This article describes different types of optimization metrics and how to use them to achieve your business goals.
Table of Contents
- AI Decision-Making with Optimization Metrics
- Optimizing conversion rate
- Optimizing a numeric conversion value
In the world of optimization and data-driven decision-making, it's crucial to grasp the concept of conversion and value-based performance. These are key metrics that help us evaluate the effectiveness of our projects and make informed decisions. This article aims to clarify the distinction between these two metrics and shed light on their importance.
AI Decision-Making with Optimization Metrics
The Optimization Metric is the parameter used to measure the audience's performance selected in a project. It serves as the compass guiding the AI's decision-making. This metric can be a specific visitor behavior (event) or a scalar value. The system relies on this metric to decide which ideas should be shown to a larger audience.
Increasing vs. Decreasing an Optimization Metric
It's important to note that the system can only increase an Optimization Metric. This means that you can only optimize towards goals like "Increase the number of videos watched per session" but cannot optimize towards decreasing a conversion event or value, such as "Decrease the number of subscription cancellations."
Optimizing conversion rate
Conversion-based optimization projects use the Conversion Event to measure the success of experimental ideas.
The conversion event is a specific user behavior you want the system to optimize. A conversion event is measured once per session and is a fundamental aspect of your optimization strategy.
Here are some examples of conversion events:
- Lead Confirmed
- Engaged
- Added Item to Cart
- Reached Thank You Page
Optimizing a numeric conversion value
Value-based optimization projects use the Conversion Event and a Conversion Value to measure the success of experimental ideas.
The conversion value is a scalar quantity you want the system to optimize. When optimizing a value, you must also identify a conversion event. This value is calculated for each session where the specified conversion event has occurred.
Here are examples of Optimization Metrics that include a conversion value:
- Video Minutes Watched (value) when a video is played (event)
- Lead Score when a lead form submission is confirmed
- Cart Value when a visitor reaches the cart page
- Order Value when a visitor reaches the thank you page
- Funnel progression value when a visitor makes progress down the funnel (a valuation of the distance a user goes in the funnel)
When optimizing a conversion value, the system makes decisions based on the average value of a session. This means that you may increase either or both the number of conversions and the value of a session.
For example:
- The average session value can be improved by increasing the number of sessions that fire the conversion event.
- The average session value can be improved by increasing the value of more sessions.
Learn more: Optimizing the value of context attributes
Consider financial Impact
It's important to consider using a metric with a clear financial value attributed to it. Understanding the financial impact of your experimentation is often essential for decision-makers. When a metric lacks a clear financial value or impact on the bottom line, it can be challenging to communicate its significance, especially when financial reporting data is inaccessible to the optimization team.