At Delivery Hero we are committed to invest in areas where automation, AI and advanced analytics can boost our business performance. One of the key applications for these techniques is marketing – for example we can automate the creation and execution of many of our marketing initiatives and forecast the impact of marketing campaigns, allowing our marketing teams to spend their time and money in the most efficient way.
The Delivery Hero Marketing Team’s main goal is to promote our global brands by highlighting our unique selling points – variety, speed, affordability and our seamless ordering experience. If we’re successful in creating demand, new users will order. Once new users place their first order, our communication evolves – meaning, we start segmenting our customers to contextualise and personalise our communication with the ultimate objective of providing an amazing delivery experience.
To achieve these objectives, the Marketing Tech Team supports our Marketing Teams in the following ways:
- Integrating customer data to leverage data signals and decide what is the best content to send to a given customer for a given time and channel;
- Translating data insights into improved decision making and marketing performance by creating technology that allows us to better assess our marketing campaign’s effectiveness and optimise live campaigns;
- Creating a marketing management platform that enables setting up marketing campaigns across different channels and communication types – imagery, video, text ads, etc;
Improving customer relationships
One problem our business deals with is negative user experience caused by various operational constraints such as platform, logistics or restaurant issues. In these unfortunate cases, customers have their orders cancelled. This has some obvious negative effects on the business:
- The probability that the customers may never return increases.
- Contacts with customer care increase, which impacts operational costs.
The solution
One of the key marketing functions is to manage the relationship with our customers (CRM). So, if a customer has an unsatisfactory experience such as a cancelled order, it is crucial to engage with him/her as quickly as possible, showing them we care and are determined to find a solution tailored to the issue they’re facing.
In order to create an automated system, we started by evaluating the signal of what happened to the customer – a cancelled order. We then built a system that processes these signals and decides, in real-time, what communication, and possible compensation, we can send to a customer based on the different data we have available at the time.
Below you can see a diagram of how the system works:
System Architecture & Technologies
In order to create this automated process, we use a lightweight Django web application hosted on Appengine, which receives details about cancellations that happen from another service (via Amazon SNS subscription). Then, it generates asynchronous events processed with Cloud Tasks which will analyse the data and extract information necessary for the next steps.
For our decisioning system, we use Cloud Functions, while the data we process is stored along the way in PostgreSQL (hosted with Cloud SQL) and BigQuery. The results of the processing are then pushed to various external services in order to credit and notify the user of their compensation.
Finally, we use Google DataStudio dashboards to visualise statistics and track performance. We decided to make use of the tools in the Google ecosystem since they work well with one another, and provide a flawless integration and increased scalability.
Future strategy
We are aware of other incidents that create a negative order experience for our customers impacting their relationship with our brands. Examples are unexpected delays or unsatisfactory items flagged post delivery. We will always strive to optimise from the operational side, to minimise these events occurrence, but we can also proactively reach out to our customers and show them we care to provide the best possible customer experience. Hence our customers will have a positive and quick solution to their problem and order with us again in the future.
If you’re interested in creating these solutions for our customers among other interesting projects within Delivery Hero’s Marketing Tech, we’re hiring!
