In today’s competitive marketplace, every business is under pressure to win customers and increase margins. Various sales channels, marketing techniques and pricing strategies are at their disposal to do this – but how to make the best use of them?

This question has been valid ever since the introduction of the marketing discipline. Changing business models and increased market complexity have introduced new dimensions to this question, but the overall ambition remains the same: to ensure that a company’s marketing efforts yield the desired financial outcomes. To raise the bar even higher, companies now want to be able to see few steps ahead, outperform the competition in a changing customer landscape and steer the business towards optimized returns.

What has happened & why?

As you can read in the previous episode, high-performing Finance teams in their role of business partner and insight generator, cooperate with the business to ensure that decision-making is always supported by relevant insights.

They don’t limit themselves to traditional Descriptive Analytics methods that mainly focus on explaining the past to answer questions like “what has happened” (e.g. decrease of margins in market A / Channel B / Customer C) and “why it has happened” (e.g. lower sales volumes, larger discounts, the attrition of a key client, or higher volume of returns) because such methods are not actionable enough,  they make limited use of external data (market parameters) and they are usually constrained by available financial data in Business Intelligence (BI) solutions and hence show limited connection with non-financial data (e.g. operational, demographical, statistical data).

What will happen & how can we make it happen?

Since decision-making is always directed at the future, insights driving decisions should shed light on what will happen next. Predictive Analytics help to provide the answer. It uses more sophisticated analytical algorithms, such as regression and correlation analysis, time series analysis or Bayesian factor analysis to explore relationships between different metrics (internal and external) and to extrapolate these insights into the future.

These insights create multiple alternative scenarios for the future. Business leaders are then left to solve the puzzle of “what is the best that can happen and what can we do to achieve it”, thus exploiting the Prescriptive nature of Analytics to draw actionable conclusions from massive data volumes. Typically, this takes the form of tackling optimization problems like:

  • Marketing spend optimization: which combination of marketing efforts should the company use to maximize Return on Investments (ROI)
  • Product mix optimization: what is the next product the customer is likely to buy
  • Sales channel and retail network optimization: deciding the optimal format by location
  • Customer segmentation for targeted marketing campaigns: e.g. based on demographics, product portfolio, transactional behavior and contact history

“The speed of insights is as important as their accuracy”

The ability to solve these optimization problems gives a company a competitive advantage – and the speed of insights is as important as their accuracy. Take the example of marketing spend optimization: delayed feedback on the effectiveness of marketing investments and a lack of agility to course-correct marketing activities result in extra costs as well as missed revenue uplift opportunities.

What is the most efficient type of marketing campaign? Should the campaign be applied market-wide or to a specific geography or category of customer? Which customer groups are more responsive to a specific type of campaign?

Responding to these questions should not be a guessing game, but a data-driven and holistic process that brings financial and non-financial perspectives together to steer decision-making. The outcome is an integrated and thorough outlook of business performance that is actionable and fact-based. The power of Analytics helps to achieve this because it allows companies to:

  • Process bigger volumes of data
  • Combine structured financial analysis with non-financial and unstructured data (demographics, sales statistics, feedback)
  • Significantly increase the number of variables (with direct and indirect relationships) involved in the model
  • Generate insights at greater speed
  • Visualize the results to represent them in an ‘easily-digestible’ and apprehensible way

Where should I start?

Based on Accenture’s experience, many organizations are still laying down the foundations of their marketing analytics capability, exploring the potential benefits, defining what their vision should be, and how to integrate this new capability into their organization. The hypothesis-based or “try-and-fail” approach is an effective way to go about it. Start small, ensure rapid data discovery to prove the hypothesis right or wrong, create a pilot to ensure the value capture, develop by iteration and then scale up.

A simple example could be a company struggling to reduce customer churn rate: what if classic analysis of product portfolio, market coverage and promotion activities failed to reveal any weak spots? Why not try with a hypothesis, and broaden the analysis to also cover transactional statistics data, for example, the number of orders placed by customers, or even the time the order was logged or processed? Such an approach could reveal correlations and provide insights on specific problems with customer order handling that may explain fewer repeat orders and a higher likelihood of losing the customer.

Discover what other companies are doing to excel in customer and market analytics in this client case study, and why not join Accenture Strategy at the upcoming CFO day on March 22 to discuss how your organization can succeed with Accelerated Learning. We’d be delighted to meet you!

Stay tuned for the next episode spotlighting supply chain analytics. Meanwhile, feel free to get in touch if you want to learn more about business performance analytics and discuss how it could support your decision-making process.

This article was co-written with Philippe Vanderschuren.