Table of Contents
As published in CMSWire. Read the full article here.
On June 8, 2020, The National Bureau of Economic Research officially declared the U.S was in a recession. The pandemic pushed companies to adjust their business models and the way they reached out to and interacted with customers.
The lack of personal interaction presented many challenges. With the shift to digital and online channels, companies needed data to understand customers. Organizations leaned heavily on analytics and machine learning to try and predict customer behavior, buy or churn propensity, and tackle the growing rate of fraud.
Fintech and digital-first companies had a significant opportunity to capture the market share during all this. Data-driven decisions were the key to quickly identifying and processing the needs of customers. But there was a challenge: the data companies had in-house was not very useful.
What Happens When Your Data Is Irrelevant?
Think about it. Data teams had created and trained all their analytics and machine learning models based on specific data sets for certain market conditions. Then the pandemic changed the rule book. The B2B credit models and loan underwriting models weren’t working any more. Businesses were closed for months, and when they opened, foot traffic was scant.
Companies had to scout for alternate data sources and build new models to make smart decisions about lead generation, demand forecasting or credit and fraud mitigation in this new age. Incorporating external, alternative or third-party data became a critical step in making these decisions. For instance, fintech companies processing small and medium business loans started to leverage data signals such as online reviews, web presence, online payment capabilities, web traffic, delivery services to determine the creditworthiness of the businesses.
As we start to see some light at the end of the tunnel, expectations are rising the recession will be V-shaped — meaning a quick recovery following the steep decline. That would be great news for the economy, for employment and for consumers wanting to resume their favorite activities. However, the rebound will present similar challenges to the shutdown, making data teams question whether they have the correct data for accurate predictions. Data scientists may feel a sense of déjà vu: the models and information they have been working with last year will suddenly not be as relevant. They will need alternate data sources to improve the performance of their models.
Read the full article here.