Uber Eats: Customer Segmentation That Elevates Engagement
From batch and blast to a strategic segmentation framework
At a glance: I was one of the early marketing hires at Uber Eats, before the IPO, and I built the lifecycle framework for the Southern California market. I used SQL to define our customer segments, which fed the journeys and campaigns that supported $20M+ in average monthly sales. I still use the framework today.
The Challenge
Uber Eats launched in 2016 as the food delivery extension of Uber. Six months later, Southern California was already one of its leading markets in the world, but email was mainly a notification channel, and not a revenue driver. Every day, up to a million sends of the same email went out to Eaters (aka users). In a competitive food delivery market, marketing to everyone was like marketing to nobody.
I was one of the first marketing hires on the SoCal team (Los Angeles, San Diego, and Orange County), armed with SQL, a ton of data, and Salesforce Marketing Cloud. No segmentation framework, no inherited program, no owner.
The Approach
A thoughtful segmentation framework is what separates a needle moving lifecycle strategy from a basic email program. You have to obsess over your customer base and understand the communities that exist within it. That’s how you create personalized, valuable messaging that inspires action. Through intense data analysis in SQL, I defined Uber Eats SoCal’s core cohorts.
It starts with the top business KPI: restaurant sales. Restaurant sales grow two ways: more active customers and more spend per customer. Breaking that down another level establishes the main Eater lifecycle stages:
Not all Eaters are the same, so I stopped treating them that way. Digging through behavior data, I found the Power Eaters, our golden users: about 15 order days a month, over 80% of sales, just 20% of active users. I studied the habits and path that got them there, then used it as a blueprint to move other cohorts toward the same status. This influenced multiple focused journeys.
| Stage | Cohort | The job |
|---|---|---|
| Activate | Prospects | Convert Uber Riders to Uber Eaters |
| Activation | Lead registered users to their first orders | |
| Onboarding | Build the habit, measured by first month retention | |
| Retain | Engagement | Progress active Eaters to Power status |
| Retention | Keep Eaters ordering within the last 30 days | |
| Winback | Winback | Bring lapsed Eaters back with a return order |
Activate
- Prospects
- Convert Uber Riders to Uber Eaters
- Activation
- Lead registered users to their first orders
- Onboarding
- Build the habit, measured by first month retention
Retain
- Engagement
- Progress active Eaters to Power status
- Retention
- Keep Eaters ordering within the last 30 days
Winback
- Winback
- Bring lapsed Eaters back with a return order
Each cohort had its own goal, its own persona, and one job to do. No more one message for the whole database.
The Engine
The segmentation framework was the map to personalization. Three layers put it to work:
- Eater journeys: Each stage got timely messages built to guide Eaters toward that stage’s goal.
- Targeted offers: Promotions went to the right cohorts to prompt actions they wouldn’t take on their own, not to the whole list.
- Conversion efficiency: The right campaign reached the right Eater at the right time. Fewer unsubscribes and spam complaints, higher conversion.
The Results
- $20M+ average monthly sales supported in Southern California
- Top 5 market globally by sales
- 80%+ of sales from Power Eaters, who were 20% of users
- 150% lift in Uber rider to Uber Eats cross sell conversion over the pre framework baseline
- 6 lifecycle stages, each with its own goal and persona
What This Means for Your App
If you send the same message to your entire customer base, you’re spending real attention on noise. New customers, regulars, and lapsed ones all get treated alike, and the acquisition budget you spent to reach them leaks out the back. I build the segmentation, automation, and testing infrastructure that replaces the blast with relevance, and I can show you where to start.
Frequently Asked Questions
What is customer segmentation in lifecycle marketing?
Segmentation is grouping customers by demographics, behavior, preferences, and where they are in their journey, so you can serve marketing messages that are more relevant and personal.
Why is batch and blast a poor approach to marketing?
Batch and blast means sending the same marketing message to everyone at once and treating every customer the same. But your customers are in different places: some are loyal, some are brand new, some have drifted away, and one message can’t speak to all of them. If your marketing is irrelevant, you lose influence.
What lifecycle stages did the Uber Eats framework use?
Six cohorts, grouped into three stages: Activate (Prospects, Activation, and Onboarding), Retain (Engagement and Retention), and Winback. Each cohort had its own goal, persona, and traits.
How do I know if my business needs a segmentation framework?
If you're treating all your customers the same, you have no lifecycle strategy