How to calculate lead cost when applications include spam

Lead cost in Metrica’s expenses report is a CPA, and spam is not subtracted from it. The CPA and ROI formulas divide spend by goal visits, goal visitors, goal completions, or purchases, and subtract spend from revenue or goal value. There is no separate term for “spam,” “off-target lead,” or “invalid click” in those fractions. On the Direct glossary, the Metrica goals page, the performance-statistics guide, and the expenses report, the abbreviation CPL is not defined: the price of an action is called CPA. Those four pages were checked, not the whole help.

How to divide spend by leads when you set the denominator yourself is covered in the article on CPA. Here the denominator comes from the report, and it is not “leads without spam.”

Four denominators, and none is called a clean lead

In the “Sources, expenses, and ROI” report, spend is the sum without VAT from the file you passed. CPC is labeled as the cost of one click.

A goal has to exist before a goal CPA. The formulas are fractions with no separate subtraction of spam, robots, or invalid clicks. For the Market source, amounts arrive in conventional units, 1 unit = 30 ₽, and can be updated for 90 days.

The goals help defines conversion as the ratio of goal visits to all visits. Visitor conversion is the share of unique visitors who reached the goal. In one goal visit the same goal can complete several times. Completions per visitor is the average number of completions per unique goal visitor. Goal revenue is set in the interface as a static amount or passed with the price or order_price parameter.

Conversion is not the number of leads: sending the form again in the same visit raises completions, but not the number of goal visits. A teaching example, not a norm. Spend of 40,000 ₽, 20 goal visits, and 50 goal completions inside those visits. Visit CPA is 2,000 ₽, completion CPA is 800 ₽. If you yourself treat 10 of the 50 completions as spam, the report does not know: the fraction has no subtraction. The 800 ₽ ratio stays while the denominator is every completion.

Goal metrics are in most Metrica reports, and there is a separate conversions report. Goal types: a goal event via JavaScript, API, or a file; number of page views; page visits; a click on email; a click on a phone; a click on a button; a form submit; a move to a messenger; a move to a social network; a file download; site search; a return from a payment system; calls; chats; a multi-goal; and a compound goal of sequential steps. Metrica creates some goals itself. The page does not set a normal conversion and does not subtract spam from completions of the “form submit” goal. The permission “use goals in Direct campaigns” allows optimization without access to the counter, but it is not enough for retargeting and audience selection.

Revenue in a cohort does not promise a fraud deduction either

For the Retention and Rolling Retention table you can choose users, visits, view depth, time on site, goal completions, and revenue per user. The “Users” metric is not available in cohort analysis. The grouping is ready-made or your own. A cell holds two numbers: a percent or an average on top, and the number of users or the metric sum below. “Revenue per user” is calculated from the goal or commerce revenue you passed. The page does not say Direct fraud has already been subtracted from it.

The help suggests looking at “good” leads through a CRM upload. The upload adds clients and orders to the statistics: tie sales and deal statuses to the on-site history, optimize Direct toward redeemed orders, signed contracts, and payment after a lead, and build segments from offline data, including lookalikes and people who did not visit the site before the purchase. Separately, it is how you look at “good” leads, because not every form fill becomes a payment. A “good lead” here is one whose status was passed from the CRM, not an automatic phone check.

ClientId for forms and orders on the site is taken with getClientID and written into a hidden field. The page asks you to look at the share of linked conversions in the counter settings on the “Data upload” page: linkage statuses in offline-conversion reports, it says, can display incorrectly at the moment.

CTR does not clean lead cost: clicks over impressions do not subtract spam from form completions. The share of spend on clicks with no buyer is counted outside CPA, as in budget loss. Robot share in a report is not a “without spam” denominator either — what the percent means is in the article on robot share.

Counting cost with spam in mind takes two layers. The report gives spend over all goal visits, visitors, completions, or purchases. A status from the CRM says which fill became a payment or a contract. The expenses page cannot subtract one from the other: there is no separate spam term in the fraction.