The opportunity graph and product map (Early Access)
This feature is in Early Access. During this time, aspects of the functionality may still be developed, and this documentation may not always be up to date. If you have any questions, contact Amplitude Support.
Opportunities in Amplitude aren't a flat list. They live in an opportunity graph: a connected model of your product work that links each Opportunity to the evidence behind it, the metrics it targets, the code that delivers it, and the screens and flows where it happens.
The graph is what lets the system reason about how everything relates: which opportunities target the same metric, which screen a drop-off happens on, and which pull request shipped a given change.
Why a graph instead of a list
A flat backlog loses context. A graph keeps it:
- Traceability: Follow an Opportunity to its evidence, its target metric, the agent that implemented it, and the pull request that shipped it.
- De-duplication and structure: Related opportunities link together, and larger themes can parent smaller pieces of work.
- Measurement: Because opportunities connect to metrics and to the product map, teams can tie shipped work back to the outcome it moved.
Relations
Opportunities connect to other entities through typed relations. Common ones include:
| Relation | Connects an Opportunity to | Meaning |
|---|---|---|
| Targets metric | A metric | The metric this Opportunity aims to move. |
| Supported by | Evidence (charts, dashboards) | The data backing the Opportunity. |
| Delivered via | A GitHub pull request | The PR that implements the work. |
| Implemented by | An agent or a person | Who did the work. |
| Intervenes on | A product node or edge | The screen or flow the change affects. |
| Related to / Blocks / Parent of | Another Opportunity | How opportunities relate, depend on, or roll up into one another. |
The product map
Part of the graph is a product map: a model of your product as users actually move through it. It's made of two building blocks:
- Product nodes: meaningful places in your product, such as a screen, a milestone, an external trigger, or an exit point.
- Product edges: the transitions between nodes, such as one screen leading to another, a flow being abandoned, or a user re-entering.
Nodes and edges carry health signals, so the map highlights where users drop off or hit friction. Because opportunities link to nodes and edges through the intervenes on relation, the map shows not just where problems are but which opportunities address them.
How the map is built
Your analytics and session data ground the product map, so it reflects how the product actually behaves rather than an idealized flow. As discovery runs, it ties findings to the relevant nodes and edges, connecting a drop-off on a specific transition to the opportunities that target it and to the metrics that measure it.
To act on an opportunity surfaced through the graph or map, go to Acting on opportunities.Was this helpful?