---
title: "Marketing"
id: "1778"
type: "page"
slug: "marketing"
published_at: "2026-09-20T21:35:45+00:00"
modified_at: "2026-09-20T21:43:20+00:00"
url: "https://xedant.com/agents/analytics/docs/templates/marketing"
markdown_url: "https://xedant.com/agents/analytics/docs/templates/marketing.md"
excerpt: "This section is about promotion: where the money went, what came back, and what a…"
---

# Marketing

[https://xedant.com/agents/analytics/docs/templates/marketing.md](https://xedant.com/agents/analytics/docs/templates/marketing.md)

This section is about promotion: where the money went, what came back, and what a customer really costs. All eight templates run on the same 12 months of demo data: tagged site visits, form submissions, orders from the accounting system, and an ad spend log. This section covers 8 templates.

Every template is a live walkthrough on demo data: open it in the Templates section of Analytics Agent and press “Use this template” to run it on your own data.

### End-to-End Funnel — Which Source Brought the Revenue

**Shows:** a “spend → visits → leads → orders → revenue” table for every traffic source, built by matching buyer and site profiles; orders that could not be linked to any source get a separate row instead of being smeared across channels.

**You get:** you see where each channel loses people, and how much you can trust the table itself: if some orders are unattributed, it says so plainly.

### ROMI, Both Conventions — Which Tool Returns Money

**Shows:** every channel’s return calculated twice — by revenue (how the ad account counts it) and by profit (what is left for you), with a label on each channel: pays back under both calculations, pays back only on paper, or burns money.

**You get:** a channel that looks profitable in the platform’s report but does not actually return its costs no longer misleads you.

### CAC / CPA / CPL Economics — What a Customer Costs

**Shows:** what you pay for a lead, for a target action, and for an acquired customer — by channel and campaign, with a separate row showing what share of leads become customers; the full cost of a customer sits next to the ad-only cost.

**You get:** it is clear why the cheapest lead does not produce the cheapest customer, and where the budget should really go.

### Attribution — Five Models Compared, One Truth Check

**Shows:** the same sales credited to five different ways of sharing the credit — last click, first click, evenly across touches, and two in-between options; it shows how much a channel’s score changes from the choice of method alone.

**You get:** it is clear that “who we owe the sale to” is your choice of rules, not a measurement — so budget moves are proven by experience, not by switching models.

### Offline & B2B Attribution — Calls, Meetings, Webinars on the Deal Clock

**Shows:** calls, meetings, and webinars linked to companies by phone, email, and domain, and a view of which touches separate won deals from lost ones; it also examines how fair it is to credit one specific manager.

**You get:** a long deal with several people involved stops being a black box: you see at which moment and which step separates success from loss.

### Attribution Quality — Can You Trust It At All

**Shows:** a data check before any channel talk: what share of deals is linked to a source at all, what is wrong with messy tags, what channel metrics look like under the worst and best assumptions about unattributed money, and whether data collection has broken.

**You get:** before changing the budget, you see which numbers you can trust, which are the result of accounting holes, and what to fix first.

### Site Funnel — Where the Site Leaks Visitors

**Shows:** the visitor’s path through the site — visit, product, cart, checkout, payment — by device and source, against usual values; a loss ranking names the steps that lose the most people, and each loss comes with a list of testable hypotheses.

**You get:** site fixes go into the two steps that actually lose buyers, instead of spreading across the whole site.

### A/B Testing — Significance, Power, Peeking

**Shows:** how to read an experiment’s results: comparing variants with a caveat about random noise, calculating what lift your traffic can detect at all, and a walkthrough of what peeking at results before the deadline looks like.

**You get:** the “keep variant B” decision is made on numbers, not on gut feel and not on the first days of the experiment, when the picture is still deceptive.

[← Templates library](/agents/analytics/docs/templates)
 · [Business group](/agents/analytics/docs/templates#wp-block-themeisle-blocks-advanced-heading-biznes)