Skip to content

Targeting and audiences

Choose where a test runs, when visitors count, who sees it (devices, traffic, cart and product conditions), traffic share and IP rules, and keep overlapping tests apart.

On this page

The Targeting tab decides who enters a test and where. A visitor enters the first time they open a matching page and stays in their group from then on.

The tab shows the two things most tests set, Pages and Who can join (devices and new or returning visitors). Everything else is folded under Advanced targeting, with a one-line summary of what it's set to: when visitors are counted, traffic sources and rules, IP addresses, keeping shoppers out of other tests, and how many visitors join. It opens by itself when something in it is already set.

Pages

  • Page type: All pages, Home, Product pages, Collection pages, Cart page, Content pages, Search, Blog or Articles.
  • Only some URLs? (optional): include or exclude URLs that contain, are exactly, start with, end with or match a pattern. Letter case, a pasted full address (https://your-store.com/...) and the language or market prefix don't matter, so /fr/products/x matches /products/x. A space in a value matches however the address writes it: contains filter.v.option.color=light blue matches Light+Blue and Light%20Blue.
  • Product pages opened from a collection: /collections/x/products/y is the product page /products/y. Product rules match it, and collection rules (like starts with /collections/x) don't: it isn't a collection page.
  • Starts with a collection, page or blog: starts with /collections/shirts covers that collection and its sub-pages and filtered pages (/collections/shirts/red), never another collection like /collections/shirts-sale. For a family of handles use contains /collections/shirts or a pattern. Product addresses stay a plain prefix: starts with /products/hoodie- matches every hoodie whose handle starts that way.
  • Test a URL: type a path or paste a full address to see whether it's in the test, even before you save. It says In the test, or Not in the test and why (which rule leaves it out, or that /collections/shirts-sale is a different collection from /collections/shirts), and gives the same answer your store will.
  • Pages that open without a reload: tests start when a page loads. Filters, sorting and variant picks that only change the query string (?filter..., ?variant=) are checked again on the spot. A theme or page builder that swaps whole pages without reloading (page transitions) starts a test only on the next full page load, so preview your theme's navigation before you launch.

Price, offer, subscription, shipping, checkout and theme tests always run sitewide. Template tests pick their pages on the Templates tab instead.

When visitors are counted

Every visitor in the test sees their group's change right away. When visitors are counted decides when they start counting in results, so visitors who never reach the change don't water them down.

OptionCounts a visitorUse it for
On page load (default)when they open a targeted pagemost tests
When the cart drawer opensthe first time the cart drawer opens (your theme's, a cart drawer app's or Crovio's) or the cart page loadschanges inside the cart drawer
When an element appearsonce the element is visible and at least half on screen for a moment (a closed popup or hidden tab doesn't count)a popup, a tab, a section further down the page
On a custom eventwhen your theme or another app calls Crovio.activate("event_name")a quiz step, a custom widget
When checkout startswhen they start checkoutshipping rate and checkout tests

Which options you see depends on the test type:

  • Content tests: page load, cart drawer, element or custom event.
  • Shipping rate tests: all five. Choose When checkout starts to count only visitors who start checkout, where the rates are shown: each one is counted once, with the group their checkout carried.
  • Checkout tests: always when checkout starts. The change only shows in checkout, so every group sees the same store until then and there is nothing else to choose.
  • Price, offer, subscription, free shipping and shipping discount tests: always on page load. A different price or offer changes who adds to cart, opens the drawer or starts checkout, so counting only those visitors would compare different kinds of shoppers and bias the result.
  • Split URL, template and theme tests: always on arrival.

A cart drawer Crovio doesn't detect

Crovio recognises theme drawers, cart drawer apps and its own drawer. If a preview shows yours isn't detected (the test counts no visitors), choose When the cart drawer opens and fill in Cart drawer selector:

  1. Click Pick on my store. Your store opens with the Crovio picker at the bottom, in browse mode, so the store works as usual.
  2. Open your cart drawer the way a shopper would (click the cart icon, or add a product).
  3. Click Start picking (or hold Alt / Option) and click inside the open drawer. The picker snaps to the drawer panel; cart icons, buttons and anything smaller than 200 × 150 px are refused.
  4. Click Use in Crovio, go back to the Crovio tab and save Targeting.

The same selector works for an Opened the cart custom metric (see Goals).

Audience

Checked once, when a visitor would join.

  • Devices: All, or any mix of Desktop, Tablet and Mobile.
  • Visitors: All, New or Returning.
  • Conditions: click Add condition and pick one.
ConditionExample
Traffic sourceis Paid ads (also Direct, Search engines, Social media, Email, Other websites)
UTM source, medium, campaignUTM campaign contains black-friday
Referrercontains instagram.com
URL parameterref is present
Landing pagecontains /collections/sale
Countryis any of United States, Canada (pick as many as you like). Where the visitor is, detected by Shopify, even on a store with one market
Market countryis any of Canada. The country the visitor shops in: the one picked in your store's country selector (Markets)
Languageis French
Customeris Logged in / Guest
Customer tagis VIP
Cart totalis more than 50 (in your store currency)
Items in cartis more than 2
Cart has product / vendoris any of the products or vendors you pick
Product, Product in collection, Product tag, vendor, typethe product on the page is any of the ones you pick
Product price, Product stockthe product on the page costs more than 30, or is in stock
Cookie (Advanced)member is present, or plan contains gold
JavaScript (Advanced)returns true: window.innerWidth > 1200

Country or Market country

Country is where the visitor physically is, detected by Shopify from their connection. Market country is the country they chose to shop in, in your store's country selector (Shopify Markets), which sets their currency and prices. A shopper in Canada browsing your US store is Country Canada and Market country United States. On a store with one market, Market country is the same for everyone, so use Country to target by location.

Text conditions can be is, is not, contains, doesn't contain, starts with, ends with or matches pattern; URL parameters and cookies can also be is present or is missing. A JavaScript condition runs in the visitor's browser when they would join, and they qualify when it returns true.

OR and AND

Each condition sits in a box. Or adds another option to the same box: the box matches when any of its conditions does. And condition adds a new box that has to match too. For example, Traffic source is Paid ads or UTM source is newsletter, and Country is any of United States, Canada. The Who's in this test summary spells it out in words.

A condition you haven't finished is marked Incomplete and ignored, so a half-filled row never shuts everyone out.

Traffic conditions use how the visit started, so someone who arrived from a campaign still qualifies on later pages.

Cart conditions are checked against the visitor's cart. A visitor who doesn't qualify yet joins as soon as their cart does, for example when an add to cart takes the total over 50, without a page reload.

Product conditions read the product on the page, so they only match on product pages. Product in collection targets a collection's products wherever they're opened, including from the collection page (/collections/x/products/y).

Currency (price tests)

Price tests run in your store's currency, shown on a Currency card on Targeting. Shoppers browsing in another currency (through Shopify Markets) see and pay control prices.

IP addresses

Show a test only to certain IPs (your team, for QA on the live store) or never to some (your office, agencies, bots). Use a single address (203.0.113.7) or a range (203.0.113.0/24). Only me and Exclude me fill in your current address.

Crovio checks IPs on its server on the first page of each visit, so your lists never appear in your store's code. The page waits at most one second for the answer.

How many visitors join the test

Under Advanced targeting, set the share of matching visitors who join the test. This isn't the split between versions (that's on the Test groups tab). The rest always see your normal store and aren't counted. Who's in this test shows the resulting split, for example 45% Control, 45% Variant A, 10% not in test.

Prevent flicker hides the page for up to 0.8 seconds while variant JavaScript runs. CSS-only changes never flicker.

How visitors are split

  • Each visitor who matches gets a random group, in the proportions you set, and keeps it on every later page and visit in that browser.
  • Logged-in customers keep their group on other devices too. A guest on a new device or browser is a new visitor.
  • Groups in one test say nothing about groups in another: being in Variant A of one test doesn't make a visitor more likely to be in Variant A of the next.
  • Bots and crawlers, the theme editor and preview visits are never counted.
  • If the counts drift from your split, results show The traffic split looks broken (a sample ratio mismatch). See Reading results.

Keep shoppers out of other tests

If two tests could affect each other (same page, same offer), put them in one group: a shopper only ever enters one of them, so results stay clean. (This used to be called an exclusion group.)

  1. Under Advanced targeting → Keep shoppers out of other tests, pick a group or create a New group.
  2. Save targeting.
  3. Adjust How visitors are split across the group if you like, then Save split.

Only running tests share the group's visitors. Within its share, each test's own traffic % and groups still apply.

Was this page helpful?