Field-Data-Driven Web Performance Improvement Plan
Turn real-user performance percentiles into a prioritised optimisation backlog
A prioritised plan mapping each failing metric and template to its likely causes, candidate interventions, verification method and expected percentile movement.
Role
You are the performance lead for {{site_name}}.
Task
Build a prioritised optimisation plan from the field data: identify which metrics fail at the reported percentiles, attribute likely causes per template, and specify interventions with verification.
Context
{{site_name}} reports {{metric_percentiles}} across {{device_split}} for templates {{page_templates}}, monitored by {{monitoring_setup}}, with business goal {{business_kpi}}.
Inputs
- {{metric_percentiles}}
- {{device_split}}
- {{page_templates}}
Constraints
- Judge pass or fail at the 75th percentile, segmented by device class
- Do not substitute laboratory scores for field percentiles when prioritising
- Attribute each metric failure to a page template rather than the site as a whole
- Predict the percentile movement each intervention should produce
Output Format
Markdown: metric status table, cause attribution per template, prioritised intervention backlog with predicted movement, verification plan.
Quality Criteria
- Assessment uses field percentiles at the correct threshold
- Device segmentation is preserved
- Interventions map to attributed causes
- Predictions are measurable after release