Resource
E-commerce metrics glossary: GMV, CVR, AOV, ACoS, and 20 more operators actually use
An operator's glossary of the metrics that actually drive DTC decisions — what they measure, how to calculate them, and where they lie.
TL;DR — Half the DTC brands we meet optimize the wrong metric. This glossary lists the ones that actually move the business (GMV, CVR, AOV, CAC, LTV, ROAS, ACoS, sell-through), how they're calculated, and where they lie. If you only act on one metric this week, make it contribution margin — not GMV.
How to read this glossary
Three groups: revenue (what comes in), acquisition (what it costs to bring a customer), retention (how often they come back). At the end, a short list of platform-specific metrics for Amazon, Mercado Libre, and TikTok Shop — relevant for any brand operating or considering LATAM.
Operator rule: never optimize a single metric in isolation. Each has a sibling that contextualizes it. GMV without margin is vanity. CVR without qualified traffic is a trap. CAC without LTV means nothing.
Revenue metrics
GMV — Gross Merchandise Value
What it measures: total value of goods sold in a period, before platform fees, logistics, returns, and ads.
Formula: sum of (unit price × units sold) across all transactions.
Why it matters: it's the number TikTok Shop, Mercado Libre, and Amazon report first because it sounds big. Useful for momentum and quarter-over-quarter trend.
Where it lies: it's not net revenue and it's not profit. A brand with high GMV and negative contribution margin is losing money. GMV alone hides that until accounting closes.
Operator take: report GMV to the board. Make decisions with contribution margin.
AOV — Average Order Value
What it measures: average revenue per order.
Formula: GMV ÷ number of orders.
Why it matters: tells you whether to chase more customers or more value per customer. Raising AOV 20 % is usually easier than getting 20 % more traffic.
Where it lies: averages hide distribution. If you have one high-ticket SKU selling rarely and many low-ticket SKUs, your AOV doesn't reflect operational reality.
Operator take: segment AOV by channel. TikTok Shop AOV trends higher with bundles; Mercado Libre AOV trends higher in heavy categories (home, electronics).
Sell-through rate
What it measures: share of inventory sold in a period.
Formula: units sold ÷ units available × 100.
Why it matters: indicates inventory rotation. 60 % monthly is healthy for DTC; under 30 % is dead capital.
Where it lies: high sell-through at low margin just means you're selling cheap. Cross-reference with contribution margin.
Return rate
What it measures: share of orders returned.
Formula: returned orders ÷ total orders × 100.
Why it matters: in beauty and fashion, above 15 % destroys margin. In food and supplements, above 5 % signals quality or listing-accuracy problems.
Operator take: push the cost of a return (logistics + restocking + damaged unit) into CAC, not fixed cost. Channel profitability shifts.
Contribution margin
What it measures: unit profit after direct variable costs.
Formula: sale price − COGS − platform fee − logistics − allocated ads − sales tax on fees.
Why it matters: the only metric that tells you whether selling one more unit leaves you better or worse. Everything else is noise if this is negative.
Where it lies: if you don't allocate ads correctly (per channel, per SKU), margin looks inflated. Use per-SKU ACoS or ROAS, not account-level.
COGS — Cost of Goods Sold
What it measures: direct cost of producing or sourcing the unit.
Operator take: refresh COGS quarterly. Raw material prices in LATAM track USD/MXN, and if you froze COGS 9 months ago, your real margin is probably 3–5 pp lower than you think. Same principle applies to US brands sourcing from Asia — currency + tariffs shift the floor.
Acquisition metrics
CVR — Conversion Rate
What it measures: share of visitors who purchase.
Formula: orders ÷ sessions × 100.
Why it matters: cleanest indicator of how well your PDP or funnel is "selling." 2 % is solid for DTC with cold traffic; 5 %+ is Amazon-grade for high-intent.
Where it lies: CVR depends on traffic mix. You can "improve" CVR by cutting cold prospecting ads and keeping only warm audiences — but you shrank the business. Always cross with volume.
Rough benchmarks by channel (LATAM + US DTC):
- TikTok Shop: 1–3 % (organic), 0.5–1.5 % (ads)
- Amazon: 10–15 % (strong reputation + imagery), 3–8 % (new product)
- Mercado Libre: 6–12 %
- Shopify/DTC: 1.5–3 %
CTR — Click-Through Rate
What it measures: of people who saw your ad or listing, how many clicked.
Formula: clicks ÷ impressions × 100.
Why it matters: measures the strength of your hook (main image, title, first line). On Amazon, low CTR drags your organic rank — the algorithm infers your product isn't interesting.
Operator take: if CVR looks fine but CTR doesn't, fix the main image before the listing copy. If CTR is fine but CVR is bad, fix the detail page (reviews, bullets, price).
CPC — Cost per Click
What it measures: average spend per ad click.
Operator take: CPC varies heavily by platform and category. TikTok Shop Ads typically run cheaper than Amazon PPC but with lower CVR — always compare CPC against CVR in the same attribution window.
CAC — Customer Acquisition Cost
What it measures: average spend to acquire a new customer.
Formula: total acquisition spend ÷ new customers.
Why it matters: CAC is the metric that kills companies. If CAC exceeds first-order contribution margin, you're paying to lose money — you only survive on repeat.
Where it lies: blended CAC (all channels together) hides that one channel is profitable and another is subsidizing. Always CAC by channel.
ACoS — Advertising Cost of Sale
What it measures: share of ad revenue consumed by ad spend.
Formula: ad spend ÷ ad-attributed revenue × 100.
Why it matters: Amazon's canonical PPC metric. If ACoS is above gross margin, you're paying to sell at a loss.
Rule of thumb: target ACoS ≤ gross margin × 0.5. 30 % gross margin → target ACoS ≤ 15 %.
ROAS — Return on Ad Spend
What it measures: revenue generated per dollar of ad spend.
Formula: ad-attributed revenue ÷ ad spend.
Why it matters: the inverse of ACoS, more common off-Amazon. 4× ROAS is healthy for DTC; 2× only works if LTV justifies losing money on first order.
Attribution window
What it measures: how long after a click or view a conversion still "counts."
Operator take: Meta Ads defaults to 7-day-click / 1-day-view. TikTok Ads, 7-day-click. Amazon Ads, 7 days PPC. If you compare channels using different windows, the longer-window channel looks artificially better.
Retention metrics
LTV — Lifetime Value
What it measures: expected total revenue from a customer over the relationship.
Simple formula: AOV × annual purchase frequency × average retention years.
Why it matters: CAC must be lower than LTV, ideally ≤ 33 % (the three-to-one rule).
Where it lies: 5-year projected LTV on an 18-month-old brand is fiction. Use 12-month LTV for operating decisions, 36-month only for investor projections.
Retention rate
What it measures: share of customers who buy again in a period.
Formula: (customers at period end − new customers) ÷ customers at period start × 100.
Why it matters: distinguishes churn businesses from businesses with a moat. Supplements and personal care: 40 % month-12 retention is a solid benchmark. Fast fashion: 15 % is acceptable.
Repeat purchase rate
What it measures: share of buyers who make a second purchase.
Formula: buyers with 2+ orders ÷ total buyers × 100.
Operator take: the most honest leading indicator of product-market fit. If 90-day repeat purchase is < 10 %, the problem isn't marketing — it's product or experience.
Churn rate
What it measures: share of customers who stop buying in a period.
Operator take: in subscription, 5 % monthly churn is already critical (50 % of base lost in 12 months). In transactional DTC, "churn" is measured as "didn't return in X months" — definition varies by category.
Platform-specific metrics
Amazon
- BSR (Best Seller Rank): category rank. BSR #1 in "Beauty — Skincare" might move 300 units/day; BSR #5,000 might move 3. Logarithmic distribution.
- Session Percentage: share of listing sessions from shoppers who viewed other products first. Tells you whether you compete with direct-search buyers.
- Buy Box %: share of time your offer wins the Buy Box. Losing Buy Box means losing ~85 % of potential sales.
Mercado Libre
- Reputation: green / yellow / red, based on claims, cancellations, and delivery time. Yellow cuts conversion ~30 %.
- Mercado Líder: badge that amplifies organic visibility. Earned via sustained volume, green reputation, and fast response.
- Publicación Premium: higher-cost listing tier with better placement and installment plans; ~5 pp higher commission than Clásica.
TikTok Shop
- GPM (GMV per Mille): GMV per 1,000 views. Set your category's floor with your own data in the first 60 days and use it to filter out creators who don't perform.
- Creator affiliate GMV share: share of total GMV from affiliate creators. In most accounts we've seen, it settles at 60–80 %.
- Cart conversion rate: sessions with a product added to cart that end in purchase. Below 30 % usually points to checkout friction (email/phone verification, payment options).
Funnel metrics worth watching
PDP bounce rate
Share of visitors who land on a PDP and leave without interacting. Above 60 % signals a bad main image or a mismatch between ad and landing.
Add-to-cart rate
Visitors who add to cart ÷ sessions. Benchmark: 5–10 % for DTC with mid trust.
Checkout abandonment rate
Carts started that didn't close. 70 % is normal; above 85 % suggests payment or shipping friction.
The three metrics most brands measure wrong
From 40+ DTC accounts across LATAM and the US:
1. Confusing landing CVR with account CVR. Account-level CVR (all sessions vs. all orders) is a weighted average without meaning when you mix prospecting and remarketing. Always CVR by audience.
2. Last-click attribution with short windows. TikTok often "doesn't get credit" for a purchase it originated, and Meta or Google does, because the shopper searches the brand after seeing the video. Install Meta Pixel + TikTok Pixel + GA4 — and consider a third-party tool (Northbeam, Triple Whale) once your volume justifies it.
3. Reporting GMV in mixed currencies. If you operate in MXN and USD without converting to a single currency, quarter-over-quarter comparisons lie. Fix exchange rate at period end, not at average.
Quick reference
| Metric | Formula | Cadence |
|---|---|---|
| GMV | ∑ (price × units) | Monthly / quarterly |
| AOV | GMV ÷ orders | Weekly |
| CVR | orders ÷ sessions | Daily (ads), weekly (organic) |
| CTR | clicks ÷ impressions | Daily (ads) |
| CPC | ad spend ÷ clicks | Daily |
| CAC | acquisition spend ÷ new customers | Monthly, by channel |
| ACoS | ad spend ÷ ad revenue × 100 | Weekly |
| ROAS | ad revenue ÷ ad spend | Weekly |
| LTV | AOV × frequency × retention years | Quarterly |
| Retention rate | (end customers − new) ÷ start customers | Monthly |
| Repeat purchase | buyers 2+ ÷ buyers | Monthly |
| Return rate | returns ÷ orders | Monthly |
| Sell-through | units sold ÷ units available | Monthly |
| Contribution margin | price − variable costs | Per SKU, always |
Next step
If you're building the operating dashboard for your brand and want a real template with cross-formulas (not a generic Looker Studio dump), book a working session with Roikon. The first deliverable is always an operator dashboard that separates what you watch from what you decide on.