All answers

The Prevailing
Answers for
Business model:
Business model, pick up to two

How should we build and read cohort reports?

The PrevailingPublished Oct 29, 2026

Our answer

Build cohorts by first-order month in Shopify or Klaviyo, grouping by quarter below about 100 new customers a month. Read rows, columns and diagonals. Judge a cohort only after 90 days, against your own earlier cohorts.

Confidence
Medium
Strongest proof
4 official rules
1 practitioner, 8 opinions or claims
Sources
11 sources
13 statements

New here? Take the 1-minute tour

Why this is the answer

3 reasons and our judgment

Your store or email platform builds the grid for you. In Shopify and Klaviyo, rows are first-order months and columns count months since. Klaviyo bases percentages on the cohort's original size.

Small cohorts swing by several points. One person moves a cohort under about 30 customers by several points (Coupler.io, data connector maker). Fullmetrix (ecommerce analytics software maker) wants at least 100 customers, and Impremis Marketing (performance marketing agency) wants 200 to 300.

Cohorts keep maturing for months. Lifetime value differences need at least 90 days to show (Dotidot, product feed tool maker). AppLovin's holiday cohort rose from a 9.4% repeat rate by day 90 to 14.5% by day 270.

Our judgment

The floor of about 100 new customers for a monthly cohort is ours. We take Fullmetrix's minimum as the middle of the sizes sources give, above Coupler.io's warning and below Impremis Marketing's floor for channel cohorts. Preferring the store or email platform over GA4 for purchase cohorts is the side we take, since Shopify's and Klaviyo's reports group customers by first order and Fullmetrix says GA4 falls short for purchase cohorts, while Dotidot shows a GA4 setup. The 90-day wait and the comparison with your own earlier cohorts at the same age combine Dotidot's maturity rule, AppLovin's Black Friday and Cyber Monday cohort still adding repeat buyers at day 270, and Fudge's advice to compare against your own October and September cohorts rather than a benchmark.

Our judgment, built from Fullmetrix, Coupler.io, Impremis Marketing, Shopify, Klaviyo, Dotidot, AppLovin and Fudge: our reading of these sources, not a finding from any one of them.

What the sources say

13 statements from 11 sources
  • ShopifyUndatedSupports

    Shopify's cohort report groups customers by first order month, week or quarter and shows repeat activity as a heatmap or retention curve, with a period 0 column for repeat orders in the first period.

    “The period 0 column captures returning orders by the cohort's customers in the same period as their first order.”

    Read it on help.shopify.com

  • ShopifyUndatedSupports

    Shopify's cohort report groups customers by first order month and tracks retention, sales and order value; projections need 24 months of store data and are not guarantees.

    “customers are grouped into cohorts based on the date that they placed their first order”

    Read it on help.shopify.com

  • ShopifyUndatedSupports

    Shopify's customer cohort report groups customers by first order date and can filter cohorts by the first order's sales channel, marketing channel, product or subscription.

    “customers are grouped into cohorts based on the date that they placed their first order”

    Read it on help.shopify.com

  • AppLovinOct 5, 2026Caveat

    AppLovin says in BFCM 2025 new-customer checkouts hit 3.3x the normal baseline versus 2.9x for returning customers, and suggests separate new-customer campaigns started a week or two early. Its own data.

    “Avoid cutting or reallocating their budget based on the first few days”

    Read it on applovin.com

  • DarkroomSep 4, 2026Caveat

    Darkroom, which sells retention services, says to split repeat-purchase curves by first channel, discount depth and entry product; often one channel or welcome offer causes most of the stall.

    “that hero SKU often has the lowest reorder rate in the whole catalog.”

    Read it on darkroomagency.com

  • FudgeAug 22, 2026Supports

    Fudge, a Shopify tool vendor, says after BFCM measure new-customer share, contribution per order by discount tier, the cohort's 30, 60 and 90 day repeat rate against earlier cohorts, and final returns.

    “If your deepest discount tier contributed close to nothing per order, you bought revenue, not profit.”

    Read it on fudge.ai

  • Polar AnalyticsAug 11, 2026Caveat

    Polar Analytics, which sells retention reporting, says retention belongs to cohorts, not months: a blended rate can fall while every cohort improves, and fast acquisition lowers it mechanically.

    “Acquire aggressively and your blended retention rate falls, mechanically”

    Read it on polaranalytics.com

  • DotidotAug 6, 2026Supports

    Dotidot, which sells feed and ad software, says to build purchase cohorts in GA4's Cohort exploration, start weekly over 12 weeks, and wait at least 90 days before judging LTV differences.

    “Judging cohorts before they mature leads to poor decisions. Most ecommerce businesses need at least 90 days to see meaningful LTV differences.”

    Read it on dotidot.io

  • EightxJun 1, 2026Disagrees

    Eightx, a CFO firm, says a healthy DTC brand targets 20 to 30% month-3 cohort retention, below 15% signals a product problem, and 90-day cohort contribution sets the CAC ceiling.

    “A healthy DTC brand targets 20 to 30 percent month-3 retention; below 15 percent signals a product or onboarding problem, not just a CAC problem.”

    Read it on eightx.co

  • Impremis MarketingMay 10, 2026Supports

    Jordan Glickman of Impremis Marketing, an agency, says to judge channels by 90-day customer value against acquisition cost, and to group small channels until each cohort has 200 to 300 first-time buyers.

    “Blended LTV tells you the average. Cohort analysis by channel tells you where to actually invest.”

    Read it on jordanglickman.com

  • FullmetrixApr 14, 2026Supports

    Fullmetrix, which sells cohort analytics, explains reading cohort tables across and down, and suggests at least 100 customers per cohort, with no method given for that floor.

    “For statistically significant trends, aim for at least 100 customers per cohort. Below that, individual variations skew percentages.”

    Read it on fullmetrix.com

  • Coupler.io2026Supports

    Coupler.io says Shopify's cohort report cannot split by UTM campaign, so tag first orders by order index, count value in a fixed window such as 90 days, and compare only cohorts that have finished it.

    “mislabels existing customers as new”

    Read it on coupler.io

  • KlaviyoAug 6, 2025Supports

    Klaviyo's cohort report shows the last 12 full months of a conversion metric by first-activity month. In its worked example, the most common time for repeat purchases is 3 months after the first.

    “Averages and percentages use the original cohort size, including profiles that did not repeat the activity.”

    Read it on help.klaviyo.com

In the full answer

For readers who sign in
  • What to do this week3 steps
  • The numbers behind it2 figures from cited sources
  • Where each source standsA chart placing each source on the question
  • The strongest disagreementThe best case against the answer, from a cited source
  • How sure we areWhat its confidence rests on, and what would change the answer

Read next

More in Measurement