FoodRocket is a dynamic food delivery app for fast grocery delivery within 15 minutes of the order’s placement. FoodRocket relies on AI technologies to manage warehouse stocks, forecast demand and optimize delivery time and guarantee the lowest delivery costs for their users.FoodRocket needed to optimize their communication strategy and boost customer engagement. This case study outlines the challenges faced by FoodRocket before integrating NGrow, the testing goals, and the impressive results achieved through AI-powered push notifications.
The Problem 🧐
Prior to partnering with NGrow, FoodRocket relied on a mix of basic OneSignal push notifications and trigger-based communication for promotional offers and updates on new stores/restaurants.However, the process posed multiple gaping problems:
- The process was primarily manual
- Lacked ROI, transparency and systematic metric measurement systems
- Repetitive messaging with little to no variation leading to a steady decline in the efficiency of marketing campaigns
This led to major complications starting with the absence of fresh push content which resulted in decreasing app opens from push and customer interactions, lack of a system that tracked unsubscribed rates, and so on.
Enter NGrow! 😎
NGrow integrated seamlessly with FoodRocket's existing mobile analytics and here’s how things started moving upwards. FoodRocket had Amplitude mobile analytics and Firebase Cloud Messaging in place which made integrating NGrow very easy and fast. The no-SDK setup was up and ready for use in a matter of days as against the normal timeframe that takes months with other tools.
This enabled FoodRocket to quickly progress to planning their future push campaigns. Before launch, we also dedicated our Customer Success & Analytics team to additionally research & evaluate FoodRocket’s communication strategy:
- By analyzing user data (including purchases, product views, and searches) NGrow developed individual CRM strategies for different customer groups
- After validating the settings for each user group, the CS team created multiple versions of highly-targeted push texts, tailored to each particular engagement campaign
- Lastly, we identified the scope & expected outcomes of the testing. Things we tested included churn prediction, product category affinity, RFM (Recency, Frequency, Monetary) segmentation and abundant clicks/searches/cart/checkout to enhance user engagement
So, what did we target and what did we measure? 🔍
The key target metrics for FoodRocket were order growth and Gross Merchandise Volume (GMV).
The team aimed to assess the impact of push campaigns both individually and in conjunction with email campaigns.
To measure incremental uplift of push notifications, we used “local” control groups for each push campaign, alongside a “global” control group to see the overall picture for all campaigns in scope. Control groups allow tracking both positive (like orders, GMV, ROI) and negative metrics (bounces, uninstalls, bounce sessions).
The Results? Magnificent! 🔥
3x Marketing Campaigns Effectiveness Improvement. By optimising push campaigns with AI, NGrow streamlined user communication and identified inefficient practices:
- Optimised the time of delivery of communication to the user
- Found optimal frequency of communication for different user groups
- Predicted potential user churn & helped to decrease the number of uninstalls
Enhanced Engagement: NGrow's AI-powered text generator along with the multi-armed bandit algorithm allowed for dynamical, simultaneous testing of hundreds of hypothesis. This enabled NGrow to continuously test newly created and refreshed push content along with new promo campaigns, resulting in increased app opens and engagement rates.
Optimized Communication: we took an "auditor" approach and scrutinized every aspect of FoodRocket's user communication. This led to a reduction in push frequency for active users and the incorporation of upsell mechanics, resulting in a more engaging and less intrusive user experience.To elaborate, the NGrow audit enabled FoodRocket to take actionable decisions such as
- Reducing the amount of communication for active users, because they were receiving it too often (which had had a negative impact)
- Tested new upsell mechanics & promos, improving ARPU
- Targeted inactive users who still had the app installed to reactivate them again
ROI Measurement: NGrow provided detailed insights into the performance of each push campaign, enabling FoodRocket to track ROI, unsubscribe rates, and other vital metrics. In addition to the built in-reporting directly in the platform, NGrow provided a dedicated account and analytics team to FoodRocket to ensure they succeed.The analysts helped to prepare regular reports so FoodRocket could track incremental campaign influence reports, as well as shared data in Amplitude ensuring a 100% transparency in the exercise.
TL;DR in numbers
+33% D21 Cumulative retention uplift (Daily Retention)
+36% 14-week cumulative retention uplift (Weekly retention)
+22% iOS and +29% Android devices Funnel conversions (intent conversion into a purchase)