---
title: Performance
description: Investigate LCP, INP, and CLS field evidence with page, environment, and deviation breakdowns.
slug: performance
---

# Investigate storefront performance

Performance uses Dozenfold field measurements captured from permitted storefront sessions. It is
not Chrome UX Report data and should not be compared as if the sampling populations were identical.

## Read metric quality

Dozenfold reports LCP, INP, and CLS using the p75 value for the selected metric and evidence window.
The current rating follows the metric thresholds shown in the product.

- Fewer than 30 eligible samples is **early evidence**.
- At least 30 eligible samples is **assessable evidence**.
- Missing evidence is shown as unavailable or collecting, not as a zero score.

## Find where performance deviates

Start with the overall distribution, then compare page groups and environment segments. Browser,
device, operating system, and viewport breakdowns show both the segment value and its deviation from
a comparable baseline when one exists. Connection type and device-memory tabs show each metric's p75
on, for example, 3G compared with 4G, or on low-memory phones. Only Chromium browsers report these,
so samples from Safari and Firefox appear as Unknown.

A large segment is not necessarily the cause of a poor metric. Prioritize segments that combine
meaningful traffic, a material deviation, and affected journey evidence.

## Connect performance to the journey

Open the affected page or anonymous sessions to see where a poor vital appeared relative to funnel,
error, and friction events. This is supporting evidence; Dozenfold does not claim a field vital
caused every abandonment.