The COVID crisis and the growth of digital commerce together accelerate the pace of change in Latin American business, starting with consumer needs and preferences. Companies have to move faster to implement both strategic and tactical changes to their marketing approaches. Marketing professionals are required to deliver short-term results and go beyond their traditional role of setting strategy and building brand loyalty. Today, marketers are expected to demonstrate a direct correlation between marketing spend and sales growth.
According to the 2020 Brand-Z report from Kantar, top Latin American brands such as Bradesco, Corona, Telcel, Falabella, and Mercado Libre, say that “creativity helped brands expand, even in a crisis, because marketing departments directly contributed to business performance.” The real effective growth of marketing “begins by identifying which consumer behaviors should be influenced by brands in order to increase market share today and over the long haul.”
Similarly, the role of the Insights (in-house market research) professional is also being redefined. Organizations are demanding that each market study contribute more than just observations. Research is now evaluated on the measurable business impact that its insights can deliver.
Some examples of measurably impacting insights include:
- The sales generated from recommendations provided by a research team or study
- Change in market share after research recommendations are implemented
- The cost savings stemming from specific research recommendations
Fewer Surveys — and More Analysis of Consumer Data
The pace of change in digital commerce, the preferred buying channel at present, coupled with the volatility generated by the pandemic creates a new working environment where traditional research timelines and deliverables lose tremendous value. Today, research has to be fast, focused, and quickly interpreted so that real time changes can be made to marketing decisions from pricing to messaging to channel placement. As a result, companies conduct fewer surveys, reserving infrequent surveys to annual planning schedules. Surveys, are by their very design biased because people only respond to our questions. Today, the smart money is on monitoring and analyzing customer behavior. How we act as consumers is far more truthful than how we respond to a survey.
Enter artificial intelligence (AI), machine learning and natural language processing technologies into the realm of market research. All three technologies help companies monitor, analyze, predict and yes, shape consumer behaviour. These methods can be applied to newly fielded research that is digitally collected or applied to consumer data already stored by companies, or both. Much has been written about how Netflix, Facebook and Amazon track your behaviour, adjust their messaging and product line to suit your preferences and otherwise upsell you on everything. But a lot of other traditional businesses that began selling online years ago still do not mine their own customer data for insights.
3 Key Steps
Housing reams of customer data is just the beginning. To turn that data into knowledge with the use of AI & machine learning requires three crucial steps:
#1 Define the objective. What do you want to know? You need to craft a thoughtful, detailed response to this question, because it will guide any AI/machine learning-driven data analysis.
#2 Collect the data. This means going after two specific types of information: structured and unstructured. The first type is what is available in conventional systems and formats based on rows and columns such as Excel, CSV, or customer data in a CRM (Customer Relationship Management) system such as Salesforce. The unstructured information is found in materials such as audio files (customer service complaint calls), video files, etc.
#3: Clean up and model. Once you have extracted your data, the insights you obtain from it will only be as good as your next steps. First, you have to clean your data and perform feature engineering, as well as develop the proper machine learning and deep learning models. Those models are what make sense of your data and reveal the patterns. Once you deploy the models, you then have to monitor them to see if they drift over time and reprogram them as required. It can be faster and more profitable to run the entire process in the cloud. However, getting started is the hard part. Using machine learning to spin useful insights from your unstructured data requires a steep learning curve—and that’s where an outside company experienced in this kind of data analysis can be a crucial resource.
At AMI, our senior consumer practice leaders are trained and experienced at designing customer data analysis tools based upon AI and machine learning techniques.
Contact us to learn more about how we can help you quickly obtain and analyze game-changing insights from the customer data you have in-house or you plan to research soon.