Digital Marketing KPIs: What to Measure at Each Funnel Stage

Brazil alone poured R$ 42.7 billion (about US$ 8 billion) into digital advertising in 2025, up 12.7% in a single year, according to IAB Brasil. Now the uncomfortable part: in HubSpot's State of Marketing 2025 survey, only 35% of marketers say they use data to guide strategy. Put those two numbers together and the picture is stark. Spending has never been higher, and most decisions are still made in the dark.

If you run a marketing operation, you know the symptom. The monthly report has 40 indicators, the dashboard has eight tabs, and the question "should we increase the media budget?" still has no clear answer. The problem is almost never a lack of data. It is an excess of metrics and a shortage of KPIs.

Digital marketing KPIs are the key indicators that show whether your strategy is hitting business goals. Every KPI is a metric, but not every metric is a KPI: a number only becomes a KPI when it is tied to a target and can support a decision. The right ones vary by funnel stage, from CPM at the top to CAC, ROAS, and LTV at the bottom.

What is the difference between a metric and a KPI?

A metric is anything you can measure: likes, impressions, sessions, time on page. A KPI (key performance indicator) is a metric promoted to performance indicator because it is anchored to an objective. If this quarter's goal is lowering acquisition cost, CAC is a KPI. Likes remain metrics: useful for diagnosing creative, irrelevant in the boardroom.

The test is simple. If the number doubled or halved tomorrow, would you do anything differently? If the answer is no, it is a vanity metric. Total impressions, followers, and pageviews almost always fail this test. In practice, bloated reports exist because those metrics grow easily and make the work look successful, even when revenue does not follow.

Which digital marketing KPIs should you track at each funnel stage?

The logic is always the same: each funnel stage has one job, and the KPI measures whether that job is being done at an acceptable cost. A lean list, three indicators per stage, covers the vast majority of operations and fits on a single screen.

Top of funnel: acquisition of attention

The job here is putting the brand in front of the right people at a low cost. CPM (cost per thousand impressions) tells you what it costs to show up; new sessions tell you how many strangers reached your site; cost per session ties the two ends together. Raw reach alone is misleading: 2 million impressions to the wrong audience are worth less than 100 thousand to the right one.

Middle of funnel: consideration

Now the job is turning visitors into identified contacts. The three KPIs: landing page conversion rate (visitors who become leads), CPL (cost per lead), and the percentage of MQLs, the leads that actually fit your buyer profile. A low CPL with a low MQL rate is a classic trap: a cheap form attracts the curious, not buyers.

Bottom of funnel: conversion

This is where you measure money. CAC, what it costs to win a customer (the full calculation is in our guide on how to calculate CAC), ROAS, the revenue generated per dollar spent on media, and close rate, the share of proposals or carts that become sales. If you could only track three numbers in the whole operation, start with these.

Post-sale: retention

The stage most dashboards ignore, and exactly where the margin lives. LTV (the value a customer generates over the relationship), repeat purchase rate, and churn. A US$ 100 CAC is great or terrible depending entirely on LTV: without that pair, the bottom of the funnel means nothing.

The same list as a table, with formulas:

Stage	KPI	Formula	Question it answers
Top	CPM	(spend ÷ impressions) × 1,000	What does it cost to show up?
Middle	CPL	spend ÷ leads generated	What does a contact cost?
Bottom	CAC	(media + team + tools) ÷ new customers	What does a customer cost?
Bottom	ROAS	media revenue ÷ media spend	Does the media pay for itself?
Retention	LTV	margin per order × orders per year × years retained	What is a customer worth?

One operational detail that kills good KPI lists: where the data lives. CPM and CPL sit in the ads manager; conversion rate in GA4; CAC, LTV, and repeat rate require joining media data with your CRM or ERP. When every indicator lives in a different tool, the team spends the first week of the month building the spreadsheet and has little time left for the part that matters, which is interpreting it. Before you settle on the list, decide who consolidates the data and how often.

How does this work in practice?

Take an online fashion retailer spending US$ 30,000 a month on Meta Ads and Google Ads. The campaigns generate 25,000 clicks (US$ 1.20 average CPC) and 300 sales at a US$ 240 average order value. The bottom-of-funnel math: US$ 72,000 in revenue, a 2.4 ROAS, and a US$ 100 CAC.

Looks healthy, until margin enters the picture. At a 40% gross margin, each sale leaves US$ 96. In other words: the first purchase does not cover the US$ 100 CAC. The operation only stands because customers reorder 2.5 times a year on average, generating US$ 240 in annual margin and a first-year LTV/CAC of 2.4. That is the conclusion no vanity metric will ever hand you: this business does not live on ROAS, it lives on repeat purchases. If the repeat rate drops 20%, the same 2.4 ROAS starts destroying value.

Notice what stayed out of the analysis: impressions, likes, followers. Not because they do not exist, but because no decision depended on them. For a deeper look at which media indicators deserve daily attention, we published a dedicated guide on paid traffic metrics.

How many KPIs should you track, and how often?

Five to seven per decision level. The media buyer can check CPC and CTR daily; leadership needs CAC, ROAS, LTV, and revenue weekly or monthly. When both groups stare at the same 40-number dashboard, nobody decides anything. An excess of indicators is a sophisticated way of having no indicators at all.

Worth admitting: this rule breaks for very young operations. In a channel's first months almost everything is a test, and the CPL/CPM pair is enough. Retention KPIs only make sense once the customer base is large enough to measure repeat purchases without statistical noise. Building that panel is the core of any serious marketing data analysis project.

There is real pressure behind this discipline. According to Gartner's 2025 CMO Spend Survey, marketing budgets have flatlined at 7.7% of company revenue. A flat budget means growth is now synonymous with efficiency, and efficiency only shows up in KPIs, never in vanity metrics.

Frequently asked questions

What are the main digital marketing KPIs?

It depends on the funnel stage: CPM and cost per session at the top; CPL and conversion rate in the middle; CAC, ROAS, and close rate at the bottom; LTV, repeat rate, and churn after the sale. A lean operation tracks five to seven of them, chosen by the quarter's goals.

What is the difference between a KPI and a metric?

Every measurement is a metric. A KPI is a metric tied to a business objective and used to make decisions. Sessions are a metric; sessions inside a quarterly target for qualified traffic growth become a KPI.

What are vanity metrics?

Numbers that grow easily and look impressive in reports but cannot support a decision: followers, likes, total impressions, pageviews. The test: if the number changed drastically tomorrow, would anything in the operation change? If not, it is vanity.

How many KPIs should a company track?

Five to seven per decision level. Fewer than that tends to leave blind spots; more than ten dilutes attention and turns the dashboard into scenery.

How do you define KPIs for a sales funnel?

Start from the business objective (revenue, margin, acquisition), identify which funnel stage limits that objective most today, and pick two or three indicators that measure exactly that stage. Review quarterly: a good KPI is one that expires.

Fewer numbers, more decisions

This article's stance is unpopular: track fewer numbers. A funnel with 12 well-chosen KPIs beats a dashboard with 60 metrics, because KPIs exist to force decisions, and decisions require focus. If your struggle is not choosing the indicators but pulling them together (Meta Ads on one screen, GA4 on another, the CRM in a spreadsheet), that is exactly the problem Sherlok solves: it connects your data sources and answers plain-language questions like "what was my CAC by channel this month?", no spreadsheet or on-call analyst required.