超越 A/B Testing：使用 Surrogacy 和 Region-Splits 衡量市场中的长期效果

Image generated with Gemini 3 Pro (Google), 2026.

使用 Gemini 3 Pro (Google) 生成的图像，2026。

Written by Amber Wang and Yoonji Kim at Lyft.

Background

背景

Whenever you use the Lyft app, there is a complex balancing act happening behind the scenes. Various levers are used to keep the marketplace running smoothly; Base prices and coupons for riders affect demand, while driver pay and bonuses impact the level of available supply. Since every change to prices and payments impacts Lyft’s costs and revenue, they lead to key optimization problems, such as:

每当你使用 Lyft 应用时，幕后都有一个复杂的平衡行为在发生。各种杠杆用于保持市场平稳运行；乘客的基本价格和优惠券影响需求，而司机的薪酬和奖金影响可用供给水平。由于每次价格和支付的变动都会影响 Lyft 的成本和收入，它们导致了关键的优化问题，例如：

How should we allocate budget between driver incentives and rider incentives?
我们应该如何在 driver incentives 和 rider incentives 之间分配预算？
How do we invest resources to achieve x% rides growth, and how much does it cost in terms of short term profit?
我们如何投资资源来实现 x% 乘车增长，以及这在短期利润方面的成本是多少？

These are the questions the Foundational Models team at Lyft tries to answer in a systematic way. A key ingredient is understanding the effects of different types of investments — for instance, what will happen if we increase the total budget for driver incentives by x%? What will happen if we increase the rider price of all rides by y%? It’s worth noting that the long term effects of such decisions tend to dominate the short term effects: we may earn more short term profit from a ride if we charge riders more and pay drivers less, but lose riders and drivers in the long run.

这些是 Lyft 的 Foundational Models 团队试图以系统方式回答的问题。一个关键要素是理解不同类型投资的效果——例如，如果我们将司机激励的总预算增加 x%，会发生什么？如果我们将所有行程的乘客价格增加 y%，会发生什么？值得注意的是，此类决策的长期效果往往主导短期效果：如果我们向乘客收取更多费用并向司机支付更少，我们可能从一单行程中赚取更多短期利润，但长期会失去乘客和司机。

Estimating the long term effects of resource allocation decisions is challenging in a multi-sided marketplace such as Lyft. Because these decisions tend to be consequential, their effects go beyond first order effects on directly affected users. For example, if we increase dri...