3 Ways to Optimize Big Data and Predictive Analytics

16902 Views
1 Comments
11 Likes

I recently had the pleasure of joining the Briefing Room webinar, which was a collaboration of Eric Kavanagh from Bloor Research and Mike Ferguson, an industry analyst in Business Intelligence and Data Management from Intelligent Business Strategies Limited. The topic we focused on was around Big Data and Predictive Analytics. Coming from the world of Predictive Analytics for over the past 12 years, this was an interesting topic for me, as we now see the emergence of Big Data;these two topics go hand in hand. The reason I say that is because with the emergence of Big Data as well as new data, such as social media on top of that – it creates more questions, and more importantly, more decisions.

Predictive Analytics can play a massive role in the simplification of Big Data. The role it plays comes down to getting the most out of this Big Data, and it starts with what some people refer to this as the fourth “V” of Big Data (Velocity, Variety, Volume), and that is Value. Predictive Analytics takes all of the data available and can pull the most important elements of that data, helping to create a predictive model or create an ideal profile to help organizations in many different aspects of the business. This includes areas like Marketing Optimization, where companies determine the characteristics of customers or prospects that are most likely to respond to a certain mailing or promotion. Another area is reducing customer churn, which involved looking for the main reasons people are leaving a company or service, or even what are some of the warning signs that someone is most likely to leave. These are just a few examples of where Predictive Analytics can play a major role in affecting a company’s bottom line.

Hopefully it is apparent the role Big Data and Predictive Analytics can play in affecting the bottom line and delivering a competitive edge, but how do you actually implement the two “technologies” together? Big Data is guarded by IT, right? You have to know programming or coding, or a statistician to really be able to use Predictive Analytics, right? This all used to be true and still can be in some instances.

3 Ways to Optimize Big Data and Predictive AnalyticsThe following are three ways organizations and vendors alike can optimize Big Data and Predictive Analytics more effectively:

1. Data Accessibility

The traditional approach to analytics or data has always been from an IT organization perspective. IT is in charge of storing the data, sometimes collecting the data, and many times even managing and integrating that data. IT organizations are critical in maintaining the company’s infrastructure, as well as the integrity of the data. They are great in this role within the organization, butit isn’t IT’s job to be able to tell executives why sales forecasts were down last quarter, or why an organization has millions of dollars of extra inventory on the shelves.

Streamlining the data access process doesn’t have to be difficult. Organizations need to allow the people that know the business, know the data they need, and know the results they need to get, access to the relevant data. These business analysts need access to all type of data - not just transactional data, but all of the relevant data, such as social media data, point of sale data, give them third party data to enrich their datasets. I understand that IT is always going to be protective of the data, and that in some instances there is going to be data that an analyst can’t directly access because of privacy issues or restrictions. Giving analysts direct access to the majority of the data, not only makes the analyst’s job easier, but IT’s as well.

With this process, analysts no longer have to put a request in the queue, hoping to get the right dataset they need, while freeing IT to work on other projects and other requests, rather than being a data delivery mechanism. The one caveat to this is that with great access, comes great responsibility. IT needs reassurance that data governance will not be compromised if more and more people are going to have direct access to the data.

2. Eliminate complexity

We talked about how Predictive Analytics can impact the business in many different ways, but the next step is how do organizations make it accessible and something that can be used in an organization - especially at the analyst level? It starts by eliminating the complexity of Predictive Analytics. Predictive Analytics tends to scare a lot of people because traditionally the thinking was that it is too complex, you need to be a statistician, or need to know programming to utilize Predictive Analytics. Obviously if you are a statistician or know how to do statistical computation and programming, it doesn’t hurt, but not everyone has a PhD in Statistics or knows how to write SAS or R code.

Give users the power of Predictive Analytics wrapped up in a nice little package that can be easily adapted and implemented into existing analysis and workflows, and XX. It’s understandable that not everyone will be able to use Predictive Analytics as you still have to have a little analytic knowledge to really get the most out of any form of statistics, but by simplifying the process of producing output and results through predictive analytics, this is a big step closer to broader implementations and better business decisions.

3. Decisions anytime, anywhere

Making data accessible and eliminating the complexity lead into the third way to optimize Big Data and Predictive Analytics; allowing key business decisions to be made anytime, anywhere. How do we take the complexity of building a predictive model from an array of disparate data, so that an executive or business decision maker can determine what step they should take next? Most decision makers don’t need to know everything behind the analysis; they just want to know what they should do next and have it backed up by the data.

You can give decision makers the ability to make decisions anytime, anywhere, with just a few clicks of a button through an analytic application. With cloud technology so prevalent these days, most vendors are adapting to delivering their solutions in the cloud. Delivering an analytic application in the cloud combines the complexity of big data and Predictive Analytics into something that is simple to understand and you can:

- Give sales managers the ability to run a forecast report on the fly to determine where they are falling short or where they need to move resources

- Give a vice president of marketing a better idea of what area code might yield the best return for the next campaign

Decision makers don’t want to wait for results, and giving analysts direct access to the data and the flexibility of the output the executives need can help optimize the decision making process and improve business processes.

As Big Data and Predictive Analytics continue to become integrated into the everyday aspects of the business, it is going to be critical for organizations to determine how to use them together. These are just three easy ways that organizations can make the most of their investments around analytics and improve their decision making process. Companies don’t have to re-invest in their infrastructure to make this happen, or hire new expensive resources to help. The beauty of this process is by utilizing these three methods, organizations can take advantage of resources and personnel that they already have in place, which can save organizations time and money because they are relying on the people that know the business reducing instead of 3rd party resources or internal roadblocks..

Author: Matt Madden is a Sr. Product Marketing Manager at Alteryx, Inc. He has been helping organizations realize the benefits of analytics for the past 14 years.

Article image © thitimontoyai22 - Fotolia.com

Like this article:
  11 members liked this article
16902 Views
1 Comments
11 Likes

COMMENTS

zarfman posted on Monday, May 13, 2013 9:15 PM

Hi:

I freely admit that I am somewhat dubious of many claims made for predictive analytics and decision analysis. I suspect in many cases puffery is alive and well. However your system may very well be the holey grail of analytics

Can your system handle linear or nonlinear systems with dependencies?

Can you optimize credit default swaps or various derivatives that are common in finance?

What about optimization of discrete vs. continuous data

For those of you who may not know, puffery is the exaggeration of the good points of a product, a business, real property and the prospects for future rise in value, profits and growth. Since a certain amount of "puffing" can be expected of any salesman, it cannot be the basis of a lawsuit for fraud or breach of contract unless the exaggeration exceeds the reality. However, if the puffery includes outright lies or has no basis in fact a legal action for rescission of the contract or for fraud against the seller is possible.

Regards,

zarfman


zarfman
Only registered users may post comments.

 



 




Copyright 2006-2024 by Modern Analyst Media LLC