Ecommerce
Ecommerce metrics that matter: measure what steers growth
June 22, 2026 · 7 min · ScaleLab
Dashboards fail in two directions: too little (revenue and a feeling) or too much (forty charts nobody acts on). A store is steered with a small set of numbers that connect traffic to profit — few enough to watch weekly, connected enough that when one moves, you know where to look next.
01The profit trio: revenue is not the goal
Revenue is the headline, but contribution margin — revenue minus product costs, shipping, payment fees and ad spend — is the number that decides whether growth is real. Stores have scaled revenue into bankruptcy; nobody scales contribution margin into bankruptcy.
Alongside it: gross margin per order and MER (total revenue over total marketing spend). MER is blunter than channel ROAS but harder to fool — it catches the attribution double-counting that makes every channel look profitable while the total says otherwise.
02The efficiency pair: CAC and LTV
Customer acquisition cost only means something next to lifetime value: LTV comfortably above CAC is a business, LTV near CAC is a treadmill. The ratio also sets your ceiling — it decides how aggressively you can bid for growth before growth becomes loss.
Segment CAC by channel and LTV by cohort. Blended averages hide the fact that some channels buy loyal customers and others buy one-time discount hunters at the same sticker price.
03The store health pair: RPV and repeat rate
Revenue per visitor — conversion times AOV — is the single best measure of how well the store itself performs, because it is immune to the conversion-versus-basket trade-off that distorts either metric alone. It is the number your CRO work should move.
Repeat purchase rate is the quiet compounding engine: acquiring a customer is expensive, keeping one is cheap, and a small lift in repeat rate often outearns a large lift in traffic. If you track one retention number, track this.
04Make the numbers operational
A metric steers only if it has a cadence and an owner: weekly review of the trio and pairs, monthly cohort review, and a pre-agreed threshold for each number that triggers investigation rather than debate.
And verify the plumbing quarterly: a broken purchase event or double-firing pixel silently corrupts every number downstream. Trustworthy data is not a given — it is maintained, the way inventory is.