The Shopify analytics metrics that lie to you (and what to look at instead)

Conversion rate, average order value, sessions — the headline metrics in Shopify analytics can hide more than they reveal. Here's the analytical hygiene we apply.

Shopify’s analytics dashboard puts five numbers front and centre: total sales, sessions, conversion rate, AOV, returning customer rate. Every one of them can mislead you if you read it wrong. Here’s how to read them properly — and what to look at instead.

Conversion rate

The lie: “Our conversion rate is 2.1%, industry average is 2.5%, we have a problem.”

The truth: conversion rate is meaningless without traffic context. A store with 100% organic-search traffic from buying-intent keywords should convert at 5%+. A store with 100% top-of-funnel paid social traffic might be doing brilliantly at 1.5%. The “industry average” is averaged across both, so it tells you nothing about your mix.

Look at instead:

  • Conversion rate by traffic source
  • Conversion rate trend over 90 days (is it improving?)
  • Conversion rate by mobile vs. desktop (mobile is usually 0.5× desktop — alarming if much worse)
  • Conversion rate of returning visitors specifically (this should be 4–8% for any decent store)

Average order value

The lie: AOV “increased” 12% — celebrate the upsell strategy.

The truth: AOV can rise because you actually upsold customers, OR because lower-AOV customers stopped buying. Same number, opposite stories. The first is good; the second is collapse disguised as improvement.

Look at instead:

  • AOV alongside total order count (both up = real win; AOV up + orders down = something’s wrong)
  • Distribution, not average. Median order value tells you more about typical customer behaviour. A handful of $2,000 wholesale orders can lift the mean while typical customers spend $40.
  • AOV by acquisition channel (paid vs. organic vs. email)

Sessions

The lie: “Traffic is up 20%.”

The truth: sessions can spike from bot traffic, scraping, ad fraud, or mass-emailing the same list twice. Useful sessions are qualified sessions — humans, with intent.

Look at instead:

  • Sessions filtered to your real audience (exclude internal IPs, bot user-agents)
  • Engaged sessions (sessions > 30s, or with multiple page views)
  • Sessions by traffic source (organic up = SEO win; paid up = budget increase)

Returning customer rate

The lie: “Returning customer rate is 30%, customers love us.”

The truth: this metric counts whether a customer has made multiple orders, not how often or how recently. A customer who bought twice three years ago still counts as “returning.”

Look at instead:

  • 90-day repeat rate (of customers who bought 90 days ago, what % bought again?)
  • Time-between-orders (median days between order 1 and order 2)
  • Cohort retention curves (visualised properly — most stores never look at these)

Total sales

The lie: “Sales are up 15% month-over-month.”

The truth: sales lag — they reflect marketing spend from previous months, seasonal demand, and one-off campaigns. A 15% MoM lift after Black Friday is normal; a 15% MoM lift in May from “doing nothing differently” is unusual and worth investigating before celebrating.

Look at instead:

  • Sales smoothed over 30 days vs. 30 days prior (vs. 30-day lookback)
  • Sales by acquisition cohort (are customers acquired in March worth more than customers acquired in February?)
  • Marketing-attributed sales vs. organic baseline

The metrics actually worth tracking

If we audited a store and got to keep five metrics, we’d pick:

  1. CAC blended — cost of acquiring one new customer, all channels combined
  2. LTV : CAC ratio — should be > 3:1 for a healthy DTC business
  3. 90-day repeat rate — proxy for product-market fit
  4. Mobile conversion rate vs desktop — proxy for storefront UX quality
  5. Email-attributed revenue % — proxy for retention infrastructure

Total sales is downstream of all five. Conversion rate is one input among many. AOV is contextual. Treat the headline numbers as starting questions, not answers.

A weekly ritual

Every Monday, look at:

  • Last week’s revenue vs. 4-week trailing average
  • Last week’s CAC vs. trailing
  • Last week’s email-attributed share
  • One question from each: what changed, why?

Five minutes. More signal than the dashboard can offer in an hour.

— Read next

Setting up Shopify Tax properly (without paying $5,000 for an audit later)

— Hit a wall?

We can help building properly?