When we talk about machine learning, we’re talking about a direct application of artificial intelligence that allows a system to learn from experience rather than instruction. In the modern digital economy, it’s a vital tool for providing a first-class user experience.
In the on-demand delivery space, companies are putting machine learning and predictive analysis to use to track trends, pre-empt requests and enhance the last-mile delivery service. Our ‘always-on’ culture means that customers expect rapid and relevant user experiences. Today, delivery companies have to be agile and responsive in order to retain customers. One glitch, one delay and customers could take their business elsewhere.
If we consider the likes of Amazon and Ocado, they’ve set new standards and become the benchmark for quality customer service. Users are now realising that they also have the right to expect a similar experience when it comes to product delivery. And rightly so.
A recent report from Salesforce revealed that an extraordinary customer experience raises the bar for customer engagement - but what is the definition of an ‘extraordinary’ experience and how can we provide it?
Some ways in which machine learning is positively influencing the customer experience in delivery, are:
Tracking behaviour to understand your customer
With the ability to predict consumer buying behaviour, companies are able to ensure they have the appropriate stock to fulfil customer orders. The number of couriers needed for orders can also be forecasted thanks to recorded data, alleviating any unnecessary wait times for customers.
By tracking behaviour through specific algorithms, we have the ability to recognise order trends through time. For example, Wednesday is the most popular day to order groceries, while Sunday lunchtime is when Barcelona is craving “Pollo Asado” (roasted chicken). It’s this type of data that allows us to better understand both our customers and local trends.
Taking into account the culture of a city is vital when it comes to identifying these behaviours while balancing individual and collective needs. By allowing machine learning to examine all aspects of behaviour patterns, businesses can build a clear image of the customer they are dealing with and how to adapt the CX to their needs.
Building loyalty to avoid customer churn
Building loyalty, especially through an app, can be a difficult task as there is limited human interaction - it’s easier for customers to disconnect from an app than say a local corner shop where they build a relationship with the person running the business. By using the predictive capabilities of machine learning, businesses can quickly acknowledge and pinpoint customers that have become disengaged with a service. From here, the technology can sift through the customer’s past order history and suggest the appropriate course to reel the customer back, such as compensation or free delivery.
With technology, retailers are able to improve their customer retention by using a platform to remind customers to reorder or share special deals. For example, pizza delivery company Slice is using technology to help local pizzerias compete with the giant mobile order deliveries such as Dominoes and Pizza Hut. Their system automatically leverages data to consumers who haven’t come back to order, something not previously possible when making an order in person or over the phone.
At Glovo, we use machine learning (ML) to predict the future lifetime value for each customer which is the total net income we can expect to receive from them. By using a second ML model to predict the probability of a customer remaining active on the app, we can use the lifetime value data to dynamically adapt the compensation offered to these customers, hopefully keeping them engaged.
Identifying external factors to guarantee fast delivery
Matching delivery time with customer convenience has always been a challenge in last-mile delivery. But as the industry continues to innovate, it’s not just about getting products to customers faster than competitors, we need to also factor in problem-solving and offer a tailored approach. In this on-demand world, where almost anything is available at the swipe of a finger, people want their products delivered on tap and to a certain standard - meaning fast is never just enough.
The average consumer is time-poor and ultimately seeks convenience. We only have to look at the recent research from Barclaycard on click-and-collect delivery services being too much hassle for customers, as they just don’t have time to pick-up items they’ve ordered. Online shoppers failed to collect £228m worth of goods.
This is where the on-demand delivery experience comes into its own. When we think of time-efficient delivery experience, order tracking comes to mind, but this is just scratching the surface of an extremely complex system in place. Machine learning algorithms are used to also calculate the preparation time a store will need (based on factors such as store type, order contents and time of day), match the order to the best-placed courier and then synchronise the orders preparation time with the courier's arrival - all in real-time.
Delivery companies also have to consider each city’s geographic layout, traffic and demographic distribution. During peak hours in cities, there will be tens of thousands of couriers delivering products at any one time, from food to groceries, to forgotten items such as keys. Technology enables these services to process hundreds of orders per minute, and most importantly ensure they arrive with the customer on time.
Looking to the future
The development and improvement of customer experience is a never-ending task. Combined with the on-demand delivery industry continuing to grow and innovate, technology plays a key role in automating services to make the customer experience as streamlined and easy to navigate as possible.
In the past year in the US alone, online grocery shopping has increased by 35 million shoppers, proving that on-demand delivery is not just a trend but a new wave of consumer behaviour. By harnessing the power of machine learning we can take huge steps into automating and streamlining the delivery process even further. Our customers are changing and so must we.
Bartek Kunowski, VP of Product, Glovo