Insight, Impact, Innovation
HomeNewsroomHow to Overcome Challenges for Analytics in Emerging Markets

How to Overcome Challenges for Analytics in Emerging Markets

Apr 16, 2015

Published By : CPGmatters

By Anuj Kumar and Divya Agarwal
Growth in the consumer packaged goods (CPG) industry is largely coming from emerging markets at present. This trend will continue for the next decade or so with consumer spending expected to grow three times that of developed markets.
To grab the highest share of this growing consumer pie in these markets, CPG companies are investing huge sums of money into marketing. However, marketing managers are flummoxed. There is no right way to determine the ROI of their marketing investment – unlike developed markets – due to the lack of good data in emerging markets. Marketing Mix Modeling (MMM), used to calculate the ROI of marketing investments, is one such analysis that needs to be treated differently when working in an emerging market.
The challenge of structured and clean data in such markets is mainly due to the lack of established processes to capture data. As data forms the backbone of any analysis, any misses with its quality directly impacts the quality of the outputs.
Data quality challenges can be grouped into three broad areas: Accuracy and Granularity, Coverage, and Market Changes. An analytics exercise that is valuable for marketers cannot and should not be conducted without correct data that is representative and accounts for all of the changes within it. Let’s look about these challenges in more detail.
Accuracy and Granularity
Due to the lack of a large, modern trade presence, agencies often collect data through retail audits instead of scantrack, which is typically used in developed markets. In addition, sales channels are long and unorganized. Due to these manual processes, data collected is not as accurate, does not cover all aspects of sales (such as promotions), and is not granular enough to enable a deep understanding. This leads to fewer data points for modeling analysis and insufficient variables that are typically required to make a model robust. 
The other issue that comes up, primarily due to unorganized sales, is that data sometimes does not get collected at all. Therefore the sales numbers seen in results don’t accurately reflect the actual sales. It becomes difficult to convince business owners to conduct any analysis as results with less coverage data may not apply to all markets.
Market Changes
Most emerging markets still do not have stable economies or governments, and rapid, unanticipated changes in the market environment can impact category sales. If the outputs from the ROI measurement process cannot be overlaid with the needs of a dynamic market, the findings have limited actionability. 
However, there is a way to overcome these hurdles through three key approaches. Each approach acting as a step towards arriving at a robust, statistically significant and actionable modeling result for emerging markets. 
Call for Action
We advise using harmonization tools to combine the data through all available sources and visualize using reporting tools to see any gaps in data, identify trends, create factors to explain trends, and hypothesize. These should be then called out to the marketers to overcome any data anomalies. Similarly, the model strength can be demonstrated by applying a model run on a smaller data for larger data sets with more coverage.     
Change the Rules
Companies need to test and adopt different modeling techniques based on the challenges at hand in a particular market. For example, instead of doing standard linear regression modeling for MMM in a market with less granular data, use a probabilistic approach like Bayesian regression. The advantage of Bayesian is that is it helps to stabilize the model, even with fewer observations, identifying outliers and removing them while calculating the model estimates. It also reduces overestimation and gives more accurate coefficients, which helps to overcome the issue of missing data.
We advise using a marketing return measurement study to create a decision support tool rather than one set of decisions. The tools should be able to use the model outputs across all parameters and then allow the user to enter any planned or foreseen changes in the parameter to simulate the impact across business measures such as volume, value, and profit and % margin. This helps in feeding the changing environmental factors in the decision support tool to identify the optimal decisions which leverage the model outputs. Using a tool makes it easier and more accurate to scale and implement marketing decisions faster across more markets.
CPG companies can drive marketing decisions with far more accuracy and actionability and transform their marketing analytics in emerging markets through the use of improved data and new modeling techniques available on scalable platforms. Fractal Analytics has successfully used these approaches in the past for many CPG companies in emerging markets through the use of advanced analytics methodologies and proprietary tools.
To know more about the specific challenges and solutions to apply marketing returns measurement analytics in emerging markets, please download the whitepaper.
Anuj Kumar, Vice President at Fractal Analytics and Divya Agarwal, Senior Consultant, Fractal Analytics