I love new years! After having taken a long delayed break towards the end of last year, finally spent some quality time with the family, finally had a time to just think about where I have reached and where I want to go, like most of you, I am all charged up to start another year.
With a fresh start feeling that 1st of January brings; a lot of us plan to figure out what we would do differently this year. Whether it’s a healthier lifestyle or taking the long planned entrepreneurial leap on the personal side… Or getting more done, finally getting to that long awaited innovation project on the professional side – we all want this new year to be better than the last. ‘This is going to be ‘the’ year’… we tell ourselves.
As I wrap up my first week of work this year, I notice another trend in most of my customer meetings this year. A bulk of New Year discussions have been around the theme of ‘this is going to be the year when we finally do something about big data. We have a ton of it and everyone seems to be doing interesting stuff with it. I bet we can pull out something amazing out of it too. But where do we start?’
While I cannot help you get a start on a healthy lifestyle, I can offer my two cents on how to start with big data this year. Most impactful data science insights come out of large data sets of user interactions – purchases, views, swipes, clicks, emails, documents, images, receipts, etc. Here is a four-step outline to get started with big data this year:
- Figure out what is your largest and fastest growing data set around user interactions – emails, expense records, TV program views, purchases, profiles, etc.
- Extract 10% of this data and engage a data science company to go through it to extract meaningful patterns
- Look at other publicly available information online or from social media, etc. to find opportunities to enrich this data.
- Get the data scientists to look at this combined data to create 5 models:
- Marketing – What is in this data that can help me get more customers?
- Sales – How can this data help us sell more to our existing customers?
- Customer Service – What does this data have to serve our customers better or make their experience better?
- Finance – Are there any cost savings or fraud prevention opportunities available in this data?
- Risk – Can this data help us predict or mitigate any business risks?
With a good data science partner, you should be able to find at-least one model that can help your business do significantly better than it did in 2015. This should give you a direction to get started with your data science work.