How To Make "Data Science" A Reality

A few thoughts on how data science can help any organisation (big and small) improve performance, regardless of their in-house capabilities.

Introduction

"Data science", "machine learning" & "artificial intelligence" are terms bandied around somewhat indiscriminately - but for the vast majority of organisations they belong in the 'sounds exciting, no idea where to start' box.

There are broadly three possible approaches you can take, which apply whether you work at a public sector organisation, a startup or a global multinational.


Hire a data scientist

The obvious choice would seem to be to hire a data scientist. Which is a great idea, but:

  1. It really takes a good data scientist to know one, so you have a chicken and egg problem, and
  2. Good data scientists are expensive (with good reason - it's hard)

Buy an AutoML tool

In the last couple of years a new option has become available, which allows technically-minded analysts to create machine learning-based models, without needing any data science training.

These so-called "citizen data scientists" use out-of-the-box data science tools, often known as AutoML (automated machine learning).

It's a much more accessible route in for many organisations, both financially and time-to-output - but it's not a panacea. You still need to define a clear strategy and your "citizen data scientists" actually need to be pretty clued up on data science to get the best out of the tool.


Outsource

The alternative, of course, is to outsource - be that through a consultancy who include data scientists on their team, specialist agencies who will build and maintain a model for you, or freelancers to get you started.

Done right, you get the benefits of data science's insight without the long-term commitment. The trick is in getting the brief right at the start, and finding the right balance between cost and quality.

Consultancies and agencies will normally offer to build you a model to solve any number of different problems - from customer retention to stock levels to staffing numbers. Others (like our good selves) are specialised in a particular problem such as demand forecasting


Conclusion

What's right for you depends on your budget, aspirations and timescale. One piece of advice we would offer though, is take the time up front to be clear on what you're trying to achieve.


If you'd like to learn more, these articles might be of interest:

Thanks for reading.

What we do

Skarp uses machine learning-powered predictive analytics to generate accurate, automated demand forecasts - and an explanation of what is actually driving performance.

By removing uncertainty and quantifying the impact of factors affecting performance, Skarp can reduce costs and improve customer satisfaction.

We offer a fully-managed service, designed for organisations with limited in-house data science resources.

IDENTIFY PERFORMANCE DRIVERS
OPTIMISE INVENTORY & STAFFING
MEASURE MARKETING ROI
Find out more

Get started for free

There is no setup fee or minimum contract term with Skarp, and we offer all new clients a proof of concept free of charge. We believe the accuracy of our forecasts will speak for itself.

Book your demo