No Data Scientists Required: Big Data is All About Business Users - GoodData
BlogSupportDevelopers

No Data Scientists Required: Big Data is All About Business Users

Posted February 4, 2013 by in Big Data, Business Intelligence (BI), Customer Support, Thought Leadership

Big data doesn’t sound like a provocative field. In a recent panel at the Global Big Data Conference, however, I said that the hype that has been building around the data scientist is over. Judging by the reaction, you’d think I’d have said that Abraham Lincoln is alive and living in Ohio.

The issue was that I was talking about the new school of big data, the one in which businesses use intuitive, interactive UIs to derive value from big data and avoid the dependency on data scientists. The people taking issue with this didn’t see beyond big data infrastructure — and the data scientists who interpret what comes out of those systems.

The  panel I attended was called “Big Data Analytics: What’s the Next Big Thing for Data, Both Big and Small?” It featured one data scientist, three database infrastructure experts, a senior infrastructure architect and myself.

I stuck out like a sore thumb. As the Senior VP of Customer Success for GoodData, I was the only panelist focused on customer-facing activities. The tech guys were mainly focused on infrastructure issues, like speed, scalability and latency. I’m close to customers who are actually using the technology. It’s my job to be aware of how customers use big data to gain business value. So I see the world of big data in a different way.

I made three points that were seen as particularly controversial:

1. Companies need to provide solutions that a business user can understand.

Business users are hungry for expertise on how to use the massive amounts of data at their disposal. They often don’t know where to start, however, or who to turn to. Big Data providers need to educate business by providing best practices on how to derive value from their data. That involves metrics and visualizations that measure success, layered over the high-performing back-end solutions that everyone’s still talking about today. In other words, you need to know how to measure your marketing funnel, track customer engagement, benchmark yourself against the competition and maintain the health of your user base, among other functions. These are all business functions, and they must be presented in a way that anyone can understand.

2. Data in spreadsheets isn’t as powerful as a graph or a chart.

Spreadsheets are functional, but they don’t encourage deep interaction. To derive the greatest value from big data, users need to view and engage with their data on an ongoing basis. A spreadsheet doesn’t offer any motivation to do that. An intuitive UI, on the other hand, promotes regular useage, enabling companies to befriend their big data. Those visuals, in turn, encourage companies to make smarter, data-driven decisions. You just don’t get as much out of a quarterly report as you do from daily engagement.

3. The data scientist concept will die.

This is the one that really got people. Companies need solutions that enable them to use and customize their data easily, because it is the whole team, not just the individual analyst, that knows the business best. By offering business users intuitive data solutions, we bypass the need for the data scientist, who works in isolation. In fact, most data scientists are associated with the old school of business intelligence, where systems were so complicated that they needed someone with a data science background to run and get value from them. The new generation of solutions, on the other hand, is making it easy for business users to engage big data. An interdisciplinary team will see and use the visuals provided, and collaborate on the best decisions on a regular basis.

For me, speaking on the panel triggered a larger perspective of the market for big data. A new generation of data solutions is emerging. They are making big data accessible to the masses. That’s a hard fact for the old school to accept.

It’s time to move beyond infrastructure. We need to deliver applications that everyone can use, and turn big data into a democracy.

Like this post? Subscribe to our blog.

To learn how GoodData can help you monetize your data, ask for a demo, or call us at (415) 200-0186.

5 Comments

A Business User
Hmmm... Let me start by saying that I am not a Data Scientist. I am a business person who has made a living bridging the gap between business people and data. Helping them access, understand and manipulate data. This has always been focused on extracting "value" out of the data and making the insight to action process more efficient. I too used to be of the belief that the right tool will unlock enormous value and in many ways it can, but not all. Interrogating data is about the questions that you ask and the interconnections between data that is not obviously related. Visuals are great but if you ask the wrong questions or can not conceptualize these relationships between unrelated data, the unfortunate end-user will be stuck in relational silo that will likely only lead to small, incremental value. The real value, the game changing insights that Big Data promises, this sits outside of the standard business users conceptual realm. Yes, education can help, but there is an epic force pulling against any business person trying to wear the Data Scientist cap - their job description and very limited time they can commit to performing this role (at least well). The other key point - selling your insight so your superiors buy-in to your findings. Insight is useless unless that insight can be turned into a consumable "story" that is compelling enough for management to want to upset the business and risk losing money. Data Scientists, whether they be Analysts, Journalists or Graphic Designers are just as skilled at telling the story as they are in finding the nugget. Context, hidden relationships and objectivity. These are the forces behind the patterns and behaviors of consumer activity that is fueling Big Data and is what will lead to true value, not just nice visuals in the hands of busy people focused on their job, not necessarily their companies strategy. I am not entirely disagreeing with you, I just think it is a big call to think the right "tool" will bring the end of the Data Scientist and deliver the same value that these skilled people can deliver.
Rich Homich
Agreed, effectiveness improves when users with business and market knowledge have more direct access to data. However, data scientists still have a role in organizing data and finding patterns in data with advanced analytic methods.
Guy Cuthbert
Great article Cliff - the sentiment behind gets my vote, 100%. I read it on Sunday, and what's been nagging at me since is the capability gap that I perceive in many businesspeople; the every people who *should* be keen to use great [accessible] analytics. At the risk of sounding patronising and/or overly-negative - and having spent years working with highly accessible, visual analytics tools and techniques - I'm dismayed by the laziness and unthinking approach that permeates from many in business. Too many people want a magic wand - they're not prepared to *think*. Too often, they expect the computer not only to do the extraordinary heavy-lifting of grabbing, crunching and presenting huge volumes of data in milliseconds (which, thankfully, computers are good at); but they also want it to think for them and tell them what to do, without them having to think about the information available or how it should be applied to their understanding of the business. Worse, there are cohorts within businesses who have only a feint grasp of how their business operates, whose understanding is flimsy at best. An effective data scientist remains a necessary catalyst in connecting many businesspeople with their data... but to be truly effective, the data scientist needs to be an excellent communicator and - often - more of a 'data animator' i.e. they need to bring data to life, and to use it to tell interesting, compelling stories. Stories which persuade, educate, amuse and impart understanding and knowledge - often challenging received wisdom and prejudice. So perhaps I agree that the data scientist is doomed... but that the data animator is needed for quite some time to come.
Griffin Schultz
I couldn't agree with you more. This shouldn't be that hard. We are trying to offer powerful safety data science to our customers in a manageable way that they can understand and use every day. The fourth item I would add to your three above is that solution providers need to be more open to sharing data across systems. After all, its not our data, its our customers' data. And they don't really want data, they want answers! Many times, one solution provider's data set only has part of the anwer, and it needs to collaborate with other data sets to derive the complete and comprehensive answer. We, as an industry, need to break down these barriers and data silos for our customers. Great post!
ashish
If your thesis is right then why only Nate Silver could make the correct projection of 2012 elections. There were other data scientists doing the same thing. I think what you wanted to say is if it is not an exception data scientist then let the users figure it out. Because anyone can make wrong prediction, read about failed Boeing Dreamliner.

Leave a comment

(required)