Earlier this week we had the pleasure of hosting a focus group with retail leaders from across Asia, together with SAS, Intel, CT Global, who had their industry experts joining us all the way from the US.
Prior to the roundtable, we did a poll among the industry and found that 100% retail leaders have seen changes in customers purchasing behavior since the start of 2020.
This change is reflected in multiple ways:
Physical to Digital Engagement
An example given by Patrick from NTUC Fairprice Food Services: online delivery sales from their food courts and coffee shops contributed 30% of total revenue during the peak of the lockdown in Singapore, an increase from pre-lockdown levels of 5% contribution. Although there was a dip when social distancing eased, it remained above pre-lockdown levels at roughly 10%.
Kai Ming from DFS recalled how it was double whammy when travel grinded to a halt and retail stores had to be shut. DFS’s largest customer base was from China and through this period they engaged their customers digitally by running live streaming events on wechat to continuously engage them. Kai Ming noted that there was pent up demand in the market and it was not a matter of demand levels falling off, but it was a matter of how consumers are going to satisfy their demand and the availability of channels that can help fulfill this demand.
Similarly, Syafiq from Sephora, shared how the pandemic hit just after Sephora opened their largest flagship store in Malaysia earlier in January 2020 and their stores had to be closed to comply with the Movement Control Order (MCO). Even after the MCO was lifted, no testing of cosmetic products was allowed in store. Sephora moved to provide consultations digitally, where customers were given recommendations based on the preferences and the types of products they were looking for. Sephora found that customers trusted the recommendations and were making purchases based on these recommendations.
Separately, Jardine Cycle and Carriage, an automotive group, saw a dip in demand for new cars. Jodi spoke about the need to pivot online and grow their digital engagement with customers: “typically when you buy a car, you want to touch it and open the door, but now it’s all about engaging the customers online. The shift that we see is the willingness of the consumer in this market wanting to be engaged digitally – through WhatsApp or different channels online.” Having rolled out multiple engagement channels online, the next step for them is to understand how customer behaviour has changed, and how they can re-segment and re-target their marketing efforts.
Shifts in Demand
Both NTUC Enterprise and MYDIN shared similar observations. Over the past few months, there has been a sales spike in food products/ fresh produce. For MYDIN, a family run business, it was important to validate the kind of promotions to run using data. In the face of changing customer requirements, Alwan from MYDIN pointed out that they had to use data to make strategic decisions on the kind of products to focus on to balance the sales of goods with high and low margins. These decisions are validated by both walking the ground in stores and sales data.
The different panelists highlighted how engagement has shifted digitally. Demand sensing now comes from both online and offline channels. For Sandeep who leads the business analytics of NTUC Enterprise, the challenge was mainly on combining both online and offline data using machine learning to make better predictions.
One key question raised during the discussion: How are retailers supposed to treat the customer behavior data during the COVID-19 period? Should it be treated as an anomaly / outlier?
Charles Chase, Industry Consultant from SAS addressed this by explaining that in addressing the spike during COVID-19, short term demand sensing is key. “Data should be collected and monitored on a real time/ daily basis – with this you would be able to see the demand patterns almost immediately.” Leveraging machine learning and artificial intelligence, retailers can be more agile in an unstable marketplace by more accurately predicting short-term shifts in demand patterns, quickly reforecasting the mix within the average market basket.
In modeling the return to recovery, sentiment and economic indicators and other external variables can then be incorporated to cleanse abnormal demand patterns out of the demand history to reflect normal demand patterns.
Advanced Analytics to Model, Predict and Adapt to Changing Consumer Demand Patterns from Demand Planning to Effective Targeted Marketing
In conclusion, Dan Mitchell, Business Director, Global Retail/CPG from SAS summed it up beyond the models and forecasting techniques. The real take away from COVID-19 is the innovativeness in which enterprises have approached their challenges to understand their data.
“Just like what Monica (who led the Far Eastern Group’s digital transformation) demonstrated, the most important thing is the skills that we acquired through the whole process, the agility to add in extra data – test, try, refit – and even, go to other parts of the company to get data that can be used to enrich the current model.”
We will continue our discussions on understanding how businesses are prioritizing their technology investments across Southeast Asia. Drop us a note at email@example.com if you are interested.
16 September 2020