Storefront reliability for commerce teams

Find the bugs that cost sales.
See the fix bring them back.

Dozenfold measures what each broken request, slow page and failed step costs in purchases, against comparable shoppers who didn’t hit it. Then it hands your team the evidence and follows the result after the fix.

14-day trial · From $49 a month · Billed through Shopify · Custom storefront? See how setup works

Shopify & custom storefronts Revenue per day, against a control Bots never counted

EXPLORE WHAT DOZENFOLD DOES TODAY

One storefront.
A connected picture.

Audience, journeys, performance, errors and releases.
Start with the question on your mind.

Explore your storefront from every angle.

Nnorthstar/Your storefront
MonitoringLast 30 days Auto · daily
AUDIENCE & SEGMENTS

Who shops here, and who struggles?

Each session counted once
Observed sessions24,891Bots & crawlers excluded
Started on a product page41%Landing page
From paid social22%Channel
On mobile71%Device
Landing page
  • Product page41%
  • Home27%
  • Collection19%
Channels
  • Organic search31%
  • Paid social22%
  • Direct18%
Countries
  • United States46%
  • Canada14%
  • United Kingdom11%
Devices
  • Mobile71%
  • Desktop26%
  • Tablet3%

Also campaigns, browsers and operating systems. Every row opens its sessions.

Illustrative product dataOne storefront. Connected evidence.
Go on, follow a signal.Audience, journeys, performance, errors and releases.
Useful on their own. Better together.

Illustrative product views. Take the full product tour ↗

Know your shoppers

Every question, split the same way.

Landing page, channel, campaign, country, device, browser, connection and device memory travel with every session, so a pattern in your audience can be followed into an issue, a journey step, a page or a Core Web Vital. Audience & breakdowns →

Breakdowns in the product today
  1. OverviewWho are this week’s shoppers? 7 Landing pageChannelCampaignCountryDeviceBrowserOS
  2. IssuesWhere does this error concentrate? 11 BrowserDeviceOSViewportCountryConnectionDevice memoryPage groupURLChannelCampaign
  3. JourneysWho leaves at this step? 3 DeviceBrowserChannel
  4. PagesIs this page worse for someone? 2 EnvironmentChannel
  5. PerformanceWho waits the longest? 5 BrowserOSViewportConnectionDevice memory

Bots, crawlers and headless browsers never enter a segment.

Illustrative issue · TypeError in reviews-widget.js

Not every error is yours to fix.

Each issue shows how much of it comes from the shopper’s environment.

  • Browser extensions46%
  • Instagram & TikTok in-app browsers31%
  • Offline when the request failed7%
  • Your storefront16%

Mostly the shopper’s environment. Check before anyone starts debugging.

Revenue impact

Revenue you can defend.

Not affected sessions × conversion rate × order value. Dozenfold compares shoppers who hit an issue with shoppers at the same step, on the same device and environment, who didn’t, and shows the gap with a 95% interval.

Every amount reads per day the issue was seen, over at least a week of shoppers. Stores with few orders still see who fails to reach the next step.

Inspect an example calculation →
Cart update fails after a delivery changeIllustrative issue
$614/dayMeasured loss
Hit the issue10%
Same step, no issue34%

Purchases within 30 minutes · gap 24.0 pts · 95% interval 21.2–26.8 · 480 compared sessions

  • Measured the whole interval is above zero
  • Directional positive, still uncertain

Error-session replay · Opt-in

See the moment it broke.
On your terms.

Replay starts off. When you turn it on, only sessions with an error upload a recording: the minute before it and the rest of that visit.

You choose the sample rate and how much of the page stays visible. Form fields are always masked.

  • Opens 10 seconds before the error
  • Says why when there is no replay
  • Kept 30 days
See error-session replay →
Settings · Session replayIllustrative settings
Session replay for errorsRecords only sessions where an error happens.
Sample rateShare of sessions that keep a one-minute buffer.
Masking
Mask all text and inputsLayout and interactions only. Recommended. Mask inputs onlyProduct names and prices stay visible.
LAST 7 DAYS · REPLAY REACH
Allowed analytics
4,920
With an error
388
With a replay
94

At 25%, about a quarter of sessions with an error have a replay. Raise the rate for more coverage.

Where your team already works

Where your team already works.

Alerts in Slack or Telegram, a weekly summary by email, the fix in your tracker, the release from CI, and the same evidence in your AI coding tool.

Slack and Telegram alerts, per channel

Each Slack channel or Telegram chat picks what it receives: checkout errors, cart failures, new issues, regressions, spikes or slower Core Web Vitals.

checkout-alerts

New checkout issue · Cart update fails after a delivery change

Est. $614/day · measuredLikely owner: theme

Set up alerts →

A weekly report, every Monday

Store owners get one email with the week at a glance, and what is still worth setting up.

Your storefront this week
Sessions
5,812
With errors
4.1%
LCP p75
2.6s
  1. Cart update fails after a delivery change$614/day
  2. Payment button script times out≈ 3 orders/day
  3. Reviews widget throws on Safaridirectional
What’s in the report →

Linear and GitHub issues

The Developer brief opens as a prefilled issue in your tracker, under your own sign-in. Dozenfold stores no tracker credentials.

Developer brief →

Releases from CI

Mark each deploy and upload its private source maps from CI, so the next release is followed from its first session.

npx dozenfold releases create Release setup →

Available now · AI tools (MCP)

Your AI coding tool reads it too

Add a read-only key to Claude Code, Cursor or another MCP client. Ask what costs the most, or what breaks in the file you are editing, and check whether a shipped fix stopped it. Included in every plan.

Connect your AI tools →

Next: an AI reliability engineer that works on the fix. See the direction

Illustrative MCP session · Available

What breaks in assets/cart.js, and did the fix work?

list_problemsfile: assets/cart.js · 2 open, ranked by impact

get_problemDelivery change fails · cart.ts:184 · theme

verify_fixstorefront@2.18.5 · 0 since vs 41 before · stopped

Safety & privacy

Built for a live storefront.

How the script stays safe →

Out of your store’s way

Our requests use the browser’s own fetch, not your theme’s, and leave room for your analytics and ad beacons.

Shoppers, not bots

Crawlers, headless browsers and synthetic monitors are never collected or billed.

Privacy by default

No fingerprinting, no input values, no cross-session profile. Replay only if you turn it on.

Backs off under load

Replay stops on a flood of page changes or a rate limit instead of slowing the page.

Shopify app with theme embed and PixelAny storefront with one scriptIntegrations →

Common questions

Straight answers.

Does it show revenue lost or revenue at risk?+

An estimate of missed purchases and their value, shown per day the issue was seen. Dozenfold compares sessions that hit the issue with clean sessions at the same journey stage, device and environment, over at least seven days, and reports the gap with a 95% interval: a measured loss when the whole interval is above zero, a directional signal when it is positive but uncertain. It does not present the estimate as booked loss, project a short window into a monthly loss, or add overlapping issues into a store total. See how the estimate works →

Is it safe to install on a live store?+

Yes, and it is built for that. The collector adds no shopper-facing UI. Its requests bypass your theme’s fetch wrappers, it uses at most half of the browser’s page-exit request budget, its wrappers return your code’s original result and errors, and it never blocks a popup. We can also switch collection off remotely. How the script stays safe →

Exactly what shopper data does Dozenfold collect?+

Anonymous error, performance, page-group, bounded diagnostic and funnel evidence, plus coarse context: country, connection type, device class and whether a page ran in a social app’s browser. We do not ask for customer identities or order payloads, and we do not record input values or cross-session shopper identities. Screen replay stays off unless a store turns on error-session replay.

Is every error a bug in my store?+

No. Some come from browser extensions, from shoppers who lost their connection, or from social apps’ built-in browsers. Each issue shows those shares, so you can tell a store bug from an environment problem before anyone starts debugging. How issues are grouped →

Do bots count toward my plan?+

No. Googlebot and other search-engine renderers, headless browsers, Lighthouse and uptime monitors are not collected: the script does not start in them and the collector drops what slips through. They never become sessions, never open an issue and never use your allowance. Test your installation in a regular browser for the same reason. Automated traffic →

How do we get started?+

On Shopify, install Dozenfold from the Shopify App Store, approve a plan and switch on the theme app embed. The 14-day trial starts there and billing goes through your Shopify invoice. On a custom or headless storefront, book a setup call: we create your site and send the script, usually the same day, and your developer adds it to the storefront. Booking a demo does not start a trial or subscription. See plans and pricing.

Put Dozenfold to work

Keep the path to purchase
working.

Explore your storefront from every angle: shopper journeys, performance, technical issues and what changes after a release.

  1. 01
    Find where shoppers struggle

    Journey drop-offs, slow pages and recurring failures.

  2. 02
    See the developer handoff

    The request, browser, release and source context in one place.

  3. 03
    Watch what changes after shipping

    Release comparisons, buying flows and issue recurrence.

DOZENFOLD / PRIVACY

Make yourself at home.