Jeanne Harris on Building an Analytical Culture for Big Data

by   |   September 25, 2012 5:55 pm   |   0 Comments

Jeanne G Harris 2012 254x300 Jeanne Harris on Building an Analytical Culture for Big Data

Jeanne G. Harris of the Accenture Institute for High Performance

In a recent blog post for Harvard Business Review, analytics and performance expert Jeanne G. Harris argued the importance of “data literacy” in companies and laid out what some enterprises are doing to ensure that their workforces gain the necessary skills to analyze big data.

Harris, a senior executive research fellow with the Accenture Institute for High Performance in Chicago and the author of two books on analytics—Competing on Analytics (with Thomas H. Davenport) and Analytics at Work (with Davenport and Robert Morrison)—spoke with Data Informed about the challenges of fostering data literacy in the workplace and how employees can ensure they have the necessary big data skills to help the businesses and increase their own value.

Related Story

Insights from experts on how to start building your big data team.

       Read the story»

Data Informed: In your blog post, you discuss the importance of managers and business analysts becoming data proficient. But it seems like quite a  challenge for enterprises to teach all those managers and business analysts how to conduct data-driven experiments, interpret data, and create innovative data-based products and services. How will they go about doing this?

Harris: I decided to write the HBR blog because in big data there’s so much focus on data scientists and who they are and what their skills are. That’s all certainly important and appropriate. But in an organizational context, the data scientists, unless you’re a company like Google, are a relatively small percentage of the population. And there’s a much bigger employee group which has to work with those data scientists, use and interpret their data, and so we decided we needed to say something about all of those folks as well.

When I was writing my book, how you build an analytical culture was a really hot topic. And that means much more than having a cadre of skilled analysts. It’s really all about the sponsorship, the leadership, the managers supervising people, the managers who take the results of these analyses and incorporate them into decision making, and all the information workers in a business. It’s kind of the 80 percent versus the small percentage we call analytical  scientists or data scientists in an organization. This is really an outgrowth of the research we did around how do you build an analytical culture, what are some of the attributes of an analytical culture.

The first part of the answer to your question is, if you already have a very analytical organization, then you probably hire people who are culturally comfortable thinking about data and using data. They’re not statisticians themselves, but they certainly understand the basic concept, they understand when somebody presents an analysis to them, how to ask smart questions. They’re good at creating questions and working collaboratively with the data scientists.

So that’s all part of an analytical culture, but if you aren’t blessed with an analytical culture and you want to try to create one, then you probably have a workforce including management and top leadership who are less comfortable with some of these concepts. They don’t have academic training in the math and sciences and technical fields that you would have if you were in a different kind of business, and so it becomes a challenge to help build the analytical literacy of your workforce.

Companies are doing this in a lot of different ways. Obviously whenever you try to change a culture, it’s a long-term process, it’s not something you can have happen overnight. But going forward you certainly can hire more analytical managers, you can start to hire for some of the skills you’re looking for. You can try to identify people who not only have academic backgrounds, but also have demonstrated a logical, analytical approach to making decisions.

Big data is just a way to help organizations make better, more fact-based decisions, more data-driven decisions, to get insights and to be more innovative with data. So even in an analytical organization, you may find they are very analytical but they don’t use analytical for creative purposes. They don’t use it generating new insights, but they use it for justifications for the decision they’ve already made.

To me it really starts with having an analytical culture, one that is a meritocracy where the best ideas and the best insights win, and data operates in service of that. And we’ve found that analytical cultures are somewhat counter-intuitively extremely innovative and experimental in how they approach their business. So they don’t use data to justify a decision they’ve already made, they use it to gain new insights and understand their customers better. And getting an organization like that means you need a workforce with those kinds of skills.

Data Informed: Is it likely that large enterprises, at least, will retain trained data scientists to help interpret data for the lines of business, or will it generally be a case of “everyone is a data scientist”?

Harris: You’re not going to turn everyone in your organization into a data scientist. Nor should you; they still have their own job to do. But all managers and business analysts need to understand and have an appreciation for what analytics can and cannot do. So education certainly is one aspect of that. In a humble kind of way, that’s why we write books and articles, to educate management about potential of the use of data analytics in business, and to help them understand and develop an appreciation for applying the scientific method to management and the value in understanding all the different kinds of data out there. One of the big benefits of the big data trend is that it’s really helped stretch executives’ thinking about the art of the possible, about what kind of information might be out there that might help them solve existing problems or new problems in new and creative ways.

So a big part of it is educating people about what they can do with big data. Historically, executives tend to bring their questions to their analysts in terms of what they assume they’ll be able to. So if you assume the only information you’re going to be able to analyze is about, say, your ERP systems, you’re going to frame the question much more narrowly than if you start to think about all the information that’s potentially out there outside your organization.

So for executives in particular I think reading about big data, reading about applications, understanding it, is an important first step. Once they’ve got the vision for and enthusiasm for using big data analytics, then the next step is to help them be champions and sponsors of analytics within their own organizations.

Corporate education plays an important role. At Accenture, our clients often are looking to us to help them develop the skills and capabilities of their workforce, not just the data scientists, but functional users as well.

Data Informed: What’s the role of technology in teaching analytic skills to workers?

Harris: Here are a few things technology is going to do for us. As data visualization techniques get better, as the ability to access and analyze massive amounts of data gets easier, then it is more possible for people with lesser skill levels to be able to use and interpret data. Even if they aren’t building the algorithms, they might be able to use the results of the algorithms that somebody else has built. So I think we’re going to see a lot of breakthroughs on the technology side.

I teach a graduate-level course at Colombia University in New York, and one of the things that I’m seeing—I go to academic conferences, there’s a lot of discussion in MBA programs and in universities generally—there are many, many more courses popping up that are designed to help non-mathematicians, non-statisticians, non-coders, to understand how to appreciate, use and interpret the results of analyses.

So I think it’s going to really kind of a three-pronged approach: Corporations are going to have to build their analytical cultures, they’re going to have to support their people with training, they’re going to have to give them opportunities to work closely with data scientists.

There’s something about that collaboration that really helps build a manager’s appreciation for data-driven experimentation and innovation. They’re going to have to build their numeracy, to kind of fill in the gaps for people who don’t have that in their academic background. And I think the technology is going to make it increasingly easier to take big data tools out of the ivory tower and into the hands of managers.

Those of us who are more quantitatively oriented would like to sit in a room by ourselves and go through the data. We really can’t. It’s critical that data scientists are able to share that data with people and be able to share their enthusiasm and their knowledge so that they can help them understand the potential that’s there.

Data Informed: What can IT pros or non-IT people do to improve their big data skills?

Harris: Those of us in IT have always understood the importance of data governance and data management, but it’s always been treated by both business people and to an extent by IT people, a little like eating your vegetables. Nobody ever became CIO by doing a great job at data governance (laughs). I think with the rise of big data, there’s going to be a new focus on the quality of data and access to data, how we put some structure around data in order to really analyze it. Those are all things executives are going to have to deal with.

Big data is not really a new phenomenon at all. It was predicted and recognized even when we were just starting to adopt SQL, that the day would come when SQL would no longer be sufficient. So if you’re an IT employee, I think the first step is to build your skills around data governance, data management, understand how you can improve data quality. Maybe the harder challenge for people, and perhaps the more significant one, is how to operate in a post-SQL world. That’s where a lot of the training and education on the IT side is going in.

For non-IT enterprise professionals, I think step one is learning about the kind of big data that’s out there. Everybody sees social media — anybody who’s, for example, in marketing, has a really intuitive feeling that “if I could just somehow harness all that social media data, I could make sense of it and understand my customers in a way I could not before.”

For years marketing people have looked to third-person or syndicated data to try to get a more holistic picture of their customer. So in a way it’s almost easier for a marketing person to think about big data because it’s the next step in the evolution. They need to start developing their analytical skills to step up.

Of course, for those who haven’t been doing qualitative data analysis, it’s like being dropped on Mars.

Data Informed: When do you see data literacy becoming the norm down through enterprise lines of business?

Harris: What’s really very heartening about the big data trend is that it isn’t quite as big a stretch as it might have been five years ago. In the past five years, analytics have become table stakes in a lot of industries. People routinely use statistics and predictive modeling for logistics management, yield management, all across the supply chain, forecasting stock-outs, on the marketing side. So really, except for places like HR, they’ve already been starting to use and learning to build their analytical skills.

I think the difference is they’ve been doing that in a pretty structured way. The analytics they’ve been doing have been using data that comes from inside the companies that they control, or from vendors where they actually went out and purchased it. So to take that to the next level, to take all the data on the Internet, they can bring in and use and mash up in creative ways. It’s kind of a logical extension of that.

Data Informed: What are the main obstacles?

Harris: I think there’s an older generation that’s used to making decisions based on their gut, and not on data because it wasn’t an option when they were getting started in their career. The young generation has always had a computer, they’ve always had Excel, they’ve always been able to go out and search the Internet. It’s just like breathing to them. I think it’s going to be quite a significant change. I think for young people it’s not going to be that significant a change at all.

The biggest problem is we don’t have the science skills, we aren’t graduating enough people with enough quantitative background. They aren’t interested in using big data because it’s not something they are comfortable with, it’s not part of their tool kit.

Companies today are going to struggle building the analytical skills of their management and workforce. I think they’re going to be demanding more from academic institutions, but that’s going to take some time. I do think over the next five years I think we’re going to see more corporations stepping in to do more to build the skills and capabilities of their workforce and changing their hiring behavior.

I think you see that already. If you look at what the top five jobs are in surveys, they’re all very quantitative in nature. Statistician, actuary, mathematician, analyst. That’s where the jobs are. Whether the workforce catches up may be a political and social issue. And I don’t think companies are going to wait. I think they’re going to get much more proactive about building skills, both through job assignments and by pairing managers up.

Chris Nerney (cnerney@nerney.net) is a freelance writer in upstate New York. Follow him on Twitter @ChrisNerney





2f4030c6 41e3 4507 9efb 8be1e3aad05d Jeanne Harris on Building an Analytical Culture for Big Data



Tags: , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>