great accuracy. Amazon recommends books that you might want to consider reading, products that you might consider buying and services that might benefit you. LinkedIn & Facebook recommends “People you may know” out of a large database. Google recommends sites while you type in your search term. It is fast, it is relevant, it is intuitive and it just is a better user experience. You now realize that this approach is now quite pervasive and effective. Then why is it that you and your BI user community today does not have access to recommended reports or dashboards? Let's dive into this to understand this better.">

Ramesh Sunder

SVP, Products and Solutions

How to maximize your BI investments from a recommendation algorithm?

It is 2017, almost everyone is familiar to the concept of recommendation engines. Netflix recommends your favorite shows and movies with great accuracy. Amazon recommends books that you might want to consider reading, products that you might consider buying and services that might benefit you. LinkedIn & Facebook recommends “People you may know” out of a large database. Google recommends sites while you type in your search term. It is fast, it is relevant, it is intuitive and it just is a better user experience. You now realize that this approach is now quite pervasive and effective. Then why is it that you and your BI user community today does not have access to recommended reports or dashboards? Let's dive into this to understand this better.

How do recommendation engines work?

Recommendation engines are one of the most widely adopted use case for Big Data. It is all about looking at a large pool of data, apply machine learning algorithms to make predictions on what would be the most relevant information in the context of a specific user. Now you may ask what has this got to do with my BI Investments? Well, perhaps there is a missed opportunity here that is worth exploring. Your BI and Big Data end-user community could benefit from a recommendation engine. While there is quite some literature on recommendation engines and what they do, I would like focus on how you can leverage this concept to extend your existing investments in BI.


What do you need?

To make proper recommendations, you may need various types of data. There are three essential components that any recommendation algorithm will need to deliver the right output:

1. Consumption data

Historic usage data provides valuable consumption patterns. Usage patterns from a larger dataset can surface the “wisdom of the crowd”.

2. User Context

A broader understanding of the user context, the department he/she works in, the business function, personal preferences etc. provide a deeper meaning and understanding to the usage patterns. This helps in making the right associations.

3. Relevant external data

These are factors that could have influenced the user behavior. It could be weather, market conditions, calendar events, seasonal factors etc. These are influencing data points that can help in improving the accuracy of predictions.

Recommendations in BI world /

Look at your BI/Big Data landscape. You have made big investments in past years to deliver the right reports and dashboards in the most modern tool that helps your business to make big decisions. You probably will have tools for analysis (Tableau, Birst, PowerBI, SAP Business Objects etc.). It is now probably filled with tons of reports and dashboards for multiple businesses. Some are relevant but many are redundant or unused. This is the ugly part. The good news is that in every company, that has made investments in Big Data and BI, has components that can be leveraged for building a recommendation engine. Reports and dashboards are continually consumed by end-users and different reporting tools and databases keep track of the consumption. User preferences and user context is also documented in a few ways within a database. Calendar events like financial period end close, quarterly close etc. is also available. It is now a matter of bringing all these data elements together to recommend reports and dashboards to users.

How will recommendations benefit your organization?

1. Increased Adoption and usage

User adoption of individual BI tools and Big Data investments will increase when consumption becomes simple and intuitive. Broader adoption results in better ROI.

2. Reduce IT Costs

When finding information in reports and dashboard becomes intuitive, IT is not burdened with tickets to locate the right report or dashboard. This is a better approach that allows IT to innovate.

3. Decrease Training Cost

When the wisdom of crowd is helping users to learn, your learning and training costs will reduce. This helps reduce the overall TCO.

4. Improved Governance

Your user community can self-govern by understanding best practices that are being followed. Also, by analyzing usage patterns, you might be able to uncover places where processes are not followed.

What is exciting about this approach is that it drives innovation and discovery within your organization that has not been seen before. Bringing such innovation to your users at lower costs will help you maximize your current investments while creating happy consumers. This allows new users to function effectively on day one and drives overall efficiencies. You will see BI tickets for new reports or dashboards reduce as end-users explore content intuitively.

Share your experiences on how report/dashboard recommendations can drive incremental ROI and help with increased user adoption of the BI investments.

333 W. Santa Clara St.,Suite #614 San Jose, CA 95113
Email: info@zenoptics.com Phone: 408-452-8500 | Fax: 408-452-8510 | www.zenoptics.com

© 2017 ZENOPTICS