Who Needs Data Scientists? Get Business Value Now

Thursday July 10th, 2014

I would like to make a proposition: Businesses can unlock the value in their big data today without the need for data scientists.

A great deal has been made about the cost of data scientist and/or how hard it is to find the skills needed to leverage big data. I don’t think this should hold you back. I am not suggesting there is no place for data scientists or highly specialised skills I am merely stating that a lot can be achieved before having to bring them in.

There is value locked up in big data

I don’t think anyone disagrees that there is value locked up within our organisations data, as well as the data that is publically available across the Internet. But how to best unlock this value?

Most organisations have people who can unlock the value

The best people to unlock the value are people who know your business. Organisations already have an array of analysts who know their business well: Business Analysts, Finance analysts, Process analysts, Engineers (depending on your industry) and IT analysts. I would like to propose that you are better off up-skilling your existing people before investing in expensive data scientist resources.

Business analysts tend focus on the business context and a business outcome.  Many analysis have already studied many of the data scientist concepts in their undergraduate degrees e.g. statistics, economics, engineering. They already know the fundamentals. Once you give them the tools they are more than capable to augment their existing knowledge to get value out of the data.

As a consultant that advises clients on best practice analytic practices I have come across many in-house analyst that have risen to the occasion. I recently worked with a process engineer with a manufacturing background. After 5 days formal training on a popular data-mining tool he was uncovering insight that had been missed by professional analytic consultants. A few years on he is now a professional data scientist. But that is another topic altogether.

Obviously to have a data scientist or analytic consultant available to advise and support them would be even better. But this shouldn’t stop you from starting with whom you already have on your team.

Modern tools enable you to focus on the business outcome rather than the math.

Modern toolsets have significantly improved the ability of business analysts to get value from big data, let’s take two examples:

Teradata Aster – Designed for loosely structured and structured data it has a simple SQL interface (A skill most analyst already have). This focuses the analysis on defining the outcome in SQL and not on complex code. You can take weblogs, identify user interaction sessions within the logs and then run a nPath query across them in a single SQL statement. Within minutes you can find the most common paths to shopping cart abandonment or what customers are doing prior to calling your call center.

SAS Rapid Predictive Miner – An Excel GUI with simple wizard interface that unleashes the power of SAS in the background. A marketing analyst can create a targeted mailing list in minutes. The beauty of this is that the model that is generated in the background can be opened up by a data scientist and be audited or refined using SAS’s more advanced Enterprise Miner tool.

There is still a place for data scientists but it should not limit you.

Now, having just upset all the data scientists I know – Yes, there is definitely a place for data scientists. Data scientists can squeeze more value out of each model and can apply many different approaches that may see results when simpler methods do not work.

Get on with it.

I am not saying you don’t need data scientists – What I am saying is you can get started with what you have and more importantly start driving value immediately. Focus on faster time to market and relevance to your business. If you do manage to hire a data scientist it is only going to add to the value you are already generating.

Gareth Clayton is a Senior Industry Consultant at Teradata with over 16 years experience in business analytics and information management. He has a diverse background in many industries but primarily in Telecommunications and Banking. Gareth is also passionate about educating the next generation where he has been a guest lecturer at La Trobe and Victoria Universities on the practical application of predictive analytic theory in a business context. Follow Gareth via twitter @AnalyticsROI or connect via Linkedin.

The following two tabs change content below.

6 thoughts on “Who Needs Data Scientists? Get Business Value Now

  1. avatarKen Neumeister

    Perhaps a better approach to discussing this is to demystify the word “data scientist” to recognize as you mention that organizations probably already have people who understand what counts most about the data: the domain expertise and the nature of the data that limits its utility. Data scientists are a misnomer when in fact they are just computer sciences as computer sciences was originally conceived in the 1980’s with a focus on algorithms instead of user interfaces. Modern software packages go a long way without needing to customize or optimize algorithms.

    1. avatarGareth Clayton

      Good point Ken. You could rephrase my article to “You already have data scientists – you just need to encourage them”. I suppose I was referencing the new breed of consultant offering their services to organisations as “Data Scientists”.

  2. avatarBilal

    Perhaps, we have been finding the wrong resource for Data Scientists. Mostly Data/ETL Engineers transit to Data Scientist position, because we have assumed that Data Science requires Computer Science knowledge. IMHO, a person with a background in physics is rather better suited because he has that instincts to play with data. If you find such people in-house, then what better than to have a Industry Expert as your Data Scientist.

    1. avatarGareth Clayton

      Bilal, Thank you for your comment. Interesting that you mention a requirement for computer science. Most of the data scientists I know have a background in inferential statistics. It definitely helps if they like to dabble in code especially if you want to use Hadoop map reduce platforms. At the end of the day they need to focus on the business outcome.

  3. avatarRon Swift

    Varying organizations have and will define the role, responsibilities, and decision making visualizations, provided by the data scientist, that they will define with great variances. This new role is sorely needed, and with multiple business, technological, accounting,man altricial, and communications or marketing skills. There is A need everywhere I gave seen in the past fifteen years, throughout the entire world, for the people with truly good analytical skills. The article may not truly be titled accurately, as the author states within the prose,the appropriate need for data scientists.


Leave a Reply

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