Uber eats itself

When Uber launched Uber Eats in Manhattan in March 2016, the expansion seemed to offer clear synergies: same drivers, same app, same customers. The efficiencies might seem obvious.

But new research reveals that platform companies face a unique challenge when cooking up new product lines: Their diversification strategies can create unexpected trade-offs with existing operations in ways traditional companies don’t experience.

Exploring the unique nature of diversification in platform companies like Uber is the focus of research by Christine Choi, a strategy and entrepreneurship professor at UNC Kenan-Flagler Business School, with Hyuck David Chung of the University of Illinois at Urbana Champaign and Yue Maggie Zhou of the University of Michigan.

They share their findings in “When Uber Eats Its Own Business, and Its Competitors, Too: Resource Exclusivity, Oscillation, and Cannibalization following Platform Diversification” in the Strategic Management Journal.

Their study has major implications for businesses that rely on third-party service providers — known as “complementors” — to deliver value. Unlike firms that control key capital and labor resources, platforms must navigate complex dynamics with independent service providers who make their own decisions.

Diversification in the gig economy

Traditional companies can reallocate resources — employees, equipment, facilities — to capture synergies across business lines and maximize returns. For example, a manufacturer might deploy the same production team across multiple product lines, sharing expertise and infrastructure to cut costs and boost efficiency.

Platform companies face different constraints. They don’t own their most critical resources: the drivers, sellers, developers or service providers who create customer value.

These complementors are independent actors with their own profit motives. When a platform expands into new business areas, it can’t simply reassign resources — it must convince complementors to participate while continuing to serve existing product lines.

For managers considering diversification, new products can trigger new competitive dynamics and resource challenges.

A natural experiment in Manhattan

Choi and her collaborators analyzed Uber’s Uber Eats launch, which created a natural experiment. Following the launch, restaurant adoption varied significantly across Manhattan. Using 66 zones (as defined by New York City’s Taxi and Limousine Commission), the researchers compared rideshare activity in high Uber Eats areas versus low-adoption areas.

Using trip data from the Taxi and Limousine Commission, restaurant databases and delivery platform information, the team tracked how rideshare volumes changed before and after the launch. Geographic variation in restaurants signing on to Uber Eats allowed them to measure the impact of the Uber Eats launch on drivers and consider other variables that could have influenced driver behaviors.

The findings revealed two distinct effects.

Within-platform cannibalization was significant: For every 1% of local restaurants joining Uber Eats, Uber lost 2.1% of rideshare trips in that area. Researchers estimated the decline represented 3.36 million fewer potential rideshare trips and about $12 million in lost rideshare revenue annually in Manhattan alone. However, this represents only one side of Uber’s calculation, as the company simultaneously generated new revenue from Uber Eats.

Cross-platform cannibalization was even larger. Uber’s competitor Lyft lost 6.8% of rideshare trips for every 1% increase in local Uber Eats restaurant adoption. This suggests platform diversification can create competitive ripple effects. For managers of these companies, that means that even a competitor’s move to diversify could have significant business impacts.

These effects varied by time of day. Cannibalization was much weaker during rush hours, when rideshare demand and pricing peaked. This timing pattern also confirmed drivers weren’t permanently abandoning rideshare for food delivery but instead deciding to allocate their time and vehicles based on relative opportunities.

The mechanism: resource oscillation

The key insight lies in understanding how complementors behave when presented with new opportunities. Rather than permanently switching from rideshare to food delivery, drivers engaged in what Choi and her colleagues dubbed “sharing-enabled resource oscillation” — moving back and forth between business lines to maximize resource use.

Drivers could share certain resources across both businesses: driving skills, city knowledge and idle time between rides. However, other critical resources — their vehicle and active working time — remained “exclusive use,” meaning they could only serve one purpose at a time. When drivers accepted food delivery orders, that vehicle and time became unavailable for ridesharing.

This differs fundamentally from traditional firms, where managers control resource allocation. Platform companies typically enforce exclusivity only at the transaction level — you can’t deliver food and transport passengers simultaneously — but can’t control organizational-level decisions, such as driving for Uber versus Lyft or delivering food versus delivering passengers.

Implications for platform managers

These findings offer important lessons for platform executives:

• Recognize diversification trade-offs: Expansion might cannibalize existing operations, particularly during low-demand periods when complementors have more flexibility to switch between opportunities. Factor this effect into business cases and growth projections.
• Understand competitive dynamics: Platform diversification can affect competitors. Managers should consider how expansion decisions might influence competitive dynamics and create advantages or threats to their own companies.
• Design for retention: Consider strategies to retain complementors in core business, such as increasing compensation during peak periods or reducing coordination costs between business lines. The research suggests “super apps” might help by making it easier for complementors to switch between services within the same platform rather than across competing platforms.
• Time expansion strategically: Given that cannibalization effects were weaker during high-demand periods, consider launching new services during traditionally slower periods or in markets with excess complementor capacity.

Rethinking platform strategy

The Uber Eats research highlights a fundamental truth about platform business models: Success requires managing relationships with individuals and businesses who make their own decisions. Unlike traditional diversification, platform expansion creates ongoing competition for complementor attention and resources.

This means managers must consider the incentives, constraints and decision-making of complementors. The most successful platform strategies will align complementor interests with platform objectives. In the platform economy, your most important assets make independent choices about deploying their time and resources — and those choices directly impact your business outcomes.