Gordian Knot Analytics helps you cross the widening data analytics skills gap

Addressing the Human Capital Gap in Data Analytics

Much has been written about the growing data analytics skills gap. How this demand is sizable and growing for individuals with strong data analytics and advanced statistics skills. The US Bureau of Labor Statistics indicated that demand for Data Scientists place it in the ten fastest growing jobs. This growth is expected well into the future, but the situation becomes more clear when examining actual hiring data.

The job posting site, Indeed, currently lists over 20,000 openings for “Data Scientist.” Adding “Big Data” to the search increases the count to over 45,000!  Postings for Data Scientists exceed 1,000 every day – that’s a lot of demand!

Marketing data analytics tools from Gordian Knot Analytics Group, Inc. can take much of the pressure off your recruitment requirements and dramatically increase the effectiveness of the human resources you already have in place, but first, let’s explore what’s at the heart of the problem.

Data Analytics Skills Gap Fueled by Marketing

A recent AMA survey of CMOs, revealed that by 2020 marketing analytics will represent 22% of their total marketing budgets. This is more than a fivefold increase from 2016! Marketing’s need to better understand what drives customer purchase behaviors is a clear source of growth of information collection. However, it will not be sufficient to simply collect and visualize the assembled data. To make matters more complicated, analyzing these rich datasets is not straightforward and current tools do not provide the insights needed to understand why consumers choose one brand over another. Moreover, the skills of most marketing professionals do not include advanced degrees in statistics, complex math, or statistical software programming.

According to Spencer Stuart, 54% of CMOs believe that Big Data and analytics skills will be essential to their marketing strategy over the long-term.

This results in the risk of making faulty decisions based on limited analyses of data when the staff or the tools to analyze this data are incapable or otherwise not up to the task. Irrelevant, inappropriate communications, poorly spent advertising dollars, or ill-timed product deliveries based on faulty analyses adds up to dissatisfied and disenfranchised customers.

When used properly, this valuable resource promises to help us present our customers with better, more relevant choices than our competition. However, even the best of those companies utilizing data to drive content marketing – companies like Amazon, Netflix, and Google – do not always get it right. How often have you received the wrong or inappropriate recommendation for a movie or a suggested purchase?

Can Education Represent a Viable Solution to the Skills Gap Challenge?

Educating more data scientists at the University-level seems a logical starting point to create and recruit more talent. According to the American Statistical Association, the number of undergraduate degrees in Data Science (including statistics) has nearly doubled in the last four years—making it the fastest-growing degree among the Science, Technology, Engineering, and Mathematics (STEM) programs.

Education alone will not solve the data analytics skills gap
We Will Be Unable to Educate Ourselves Across the Analytics Skills Chasm

This sounds promising – until we dig a little deeper.  the number of universities granting undergraduate statistics and data science degrees has grown less than 4% since 2003. Three of the top universities offering data science degrees (Michigan, Miami, and Ohio State) had less than 170 students enrolled in their programs in 2015. This problem was further reflected in a 2015 McKinsey report predicting that if every graduate from every University program were employed, by 2018 there would still be a shortfall of 1.5 Million managers and analysts with backgrounds in data.

When we look to senior, more experienced roles the problem gets worse. The entire US education system graduated less than 4,000 PhDs in 2015 across both “Computer and Information Sciences” and “Mathematics and Statistics.” And given that many new doctorates are seeking to use their newfound skills for something other than commercial data science, the real supply to satisfy this demand is much lower.

Cultivating In-House Resources

Filling skills gaps internally with on-the-job and boot-camp-style training are options that some companies have attempted. However, as with any in-house option, the newly trained staff will still be new to data science and – even after training – are likely to lack the level of expertise needed to solve the more complex problems of handling, analyzing, and modeling Big Data to drive better marketing decisions. Further, these types of in-house developed programs tend to be expensive, require dedicated training resources, and are difficult to keep current when compared with the best practices that academic institutions can provide.

Recruiting from the Competition

A third route for filling data science roles is to recruit more data scientists with relevant work experience from other companies. These recruits will come with practical, real-world experience, and can hit the ground running, making this a possible solution. However, this strategy is difficult to execute at scale and really only transfers the problem from the companies doing the recruiting to the companies that end up losing personnel. These resources also carry increased compensation costs. To switch companies, recruits are demanding pay increases of 15% to 20%, on top of the high salaries they already command (a Glassdoor.com salary survey recently reported that ‘Data Scientist’ is the 16th highest paid job title with some entry level positions coming in over six figures).

To make matters worse for most companies, top recruits from any of these potential talent pools are likely to get snatched up quickly by companies with the best hiring incentives (e.g. Amazon, Facebook, IBM), leaving the rest of us to fight over whoever is left.

We face a situation where the world of education can’t produce qualified individuals fast enough, creating qualified staff in-house is likely to produce mixed results, and taking experienced talent from other companies is expensive and not scalable.

There is no doubt that qualified data scientists are key components to develop marketing strategies and capabilities that take advantage of Big Data, but it’s clear that we need a solution to close the human capital gap for these skills – urgently!

Gordian Knot is an Extension of your Data Department

Data analytics tools from Gordian Knot Analytics Group, Inc. are specifically designed to address these challenges. Our award winning proprietary tools perform the most tedious and complex tasks for you, allowing your marketing professionals to be more productive, and concentrate on being – well – marketing professionals.

  1. Gordian Knot’s software automatically combines the best aspects of machine learning, econometrics, and non-linear mathematics to generate the optimal solutions so marketers can focus on marketing;
  2. Determining which variables to include in a statistical analysis usually requires the most expertise, but Gordian Knot’s powerful software takes everything into account and automates the variable reduction process;
  3. Our software handles ‘blanks’ and ‘unknown values’ without having to do any additional data manipulation; and
  4. There is no need to convert your volumes of data – we can handle data in any format and of virtually any type (continuous, categorical, scalar, times, dates, and even alphanumeric strings) – again, a tremendous time savings for you.

Empower Your Marketing Professionals to Do What They Do Best

Think of it this way, before the time of ubiquitous availability of the modern portable computer and powerful software, one had to be proficient in software programming to get the system to perform even the simplest of tasks. You spent all your time programming the computer and not taking advantage of the information it produced.

Similarly, statistics and complex math are the tools we use to understand complex relationships in large datasets, however, these tools don’t provide easy-to-understand solutions for your marketing challenges unless you have the expertise to decipher the math outcomes.

Gordian Knot’s software does not pre-suppose a knowledge of statistics. People without backgrounds in math, statistics, or engineering can be immediately up and running building complex explanatory models, and applying those models to solve complex marketing problems.

Using Gordian Knot to Do More with Less – A Segmentation Case Study

A well-known online travel booking company wanted to apply segmentation scoring to their entire customer base after having completed a Needs, Attitudes, and Benefits (NAB) segmentation study of the general population of the consumer and business travel markets. This study was conducted using a detailed survey instrument fielded to a panel of respondents representing the United States and Canada. Although NAB segmentation studies provide rich strategic insight, segments are difficult to use because the company does not know the segment assignments of its specific customers and prospects in advance. The most common solution is to push out ‘gearbox’ surveys, which are smaller, 5-10 minute surveys, used to score customers. The drawbacks of this approach are that additional surveys are intrusive, lead to ‘survey fatigue,’ are costly, and are only marginally effective at assigning segments to your customers.

Gordian Knot Analytics Group TARGET achieves a higher rate of accuracy without using surveys or applying over-qualified staff
GKAG TARGET achieves a higher rate of accuracy without using surveys or applying over-qualified staff

The online travel booking company’s segmentation service provider attempted to address this problem by assigning a team of three PhD statisticians to develop an algorithm that could score all customers into segments using only internally generated data, without using gearbox surveys. This team worked full-time on the project for three months and achieved a scoring accuracy of only 19%. In contrast, when the travel company used Gordian Knot Analytics Group’s TARGET tool, a single employee with a basic background in statistics was able to build a scoring algorithm with 80% accuracy in about an hour. This represented a significant savings in resources and costs, but more importantly, actually improved the accuracy of their customer segment scores!

Gordian Knot’s Solutions for Marketing Will Help Make You a Better Marketer

Gordian Knot’s marketing analytics solutions go beyond simply visualizing data in charts and graphs (although we do that quite well too). Our proprietary software builds highly accurate predictive models specifically created from your marketing datasets … without the need of having costly, highly skilled data scientists on staff.

Learn more how we can help you get more out of your marketing efforts. Visit our website (www.GKAGmarketing.com), or send us a note at info@gknotag.com.

download button

Download Addressing the Data Analytics Skills Gap article, by providing your email below:

Contact us, or subscribe today . . .finger point button

Written by Bill Taylor

Executive Leader | Change Agent | Pragmatic Futurist Bill Taylor is responsible for developing and implementing GKAG's Client Success functions – marketing operations, sales, and client support. Bill Taylor's achievements over the past 30 years include accelerating sales growth, market share capture, earnings improvements, and numerous successful new product introductions for global companies such as AT&T, Motorola, Panasonic, and Philips. Bill holds an MBA from the University of Denver.

One comment

  1. Since posting this article, HBR September-October 2017 issue wrote a brief article on how Data Science is disrupting the job market. The growth for data scientist jobs will approach 40% by 2020, and that salaries will be higher than the average for all jobs requiring a bachelor’s or graduate degree. The vast majority (81%) of the jobs will require three or more years of experience. This confirms that it will continue to be very difficult to fill openings (2.3 million listed jobs for data scientist in 2015) and will continue to be challenging for the foreseeable future.