At its essence, a product is a combination of different funnels bringing the user from one state to another. How many funnels does your startup’s product have? How many are you measuring? How many are you optimizing?

Uber’s consumer experience has two funnels. The first is user on boarding: registering a user and collecting their payment information. The second is booking transportation from opening the application to rating a driver’s performance.

Slack has many more. Onboarding users and the groups they belong to. Configuring administrative features like group owner. Integrating bots. Creating groups. Learning the / syntax for Slack commands. Joining additional domains. Creating profiles. Upgrade the account to paid. The list goes on.

One of the key responsibilities of a product manager is developing these funnels, measuring them, and optimizing them. A friend working at a publicly traded software company framed product development this way. At his company, before product managers and their teams write a single line of code, they have already defined the conversion funnel and every possible state for user to be in. And, they set initial expectations for the conversion rate through each step.

This discipline accomplishes a four goals. First, the product team must be comprehensive; they must understand and plan for all the different pathways a user will explore in the product. What happens if a user aborts the sign-up flow when asked for payment information? What happens if the user doesn’t appear for their UberX?

Second, the product measures user performance from the outset. After all, the product is built, the engineering team knows precisely how to instrument it. It captures conversion funnel data from the very first user.

Third, the documented expectations inform the team’s focus when optimizing the product, even if these initial expectations are guesses. Most often when releasing a product, there are one or two different key problems in the conversion flow. When user data conversion rates deviate meaningfully from the team’s expected conversion rates, the team knows exactly which step in the process to reevaluate.

Fourth, as teams optimize each funnel locally, these efforts combine to create product-wide improvements in user flow, satisfaction and, eventually, revenue.

When a startup is small with just a few people working on product, teams tend to focus on just one funnel. But as the product grows in footprint and complexity in order to serve more the needs of the product, incorporating these conversion funnels as part of product development lifecycles from definition to evaluation helps product teams to continuously improve their product.