Marketing KPIs by channel: what to track and what to ignore

Every ad platform, email tool and analytics property ships with a dashboard full of numbers, and most of them are accurate and useless at the same time. An owner who asks “is marketing working” gets a slide of reach, clicks and followers, none of which answers the question.

This guide sorts marketing KPIs by channel into three tiers: the primary KPI each channel should be judged on, the supporting metrics that explain why the primary number moved, and the vanity metric that is safe to drop from the report. It then shows how the numbers chain together from impression to revenue, what a one-page monthly dashboard contains, how the picture changes for ecommerce, B2B lead generation and local services, and how to set targets when you refuse to trust borrowed benchmarks.

Three tiers: primary KPI, supporting metrics, vanity metric

A KPI is only useful if someone would change a decision because of it. That test splits marketing KPIs by channel into three tiers.

• Primary KPI. One number per channel that connects to money: cost per acquisition, return on ad spend, qualified leads, revenue per send. The channel owner is accountable for it, and budget moves toward the channel with the healthier figure.
• Supporting metrics. Two to four numbers that diagnose the primary one. When cost per lead rises, click-through rate, conversion rate and cost per click tell you whether the ad, the page or the auction changed.
• Vanity metric. A number that rises easily, feels good in a meeting and does not predict revenue. Impressions, followers and total page views are the usual suspects.

Vanity metrics are not always wrong to collect. They belong in the platform, where the specialist can use them for troubleshooting, and out of the owner's report, where they crowd out the few lines that matter. A good rule: if a metric has never caused anyone to move budget, it goes to the appendix.

Paid search and paid social: the KPIs that carry budget decisions

Google Ads

• Primary: cost per acquisition for lead generation, or ROAS for stores that pass order values into the account.
• Supporting: cost per click, conversion rate of the landing page, impression share lost to budget or rank, and Quality Score on the phrases that spend most.
• Ignore: total clicks. A campaign can double its clicks by loosening match types and bring in nothing but researchers and job seekers.

Meta ads

• Primary: cost per purchase or cost per qualified lead, measured in your own CRM or store rather than only inside Ads Manager.
• Supporting: CPM, click-through rate on the creative, frequency, and the share of leads your sales team accepts.
• Ignore: reach and post reactions on ads. A video can collect a pile of reactions and never produce a checkout.

On both platforms the reported conversions are the platform's own attribution, which tends to give itself credit generously. Compare the platform figure with what your CRM or order system recorded for the same period. A steady gap is fine as long as you know its size; a gap that suddenly widens is a tracking problem to fix before any budget decision.

SEO, content and organic social: slower KPIs that still need owners

SEO

• Primary: organic conversions (leads, orders, booked calls) from searches that do not contain your company name.
• Supporting: non-brand clicks and impressions in Search Console, the number of pages receiving clicks, and positions for a fixed list of commercial phrases.
• Ignore: a sitewide “visibility score” from a third-party tool, and total organic sessions that are mostly blog readers from unrelated searches.

Content

• Primary: assisted conversions, meaning how often an article appears on the path of someone who later converts, plus newsletter or demo sign-ups started from content.
• Supporting: engaged time on the page, scroll depth, internal clicks to service pages.
• Ignore: raw page views on posts written to chase traffic with no buyer in it.

Organic social

• Primary: visits and conversions from social profiles, tagged with UTM parameters on every link.
• Supporting: engagement rate per post, saves and shares, profile visits.
• Ignore: follower count. It grows with giveaways and shrinks with nothing, and neither tells you about demand.

These channels move slowly, so judge them on quarterly trends rather than month-to-month swings.

Email and the website: the numbers closest to revenue

Email

• Primary: revenue or conversions per email sent (or per thousand sends), which folds list quality, deliverability and message fit into one figure.
• Supporting: click rate, unsubscribe rate, spam complaint rate and list growth from real sign-ups.
• Ignore: open rate. Privacy features in major mail apps load images automatically, so opens are inflated and no longer mean a human read the message.

Website

• Primary: sitewide conversion rate on the goal that matters (purchase, form, call), split by device.
• Supporting: conversion rate of the key landing pages, form abandonment, page load time on mobile, checkout drop-off by step.
• Ignore: bounce rate on its own. A visitor who lands on a contact page, copies the phone number and calls counts as a bounce.

The website is the shared bottleneck for every paid and organic channel, so its KPIs belong at the center of the report, not at the end. A drop in site conversion rate raises the cost per acquisition of every channel at once, and a channel owner will often blame the ads for it.

Reading the chain from impression to revenue

Marketing KPIs by channel make sense only when you read them as links in one chain. Every channel runs a version of the same funnel: impressions become clicks, clicks become leads or carts, leads become customers, customers produce revenue and repeat purchases.

Walk through it with assumed numbers. Say a campaign earns 20,000 impressions, a 2% click-through rate gives 400 clicks, a 5% landing page conversion rate gives 20 leads, and your sales team closes one in four, so 5 customers. Now each lever is visible:

• Double the click-through rate and you get 10 customers without touching the page.
• Raise the page conversion rate to 7.5% (up by half) and you get 7 or 8 customers from the same clicks.
• If the close rate falls to one in eight, the channel looks broken even though the ads did nothing wrong.

This is why a single metric never explains a result. When the primary KPI moves, walk down the chain until you find the stage that changed. For B2B companies the chain has two extra checkpoints: the MQL that marketing hands over and the sales-qualified lead that sales accepts. The gap between those two is often where lead generation budgets leak.

What changes for ecommerce, B2B and local services

A store that judges Meta ads only on first-order ROAS will underfund the campaigns that bring repeat buyers, so pair ROAS with lifetime value once you have a year of order data. A B2B firm should resist the cost per lead race: a channel that produces low-cost leads nobody can sell to is more expensive than one that produces fewer, qualified ones. A local service company often discovers that the weakest link is not marketing at all but missed calls, which is why answered call rate belongs on its dashboard. The ecommerce and SaaS industry pages go further into each model.

What a one-page monthly KPI dashboard should show

The owner's view of marketing KPIs by channel fits on one page. Anything that needs scrolling becomes a report nobody opens by the third month.

1. Headline row: revenue or pipeline attributed to marketing, total marketing spend, and blended cost per acquisition (all spend divided by all new customers). Blended numbers cut through the double counting that platform attribution creates.
2. One row per channel: spend, primary KPI, the same KPI last month and the same month last year, and a one-sentence comment on why it moved.
3. Website row: conversion rate by device and the one page that changed most.
4. Decisions: what shifts next month, with the reason.

The comment column matters more than the numbers. A figure without a reason trains everyone to ignore it. The companion post on what a monthly marketing report should include covers the narrative side of this page.

Setting targets without borrowed benchmarks

Published averages for click-through rate or cost per lead mix companies with different prices, margins and sales cycles, so they tell you little about your own ceiling. Build targets from two sources you control.

Unit economics. Start from the margin on a customer. Say an average customer brings gross profit you can measure over the first year, and you decide marketing may spend up to one third of it to acquire them. That share is your maximum CAC. Divide it by your close rate and you get a maximum cost per lead; divide again by your landing page conversion rate and you have a ceiling for cost per click. Every channel target now comes from your business, not from a chart.

Your own history. Take the last six to twelve months per channel and set the next target as a modest improvement on the trailing average, adjusted for seasonality. If you have no history, run a channel for one full quarter with tracking in place and treat that quarter as the baseline.

Review your marketing KPIs by channel and their targets every quarter. If your process for pulling these numbers together is the weak spot, Marketers.com offers a free marketing analysis delivered as a PDF within 24 hours, after which a senior project manager builds a monthly report around the KPIs above; the how it works page explains the steps.

FAQ