Data Analytics behind food delivery

Customers decide within a matter of minutes whether they like your service.  Give them something timely that they value, and you got a satisfied customer. Miss the mark, and they are gone…sometimes forever.

The food and beverage industry have been undergoing rapid changes since being disrupted by food delivery upstarts.  Not that it hasn’t happened before, with pizza delivery being the disrupters during 1940s.  What changed recently is the prevalent use of data and analytics. From paper to spreadsheets to cloud computing/AI/Big Data, it is appropriate to define next gen F&B organisations as ‘Data Analytics’ companies just as McDonalds is in ‘Real Estate’.

In China, the food and service industry is worth a whopping S$1.7 trillion^[1] with the leading two food delivery companies, Meituan Dianping and Ele.me worth at least a combined S$100 billion^[2] in capitalisation.  Closer to home, FoodPanda and Deliveroo lead in the mobile delivery app space. Then again, food delivery is less about food and more about the logistics and supply chain ecosystem.  That also gives a natural edge for ride-hailing services (Grab, GO-Jek) to tap on their logistics expertise to compete in the same market.

While food delivery changed how customers are eating, a new breed of brick and mortar restaurants have been sprouting.  Based on recent trends, only 38 percent of restaurant sales are eaten inside^[3], which has given rise to eateries downsizing dining areas or even turning into strategically located virtual kitchens that cater only to food deliveries.

At the heart of both food deliveries and next gen eateries, are the data analytics and AI engines that drives operations.  From personalisation of the customer menu, accurate delivery times, food preparation efficiency and enhanced customer insights, big data is becoming the key battleground to driving customer satisfaction and survivability.

Let’s take a closer look at some of the areas in which data and analytics are driving food delivery and revamping restaurant operations:

Operational efficiency:

The last mile delivery is where customers are lost because the food has been in transit for too long.  Data from food preparation time, GPS, delivery location, and real time traffic updates assist drivers to find the most time effective route to their destinations.

For foodpanda, data analytics not only shaved 50% off their delivery time for their fleet of 3000 delivery staff, it has also powered up deliveries to an average of 3 trips per hour.

“We fully switched a few years ago to a fully automated system, which is based on an algorithm that … calculates tens of thousands of operations every second."^[4] Mr Luc Andreani, Managing Director of foodpanda Singapore.

Taking into account future-looking data points like weather forecasts, predicted traffic patterns, residential density, food trends, and driver seasonality availability trend, data analytics are making significant impact on optimising operational capabilities.  For example, Uber Eats employs meteorologists working with machine learning tools provide analytical models on the impact of rain on order trends5. Driver availability is also predicted through the analysis of driver sitting, standing, driving, or strolling behaviour.

“The more detail with which we can model the physical world, the more accurate we can be,” says Eric Gu, an engineering supervisor with Uber Eats’ information staff.^[5]

Other areas of data analytics that may be used for operational insights include:

Data-driven operational efficiency is a key competitive advantage in this digital age .

Understand customer sentiments

Data from social media helps to deepen relationships with customer. Through social listening, the act of tracking online conversations of customers, competitors and industry leaders, an organisation can get a pulse on their online brand and reputation.  Customer’s emotive social media postings can be analytically categorised into positive, negative and neutral.  As a digital business, food delivery companies need a good online standing and quickly respond to negative word of mouth or take appropriate corrective actions if required.

Customer data such as what they like, the posts they share, and reviews can also lead to a more customer centric service.  Besides discovering the sentiments of customers, social listening can extend to feedback loops for new features or even for product road mapping.

“At Uber, we use social listening on a daily basis, which allows us to understand how our users feel about the changes we’re implementing. As soon as we introduce a modification, we know which parts of it are greeted with enthusiasm, and which need more work.”  Krzysiek Radoszewski, Marketing Lead, Central & Eastern Europe at Uber.^[6]

Through analytics of conversations and trends on social platforms, social listening can help to deliver impactful insights, not just for branding, but for both internal (including staff/driver/partner stakeholders) and external side of their digital business. 

Examples of other social media analytics include:

Predicting Demand

On demand side, big data analytics look at browsing history and customer orders, both fulfilled and abandoned, to predict how many customers will order a particular dish at specific day and time, and from where.  Predictive analytics are also able to piece latent demands that may be trending but not easily observable.

Shedding light on eating patterns and trends, predictive data from food delivery is shared with restaurants to capitalise on the variety of demands and menu choices that would be ordered at a specific time.  Apart from what customers are likely to order, such predictions have helps Deliveroo’s restaurant partners know ahead of time what their customer really like or if a new opportunity is to be added for that neighbourhood.

"We've taken a look at the data of customers within this neighbourhood and adjacent neighbourhoods, to determine what are the most popular cuisines, and also which cuisines we should add to the area," said Deliveroo CEO Will Shu.^[7]

Some examples of data-driven prediction are:

How is your organisation’s data analytics journey?  By looking at the right data in combination with powerful analytics, any business can benefit from better decision making and even validate the results of those decisions.

Supplement your own data by leveraging on near real-time, anonymised data from millions of StarHub data points from demographics, psychographics and location insights.  Alternatively, talk to our social media experts on your social listening needs.  Ultimately, good data + powerful analytics = better business results.

^[1] https://www.ft.com/content/acc41678-9625-11e8-b747-fb1e803ee64e
^[2] https://asia.nikkei.com/Business/Companies/China-s-food-delivery-king-feels-the-heat-from-Alibaba; https://www.bloomberg.com/quote/3690:HK
^[3] https://decidingbydata.com/2018/03/21/data-driven-restaurant-olo-ceo-noah-glass/?utm_source=indicative&utm_medium=blog
^[4] https://www.channelnewsasia.com/news/technology/foodpanda-deliveroo-food-delivery-apps-ride-on-power-of-data-10288562
^[5] https://brand24.com/case-study/uber/
^[6] https://www.channelnewsasia.com/news/technology/foodpanda-deliveroo-food-delivery-apps-ride-on-power-of-data-10288562