Most brands drown in metrics and starve for insight. The dashboards are full, but the number that would actually change a decision is either missing or wrong. This is a working guide to the handful of marketing KPIs that move the needle for an ecommerce brand - CPL, CAC, ROAS, ROI,and the one most of your competitors ignore, LTV - plus the reason every one of them is only as good as the data underneath it.
A KPI - Key Performance Indicator - is a metric chosen to measure one specific, objective thing so you can compare it over time and against competitors. The point isn't to collect them. It's to set a goal for each, execute, review, adjust and iterate. Without KPIs you drift, chasing shiny tactics and new contractors while losing the thread on whether any of it grew revenue.
Objective metrics are the difference between guessing and deciding. They let you compare past, present and target performance and see, honestly, whether your strategy is working. Fewer decisions made on vibes, more made on evidence.
Now the metrics themselves.
CPL is total ad spend divided by leads acquired, at whatever level you want — channel, campaign, or ad. Media buyers love it because it's easy to track. That's also its weakness. CPL isn't reliably correlated with revenue and it's frequently distorted by ad platforms leaning on modeled data and stricter privacy settings. A lower CPL often doesn't mean more revenue — sometimes it just means more cheap leads that never buy.
There's a sweet spot between a flood of cheap leads that don't convert and a trickle of expensive leads that do but don't return enough to justify the cost. Finding it takes a real testing budget, some trial and error and accurate multi-touch measurement so you can see which leads actually became customers — not just which were cheapest to acquire.
CAC is arguably the most useful paid-advertising metric. On average, how much you spend to acquire a customer at the channel, campaign, or account level. Add up the spend, divide by customers attributed to it.
But here's the trap, and it's the same shape as CPL - lowering CAC isn't automatically winning. You can pull in more first-time buyers while spending more overall because those customers are low-value one-and-dones, not high-LTV repeat buyers. Chasing a lower CAC in isolation can quietly make you less profitable.
The fix is to stop optimizing raw CAC and start weighing new customer acquisition cost against lifetime value — what you truly pay to win a net-new customer, measured against what that customer is worth over time. Especially on cold traffic, feeding accurate LTV signals back to the ad platforms' machine-learning models through first-party conversion feeds like Google and Meta CAPI integrations is one of the most powerful ways to optimize for long-term profit instead of short-term volume.
ROAS is revenue attributed to ad spend, divided by that spend — the metric media buyers most often use to show performance. Two things reliably go wrong with it.
First, it's usually pulled straight from the ad manager and the ad manager is often wrong about revenue. Last-click models and limited cross-channel visibility mean the platform routinely miscounts and it has every incentive to claim credit generously. This is the "every platform grades its own homework" problem in a single metric. If your ROAS comes from the same platform whose performance it's supposed to judge, it isn't an independent number. Make sure your revenue attribution is verified against actual sales before you trust any ROAS figure.
Second, ROAS takes time to build. Inexperienced buyers kill campaigns before they've had a chance to develop positive ROAS. If you're using ROAS, give a campaign at least one full buying cycle before you judge it.
Not sure how long your brand's buying cycle actually is? That's exactly what the First Click Sales Velocity Report measures — how long customers take to convert after their first click so you're not guessing at when to judge a campaign.
The difference between ROI and ROAS. ROI uses profit and includes every cost tied to a campaign — ad spend, agency fees, creative, operating costs — while ROAS looks only at revenue against ad spend. Use ROI for the overall view of paid-advertising performance, use ROAS to zoom into a specific platform or campaign.
Here's the one your competitors probably aren't tracking, which is exactly why it's an advantage. LTV is about bringing in the right customers — the ones who love the product and buy again and again, growing in value over time — rather than one-time buyers who never build any.
You can calculate LTV for an individual customer by totaling their revenue. More useful is cohort LTV: total revenue over a period divided by new customers acquired in it. And this is precisely what's impossible inside an ad manager — it can't track individual customers, so it can't tell you new-customer LTV at all.
That's why the New Customer Cohort Report exists. In 2026, cohort reporting has become essential for measuring the real impact of content and full-funnel marketing, especially as AI-driven ad bidding depends on long-term value signals to optimize well.
A concrete example of what cohort LTV reveals - say a June cohort starts at an average new-customer value of $206 on Day 0 and climbs to $267 nine months later through repeat orders. A September cohort starts at the same $206 — but reaches $303 within just five months, well ahead of where June was at the same point. That gap is the whole insight, something about September's traffic brought in higher-value customers, faster. Identify the channel, audience, or campaign behind it and you can replicate that customer journey deliberately — instead of just counting how many customers you got.
That's the edge over competitors who optimize only for acquiring customers rather than acquiring the right ones.
Metrics used well are a profit lever. Used on bad data — like raw ad-manager numbers — they're worse than useless, because they point you confidently in the wrong direction. A few principles:
- Balance cost and quality when picking a CPL target — cheapest isn't best.
- Let a new campaign run at least one buying cycle before you kill, chill, or scale it.
- Focus on high-LTV customers across the whole funnel — and start at the cold-traffic level, not mid-funnel. Building for lifetime value from the very top is the game-changer.
- Above all, make sure the data feeding every one of these metrics is accurate and de-duplicated across channels. One unbiased source beats five platforms each grading their own work.
Want these metrics running on data you can actually trust? See how it works in the platform overview, or book a demo to see it against your own numbers.
Customer Lifetime Value (LTV), especially paired with a customer acquisition cost focused on winning high-LTV customers rather than just cheap ones. That combination is what makes revenue sustainable and lets you confidently outspend competitors on acquisition.
Ad platforms report on their own performance and often over-claim, taking credit for conversions other channels also claim while relying on last-click data. Multi-touch measurement tracks the full journey across channels and reconciles it against first-party CRM and sales data for an accurate, de-duplicated view — which matters even more now that privacy changes limit platform-side tracking.
A widely used benchmark is 3:1 or higher — a customer worth at least three times what it cost to acquire them, with 5:1+ considered strong. One caveat for ecommerce specifically: DTC brands often run lower than SaaS, roughly 1.5:1 to 3:1, because ecommerce margins are thinner. As always, your real target depends on your margins, not a universal number.