Big Data & Predictive Analytics: A Conversation with Doug Levin, Quant5

Thursday Dec 19, 2013 by Natalie Nathanson - Founder, Magnetude Consulting

Magnetude recently met with Doug Levin, CEO of Quant5, a SaaS startup dedicated to improving the revenue and cost performance of customer-facing operations through an entirely new breed of Consumer Analytics.  During this interview, Levin shared with us trends he’s seeing in big data and analytics and advice on when and how startup leaders should look to leverage predictive analytics for their business. Below is a summary of the discussion.

We hear a lot of buzz these days about data and analytics, but many don’t truly understands predictive analytics. How would you describe this to a business person just getting familiar with this realm?

In its most basic form, predictive analytics tries to paint a picture of the future by unveiling insights from both historic and current business data – that sometimes can be superimposed with social data – to produce new business optimizations and actions that result in incremental revenues, costs-savings, and budgeting efficiencies.

Predictive analytics can be used, for example, to predict:

  • The next product a customer will purchase, its price point and timeframe
  • The optimal schedule for a retail store with sales and support throughout the day, month or year
  • Where to place items throughout a store to generate a high probability for these items to be consumed on their own versus in conjunction with other products

Generally speaking, understanding product relationships is very important, and being able to combine data from a system like Salesforce with your social data can improve your ability to monitor data while enabling you to gather new insights and respond more efficiently.

What trends are you seeing in the industry overall?   

Over the last two years, big data has certainly become a mainstream concept, and this trend has grown because some very large companies have put an emphasis on customer analytics, and they’ve used this data to deliver tangible insights that have resulted in significant value. Companies like Google, Wal-Mart, Bank of America, and Charles Schwab, have essentially paved the way for smaller mid and large sized companies to follow suit.  This has created a demand amongst these smaller companies to be able to replicate what the larger companies are doing.

Large companies have hired data scientists and purchased significant infrastructure to produce and sift through analytical data. In contrast, smaller organizations aren’t necessarily in a position to make the same kind of investment, though they do have a similar need. This has opened the door for companies like Quant5 to produce affordable and easy-to-implement solutions that deliver the same kind of quality customer analytics that the bigger companies have.

Another interesting trend we’re seeing is that some businesses are building an entire culture around being a data-driven company. Companies like these are introducing ways to drive their business and inform their managers, executives, and decision-makers with insightful data on where to move the company on a strategic and a granular level. When you have companies who use data so devotedly on a day-to-day basis—within and across departments—the need for effective data capture and the proven value of the data grow significantly.

Along with this comes social data, which has produced an enormous wave of new data hitting marketing departments, who are now tasked with reviewing these volumes of data and ensuring the insights feed more than the marketing department alone.  When done correctly, this data has implications for other executives who will uncover insights into the business, their customers, and their competitors.

We work with many startups who want to be data-driven, but aren’t always sure if they’re far enough along or have enough data to truly benefit from analytics. Do you have any rules of thumb regarding when a startup is ‘ready’ for predictive analytics?

Simply put, the startup must have enough data to make the models work.  Generally speaking, you need 24 months of data, which should be your starting point.  It’s also desirable to have multiple data streams such as a CRM and a marketing automation tool, like Salesforce and Marketo, for example, in order to build an initial data set that can yield more insights.  These combinations are important, but it’s also important to have the base amount of data. Secondly, they should have personnel who can support the project and continue supporting the project on an ongoing basis. Predictive marketing and predictive analytics projects are not just a one-time deal.

Back in the day, before we had Cloud-based predictive analytics, a big consulting firm would put together a team and live with a client for a month or two. They would generate a “study” with a set of spreadsheets, which were valuable, except that they had to be redone annually. Today’s analytics processes are iteratively reacting to the flow of data that’s out there, whether it’s structured or unstructured data streams. When done correctly, establishing predictive analytic processes means you are constantly coming up with insights that help guide the business.  If you are not dedicated to the idea of being data-driven or using analytics as an important element of your business decision making processes, then it is questionable whether or not you get long-term value out of it.  Being dedicated to a long-term solution and fostering the appropriate culture around it is really important to obtaining value.

Natalie Nathanson is the Founder & President of Magnetude Consulting.  You can find additional content on the Magnetude Consulting blog located here.  You can also follow Natalie on Twitter (@_Magnetude) by clicking here.

< back to all blog entries

Categories: