Per-locale conversion rate. Single-locale dropoff = local UX issue, payment-method gap, or localisation error.
At a glance
Conversion rate broken out per locale across your Salesforce Commerce Cloud (SFCC, formerly Demandware) realm, one bar per storefront / locale. A realm-wide conversion rate is almost meaningless when sites have different traffic mixes, currencies, and payment methods; this card is where conversion becomes actionable. A single locale dropping below the group average is the signature of a market-specific problem: a broken localisation, a missing local payment method, a checkout bug in one language, or a UX regression that only shipped to one site.
| What it counts | Conversion rate (orders divided by sessions, expressed as a percentage) calculated per locale / site over the window. Each bar is one locale’s CR; the card compares them against each other and against the group average. |
| Why it matters | This is the diagnostic that a realm-wide number cannot give you. SFCC’s multi-site design means problems are usually local: a payment gateway not enabled for one country, a translation that breaks a checkout step, a shipping rule that fails for one region. Per-locale CR isolates the affected market immediately instead of dragging down a blended average you cannot act on. |
| Reading the value | Read each bar relative to the group, not in absolute terms. Markets legitimately differ (a mature market converts higher than a new one), so the signal is a locale that has fallen sharply below the others or below its own prior period. The alert specifically watches for any locale running far under the group average. |
| Session source | Conversion needs sessions as well as orders. Sessions come from the connected analytics layer (for example GA4) or SFCC’s own analytics, joined to the order data per locale. If a locale has order data but no session data, its CR cannot be computed and the bar will be empty. |
| What “drop” usually means | One locale below the group is rarely a demand problem and almost always a mechanics problem: payment-method gap, localisation / translation error, a country-specific checkout or shipping rule, or a UX regression deployed to one site. |
| Currency / tax | Conversion is a ratio, so currency does not distort it the way it distorts revenue. This makes per-locale CR directly comparable across markets in a way revenue is not. |
| Unit | percent |
| Time window | 30D vsP (30-day window compared against the prior 30 days) |
| Alert trigger | any locale <50% of group avg, a locale converting at less than half the group average raises the flag |
| Sentiment key | scc_conversion_by_locale |
| Roles | owner, marketing |
Calculation
Calculated automatically from your Salesforce Commerce Cloud data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A Fortune-500 retailer runs one SFCC B2C realm with five DTC locales. The 30-day window covers 14 May 26 to 12 Jun 26, compared against 14 Apr 26 to 13 May 26.siteId / locale | Sessions | Orders | Conversion rate | vs prior 30d | vs group avg |
|---|---|---|---|---|---|
RefArch-US, en_US | 6,420,000 | 184,200 | 2.87% | +0.04 pt | above |
RefArch-UK, en_GB | 2,180,000 | 62,400 | 2.86% | +0.02 pt | above |
RefArch-DE, de_DE | 1,640,000 | 41,800 | 2.55% | -0.06 pt | near |
RefArch-JP, ja_JP | 1,090,000 | 28,600 | 2.62% | +0.01 pt | near |
RefArch-FR, fr_FR | 980,000 | 11,300 | 1.15% | -1.38 pt | far below |
| Group average | 2.61% |
- France is the story, and a blended number would have hidden it. The realm-wide CR is 2.61%, perfectly healthy. But
RefArch-FRis at 1.15%, less than half the group average, so it trips theany locale <50% of group avgalert. If you only watched the realm CR you would never see it: four healthy markets mask one broken one. This is exactly why conversion is a per-locale card on SFCC. - France also collapsed versus its own prior period (-1.38 pt). The combination of “far below the group” and “far below its own history” points to a recent, France-specific change, not a structural market difference. A market that is simply hard to convert would be low but stable; a 1.38-point drop month-on-month is a regression that shipped.
- The likely causes are mechanical, not demand. France still has strong traffic (980,000 sessions), so people are arriving and leaving without buying. The usual suspects, in order: a French-locale checkout or translation bug, a payment method (for example a local card scheme or wallet) that stopped working for fr_FR, or a shipping / tax rule that fails for France. Pair with Payment Status Distribution to check for a payment-stage failure concentrated in that locale.
- The other four locales differ but do not alarm. DE at 2.55% and JP at 2.62% sit below US and UK, but that is normal market variation, not a fault. The card’s value is distinguishing “naturally lower” (DE, JP) from “broken” (FR). Read the bars against the group, and weight recent movement over absolute level.
Sibling cards merchants should reference together
| Card | Why pair it with Conversion by Locale |
|---|---|
| Conversion Rate | The realm-wide CR. This card decomposes it per locale; read them together to see whether a blended move is broad or one-market. |
| Revenue by Site / Locale | A locale converting poorly will show a revenue share below its traffic share. Pair to quantify the lost revenue. |
| Payment Status Distribution | A single-locale CR drop is often a payment-method gap; payment status by stage confirms whether checkout is failing. |
| Orders by Currency | A locale with heavy traffic but few orders shows up as a shrinking currency slice; cross-check the volume side. |
| Active Sites / Storefronts | A locale with zero conversion may be a site that has gone dark rather than one that is converting badly. |
| Average Order Value | Conversion and AOV together describe the health of a market; a market can convert well at low AOV or poorly at high AOV. |
Reconciling against Salesforce Commerce Cloud
Where to look in Business Manager: SFCC’s admin tool is Business Manager, at a per-realm URL likehttps://<realm>.business.demandware.net. The closest native view is Merchant Tools, Site, Reports & Dashboards, where the per-site conversion and traffic reports live (often under a Conversion or Traffic dashboard depending on your Reports & Dashboards configuration). Because Business Manager reports are scoped per-site via the site picker, you reproduce this card by reading each site’s conversion report in turn.
Note that SFCC’s own conversion figures and the figures here can use different session definitions. SFCC’s storefront analytics count sessions one way; a connected analytics platform (GA4) counts them another. The card uses whichever session source the connector is configured with, so confirm the session source before comparing to a Business Manager conversion report.
Other Business Manager views to be careful with:
- Reports & Dashboards, Conversion / Traffic (per-site): the nearest native equivalent, for one site at a time.
- Reports & Dashboards, Sales (per-site): gives orders, not the session denominator, so it is half the ratio.
- Einstein / Commerce Insights dashboards: may report conversion with their own session and attribution model, expect definitional differences.
| Reason | Direction of divergence |
|---|---|
| Session definition. The denominator depends on the session source (GA4 vs SFCC storefront analytics vs another platform). Different session counts produce different CRs from the same orders. | Either, often material |
| Bot / spam filtering. The analytics layer may exclude bot sessions that SFCC’s raw counts include, or vice versa, changing the denominator. | Either |
| Time-zone. BM per-site reports use the site’s time zone; Vortex IQ uses UTC by default. | ±1 day at boundary |
| Locale vs site grouping. If a site serves multiple locales, the card may split CR by locale while BM reports by site. | Slice-level differences |
| Order status filter. The numerator (orders) can vary with status filtering; SFCC has many order statuses and the card and BM may filter differently. | Either, depends on filter |