Data Science and Analytics as a Product

Understanding the world of data science and analytics today can be a daunting task. A lot of it is because definitions are not clear. Sometimes data science is defined very broadly, sometimes very narrowly. All this can be very confusing and hard to pin down.

From a career perspective, it is not uncommon to stumble across questions like:

  • How do I break into data science without a phd?
  • Which software tools should I learn?
  • Is Business Intelligence still a thing?
  • Somebody help me please

From a business perspective, there is a lot of talk about big data and predictive analytics. But is the value of these tools clear to those doing the talking?

Here is the thing, data science is not a department on its own, it’s a cross functional ecosystem. You got no phd? You can still bring value to many other aspects of the process.

Defining data science as a single standing system is like looking at a car and thinking you need to have the entire skill set for building it. That means you would need to know everything around a car – from engineering, to distributing, to financing and project managing – to get into the automotive business. Luckily, it doesn’t work this way.

Enter data science as a product. Your job here doesn’t end at the Excel report or the online dashboard, it has to go deeper than that.

Let me make an example. Your boss has asked you to analyze customer growth development of the last 4 years. Now, here is the break down of this project.

  • Project manager – make sure all stakeholders are kept in the loop and that milestones are reached
  • Business analyst – lock down business requirements
  • Data architect – put together the data sets
  • Data scientist – put together the model
  • Subject expert – support the project by providing business knowledge
  • Salesman and marketer – sell the output of the project, get users to adopt the analysis in their work activities

The list goes on, depending on the complexity of the project. Point is, data science is not about building models and algorithms.

So back to the career questions above, what to do? Well, consider what you are doing already today and how you can further refine that to add value to the analytics product. Find out where you really want to be. Do you like spending hours tweaking a multiple regression analysis to solve a specific problem? Data science – the strict definition given here – might be a good fit. Do you like sitting with users and go through their issues and coordinating different people to solve their problems? Maybe project manager is a better fit.

What about business and predictive analytics? Do investigate the opportunities for specific problems in your company. Do not go into this with the idea that predictive analytics, on its own, will save the day. Remember, it’s about the overall product, not one part of it.

In companies, building one model won’t bring any value if users don’t understand it – if they can’t understand it they won’t even use it. Don’t even think you’ll force users to adopt data science and predictive analytics. This might work for check the box exercises where there is little thinking involved, it won’t work in analytics as there needs to be a proper understanding of the analysis in order to derive value from it.

In conclusion, everything is a product, analytics and data science too. By looking at things this way, you can decide which aspects of your business or career need to change to leverage data for a real benefit. Analysis does not end with the report – a car is not done when the project is finally on paper.

Leave a Reply

Your email address will not be published. Required fields are marked *