Strategy
Social media metrics Canadian ecommerce brands should track
Social media metrics Canadian ecommerce brands should track: CAD revenue per impression by language segment, and the thresholds that trigger action.
What to take away
- Revenue per thousand impressions in Canadian dollars, split by language segment, tells a Canadian store which audience pays.
- Act when a segment drops more than 20 percent below its own 90-day median for two consecutive months.
- Revenue per impression cannot prove that a post caused a sale. Only a withhold group can.
- Platform attribution windows are written for US accounts and overstate returns on cross-border orders.
Most Canadian ecommerce dashboards show one blended revenue figure in US dollars, set against an American benchmark. That comparison fails on currency, tax treatment, audience and shipping economics at once. The metrics below survive those differences. Each comes with a limit and a point at which you stop watching it.
What to measure
Canadian revenue per thousand impressions, shortened here to CAD RPM, is net revenue in Canadian dollars divided by impressions, multiplied by 1,000, for one audience segment and one stated window. Net means after returns, discounts and sales tax. Impressions count every render, including repeat views from the same account.
The figure is worth building because it survives currency conversion and because it can be cut by language. A buyer in Quebec reading French creative is a separate segment from an anglophone buyer in Ontario, and the two rarely behave alike.
| Metric | Definition | Window | Act when |
|---|---|---|---|
| CAD RPM | Net CAD revenue per 1,000 impressions, one segment | 30 days | More than 20% below own 90-day median, two months running |
| Bilingual split | CAD RPM reported separately for French and English creative | 30 days | French below 60% of English for a full quarter |
| Assisted orders | Orders with any social touch anywhere in the path | 90 days | Falling while last-click orders hold flat |
| Holdout delta | Revenue difference between exposed and withheld groups | Per campaign | Confidence interval crosses zero |
Impressions come from the platform, so a change in ad delivery moves CAD RPM with no change in demand. That is one reason to compare Facebook ads versus boosted posts before reading a delivery shift as a performance result.
How to read it
Read each number against the same segment's own past, never against a US benchmark. A French-language CAD RPM means little alone. It means a great deal against the same segment in the previous quarter and against the English segment in the same window.
If French CAD RPM stays below 60 percent of English for a quarter, the cause is usually creative rather than demand. Translate the offer, not only the caption. Quebec's Charter of the French Language makes French commercial advertising a baseline requirement rather than a test to run later.
Reporting breaks when paid, organic, email and affiliate results merge into one line. Dashboards need four separated measurement layers before any of these figures can be read.
No Canadian equivalent exists for the US benchmark reports that agencies circulate. Federal business resource pages describe programs and industry data rather than social benchmarks, so your own history is the only honest comparison.
What it cannot tell you
Impressions are counted by the platform, not by you. They include the same person seeing a post several times, accidental scroll-pasts and automated traffic, and the counting rules differ between networks. CAD RPM can fall while real demand rises.
The metric also cannot separate demand you created from demand that already existed. A customer who would have bought anyway and later saw a retargeting ad counts the same as one persuaded by a post. Social media analytics definitions are conventions that shift between platforms and vendors, not fixed standards.
Attribution and its limits
Platform conversion reports use view-through windows that differ by network and change without warning. A Canadian store shipping to the United States adds a second distortion, because an order placed in Ohio can be credited to a campaign run from Toronto.
Tax treatment matters as well. GST/HST collected on a sale is not revenue, and the rate depends on the province and the supply; the CRA explains which GST/HST rate applies. Removing tax before dividing by impressions keeps CAD RPM comparable across provinces.
Orders that begin in a feed and finish inside an app sit between channels. Social commerce marketing sets out how those paths are structured and where revenue should be counted.
When to stop measuring and decide
Two consecutive months below the 20 percent band is the threshold to act. So is a holdout delta whose confidence interval crosses zero. The question then stops being about measurement and becomes a choice between audience, creative and channel, which is where social media strategy explained picks up.
A withhold group is the only honest control. Without one, label the number directional and stop presenting it as return.
A quarterly review checklist
- Every metric has a named owner and a stated window
- French and English segments are reported separately
- At least one holdout ran during the quarter
- Sales tax is removed before revenue is divided by impressions
Common questions
How many impressions does a segment need before CAD RPM is usable? Enough that one order cannot swing it. Treat a segment under a few hundred impressions as noise rather than a result.
Should I trust platform-reported revenue or my own order data? Your own. Platform figures are useful for direction, but only the order record handles returns, discounts and tax consistently.
Does French creative need its own budget line? If Quebec is a real market, yes. A shared budget hides the gap between segments, which is the number you are trying to see.
What if the sales cycle runs longer than 30 days? Widen the window and say so. A 90-day CAD RPM on a considered purchase is more honest than a seven-day figure that captures impulse buyers only.
