AI can solve your online retail woes

AI use is commonplace in multiple industries, including retail, especially for digitalized retailers who provide e-Commerce.

E-CommerceDigital retail today is an omnichannel experience that allows consumers to easily maneuver and engage with a business’s platform. This, in turn, provides the information and education to facilitate purchasing decisions. 

In the second part of an interview with Andrew Martin, head of AI & data company Databricks, Tech Wire Asia sought to understand more about the challenges digital retailers face, and how SMEs can address them with data and AI. 

What are the challenges that digital retailers commonly face? 

While expectations surrounding the customer experience have certainly risen to new heights, few retailers have invested in the right technology for meeting these new standards. 

From our research, we’ve identified four customer challenges SEA retailers face. They include fraud, delivery theft, returns, and customer service. 

These aren’t only common, but significantly impact their bottom lines. 

So how can data and AI address these four challenges?

Fraud

Fraud has become too commonplace — SEA stands to lose US$260 million annually to online fraud, with Indonesia, Thailand and Vietnam expected to be the most heavily affected.

Losses associated with fraud soared to US$56 billion in 2020 globally and accompanied a huge dip in customer confidence in the brand. 

Data and AI can help retailers get ahead of fraud and avoid financial and reputational damages, especially when it comes to proactive approaches. 

A modern data architecture that brings data together from geospatial data to sales trends, across the business can effectively enable anomaly detection at a massive scale to protect losses caused by fraud in real-time with machine learning.

Delivery Theft

Logistics costs are high in most SEA nations due to geographical challenges, the imbalanced concentration of economic activities, and poor connectivity between various parts of the countries. 

The lack of adequate shipping infrastructures will make it difficult to deliver parcels within the promised delivery timeframe. This increases the risk of package theft, which is a significant operational burden on retailers, with global estimates of over a million packages being stolen or lost daily. 

Data and analytics can help logistics providers identify common sites of traffic accidents or package thefts and design their services around those. 

The global shipping industry is also using AI to enhance security measures, both within and outside of business grounds, with shipping carriers using drones to patrol the grounds around their warehouses to collect real-time information and data. Data and AI allow retailers to easily identify such hotspots and frame apt responses.

Returns

Without the tactile experience of brick-and-mortar shops, consumers are returning their online purchases at an alarming rate. According to industry data, at least 30 percent of all products ordered online are returned compared to only roughly 9 percent bought in bricks-and-mortar shops. 

The ability to uncover trends in that data with the power of machine learning allows retailers to better understand customer behaviors, spot high-return items and take action with data to minimize returns. 

To any e-commerce retailer, returned products mean additional shipping costs, which can constitute a significant portion of any retailer’s operating margin. Retailers can also incorporate predictive analytics using data and AI in their returns and reverse-logistics operations, to improve service levels with fewer queries and reported issues.

Customer Service

Customer satisfaction, customer retention, and cost to serve are three factors that can define the long-term profitability for retailers. Although customer issues can be wide-ranging, many issues will be common to reoccur amongst customers. 

For example, Natural Language Processing (NLP) tools can be used to quickly analyze service call notes and easily identify the straightforward and most common issues. These can be tackled by blending digital and call center channels and driving self-service usage for common queries. 

Finally, businesses can arm their customer care teams with visual data snapshots of helpful customer insights – an at-a-glance view of the context, key data points in their history, and next-best-step suggestions while the customer is on the call. This is especially useful while dealing with an at-risk customer identified by machine learning (ML) models who have been routed to the retention specialist.

Are there other challenges that SEA retail SMEs face today? 

Across SEA, countries are still battling uneven vaccination rates and are still dealing with outbreaks and lockdowns. While regional online retail has skyrocketed, SEA economies continue to rank poorly in being able to prepare and adapt to the pandemic’s changing conditions. 

When looking at the retail industry holistically, one needs to be mindful that logistics, supply chain and last-mile delivery are also as critical as the products and services sold and the user experiences provided. With SEA’s retail and consumer goods markets in flux, accurate forecasting that considers variations in day-to-day product demand and distribution will be essential.

Accounting for these shifting market conditions is often well beyond the capabilities of legacy, data warehousing-based tools. The retail industry and related organizations working with ever-growing, day-to-day digital data will need a centralized hub — a logistical control tower of sorts — to orchestrate the technology, tools, and processes used to capture data across all stages of the supply chain. 

The demand for more granular, timely forecasting, can be met with solutions that employ data insights derived from machine learning. These can power the retail industry to generate forecasts that move away from traditional linear models and historical-based algorithms, towards flexible inventory planning that can be precisely adjusted on an individual day and store level.

---

---

---

---