Why segment-specific benchmarks matter
A 1.6% conversion rate is unremarkable for fashion & apparel and concerning for coffee & tea. Comparing every store to one blended e-commerce average is the single most common way benchmarking goes wrong.
Illustrative ranges by segment
Coffee & tea: conversion rate around 2.0%, blended ROAS around 3.2x, email revenue share around 18%. Subscription dynamics push retention metrics higher than average.
Fashion & apparel: conversion rate around 1.6%, cart abandonment around 74%, average order value around 550 kr. Size and fit friction drives abandonment higher than other categories.
These are illustrative starting priors, not guarantees. Your own connected data always supersedes a segment default once you have enough volume to be statistically meaningful.
How these numbers are used
Segment priors are used for Bayesian shrinkage when your own data volume is thin, and as a relative benchmark once you have enough history. They live in versioned configuration, not in a spreadsheet someone forgot to update, and they're recalibrated as real cohort data accumulates.