So much has been made of Big Data that it is often challenging to view the forest from the trees. 

Many have noted the challenges in Big Data: It can be time-consuming and expensive to implement with an uncertain payoff. Further, because Big Data can be used in a myriad number of ways, it does not lend itself to a simple concrete distillation like, for example, an investment in hard physical assets that one can see and touch.

As a result, many executives are unable to articulate exactly what Big Data is, or do not have an intuitive understanding as to what problems Big Data can solve.

Big Data is complicated and requires looking at information in new and detailed ways, which can make it feel overwhelming.
Big Data is complicated and requires looking at information in new and detailed ways, which can make it feel overwhelming.

Thus, it is instructive to address the concept of Big Data from three perspectives, in the hope that it can bring clarity to those who may not view it for what Big Data actually is: a new, critical solution.

  1. What exactly is Big Data, and how is it different from previous strategy planning tools and approaches?

  2. What are the specific problems in the marketing efforts at subscription-based companies, and how can Big Data solve those problems?

  3. What are the specific applications that can be undertaken only with Big Data — applications that provide tangible, quantifiable improvements to a subscription-based company’s bottom line?

Before we attempt to answer these three questions, realise that the forest is this: For subscription-based companies, Big Data needs to solve a business problem, one that is important, tangible, and substantial. Otherwise, Big Data is an interesting intellectual exercise but nothing more.

First, what exactly is Big Data, and how is it different from previous strategy planning tools and approaches?

Big Data is such a misused term that if 100 executives were put into a room, 100 different definitions of Big Data would result. At Impact Consultancy, Big Data is an analytical approach that possesses the following unique characteristics:

  • Extremely large data sets that were previously too cumbersome to manage and manipulate with any ease or consistency.

  • An ability to drill down much more easily and deeply than previously into a large number (often 100 or more) of micro-targeted market segments.

  • Timely (including real-time) updates at a similar level of detail to facilitate much more rapid changes in strategy than was previously possible.

  • Sufficient integration of each of these characteristics to identify, pursue, and capture new opportunities more quickly than was previously feasible. 

You might have noticed the word “previously” appears in all four characteristics. This is by design, as this definition implicitly addresses one of the most common questions asked by those trying to wrap their brains around Big Data: Why is Big Data different from prior planning and analytical tools? Why did the term “Big Data” originate only recently? Did the business world not have data and computers to tackle Big Data previously?

Yes, those things existed, and there were certain instances where very large data sets could be manipulated at very low levels of granularity. However, the combination of drastically greater computing power and Internet-based access to a plethora of new and increasing data streams eliminated two key technological hurdles that prevented Big Data from becoming a broadly used, institutionalised tool until the last decade.

Second, what are the specific strategy planning problems at subscription-based companies, and how can Big Data solve those problems?

Admittedly, the definition above skews toward the abstract rather than the concrete. However, given the multitude of applications for Big Data (discussed later), such a broad definition is required to encompass such diversity.

Thus, it is useful to take the next step in detail and specificity, which is to describe Big Data in terms of what types of problems it can address in business planning processes.

Based on our experience with many subscription-based news media companies, when it comes to allocating marketing resources, most planning processes suffer from three important, inter-related problems:

  1. Problem No. 1: At most news media companies, 50%-90% of the total value of a circulation department’s non-core (i.e., new and at-risk) subscribers is concentrated in a small percentage of that circulation. Yet most companies do not know which is which because they have not done the proper analysis.

  2. Problem No. 2: Without the right data or analysis, most news media companies set acquisition and retention goals without really understanding their true economic impact at the right level, making it impossible for that company to have any chance at assembling an integrated and optimised strategy and budget.

  3. Problem No. 3: Most news media companies are unsuccessful in tracking the effectiveness of any strategy or new initiative at the optimal level of detail at the optimal frequency. As a result, no fine-tuning of any implemented strategy is feasible, even if the original strategy goal was correct.

Each of these problems can be overcome only through a formal programme that relies on the power of Big Data to parse very large data sets into dozens of micro-targeted segments in a timely way. To fully understand what is meant by this, let’s revisit each of the three problems outlined above and how Big Data tackles each with a compelling solution:

  1. Solution No. 1: Analysis

    Data is collected (often from the Tracking module described below) to perform analysis at the right level of detail to understand the true economic sources of underperformance, which guides future goals, optimal strategies, and budgets.

  2. Solution No. 2: Strategy and budget

    Using insights generated from the analysis implemented at the aforementioned level of 100 or more micro-targeted segments, an optimal strategy can be identified and translated into an explicit marketing budget, which will subsequently be tracked against actual future performance.

  3. Solution No. 3: Tracking

    Using the milestones and goals generated from the strategy and budget module, actual performance can be tracked and compared at the same level of detail, at which point discrepancies can be identified and analysed at that micro-targeted level.
Dealing with Big Data problems requires analysing the situation, assessing the strategy and budget, and setting up tracking solutions.
Dealing with Big Data problems requires analysing the situation, assessing the strategy and budget, and setting up tracking solutions.

Taken collectively, these three solutions combine to create a mutually reinforcing, cyclical solution. All three solutions are required, as each step helps to inform and improve the effectiveness of the other two solutions.

In total, Big Data is able to address each of the major shortcomings in historical strategy and resource allocation planning, which, in turn, can identify opportunities that would not have been uncovered otherwise.

Third, what are the specific applications that can be undertaken only with Big Data — applications that provide tangible, quantifiable improvements to a subscription-based company’s bottom line?

Everything just described probably makes complete intuitive sense. However, without actionable tactics and initiatives that add tangible incremental value to a business, Big Data, or any other analytical approach, does not matter.

As a result, we at Impact Consultancy want to share a handful of specific applications that have been used at one or more news media companies over the past five years, each of which represented a six- or seven-figure performance improvement opportunity annually.

Note that while all of these applications are from the area of subscriber acquisition and marketing, such an approach works equally well in practically any business function where data resides.

It is important to note that none of these applications could have been undertaken in any meaningful way without an explicit Big Data approach. Using Big Data, our news media clients were successful in the following initiatives:

  1. Identified programmes within certain acquisition channels with costs per order that were so high that an adequate rate of return could never be generated. Those programmes were terminated and marketing dollars redeployed to other acquisition channels.

  2. Expanded online starts programme, which boosted overall rate of return and generated substantial starts growth with no discernible impact on starts from offline channels.

  3. Pinpointed ZIP  codes with the greatest “head room” for increased growth, which allowed for a US$5-$20 increase in commissions paid to motivate more sales from specific vendors while still retaining a positive rate of return for those starts.

  4. Reduced emphasis on certain subsets of direct mail subscribers, shifting those dollars to greater incentives for other types of direct mail starts and campaigns.

  5. Isolated developing delivery routes that were subsidised by newspaper yet were performing at or above agreed upon levels, which allowed newspaper to curtail unnecessary subsidy payments and redeploy to needier routes. 

  6. Identified acquisition vendors that were consistently generating under-performing starts. Shared this analysis with these vendors to explain their underperformance, make them more accountable, and negotiate for better terms.

  7. Quantified the impact of printing facility’s economies of scale and its impact on how much management should be willing to spend in acquisition cost in order to secure marginal subscribers that do not generate an adequate rate of return on their own.

  8. Determined how much extra money would be earned (or lost) if the acquisition budget was raised or reduced 20%. This allowed for the selection of an optimal acquisition spend based on maximising total returns.

To summarise the differences between Big Data and prior analytical planning tools, consider the following graphic. 

Big Data allows for an increase in granularity not found in previous analytic methods.
Big Data allows for an increase in granularity not found in previous analytic methods.

Based on the viable increases in IT systems used (from which data is regularly extracted), number of segments managed, and the frequency at which this analysis can be iterated and completed, we estimate that a Big Data approach is close to 10,000 times more granular than prior analytical approaches.

Thus, it is no surprise that Big Data is more than just an academic exercise. Instead, it is a real solution with a much greater capacity to solve real business problems.