Traffic acquisition cost, user acquisition cost and customer acquisition cost are three different numbers

Traffic acquisition cost, user acquisition cost and customer acquisition cost are three different numbers with three different denominators, and most arguments about "our acquisition cost" are really arguments about which one somebody is quoting. Traffic acquisition cost divides spend by visits or sessions. User acquisition cost divides spend by registered users, installs or accounts. Customer acquisition cost divides spend by paying customers. The three can differ by two orders of magnitude on the same campaign, so the first thing to settle in any conversation about calculating acquisition cost is the denominator.

Google Keyword Planner reports 10 searches a month in Singapore for this group of phrasings, harvested 26 October 2026 against geo 2702. The phrasings include "traffic acquisition cost", "average user acquisition cost" and "calculating acquisition cost", and the reason they look like one query and behave like three is exactly the denominator problem.

The three numbers, with their denominators

Number	Numerator	Denominator	The question it answers
Traffic acquisition cost	What you paid to get visits, including media, affiliate and distribution fees	Visits or sessions in the period	What a visit costs, and whether the channel is affordable before conversion is even involved
User acquisition cost	Spend on campaigns aimed at sign up or install	New registered users, installs or free accounts	What a user costs, which matters when the product is free at the front door
Customer acquisition cost	Sales and marketing cost in the period	New paying customers	What a customer costs, which is the number that sits against lifetime value

Each number is correct for its question and wrong for the other two. A mobile app with a free tier can have a user acquisition cost of a few dollars and a customer acquisition cost in the hundreds, because only a small share of users ever pay. A content site can have a traffic acquisition cost of a few cents and no customer number at all, because its customer is an advertiser rather than a reader.

Traffic acquisition cost has a second meaning, and it is the bigger one

In financial reporting, traffic acquisition cost is not a marketing efficiency metric at all. It is a revenue share line, and the largest published example of it belongs to Alphabet. Its annual report on Form 10-K for the year ended 31 December 2025, filed with the United States Securities and Exchange Commission on 5 February 2026 and read on 28 October 2026, puts it like this:

"We expect traffic acquisition costs ("TAC") paid to our distribution partners and Google Network partners to increase as our advertising revenues grow."

The same filing says TAC "includes: amounts paid to our distribution partners who make available our search access points and other ad-supported services", and reports the figures in millions of US dollars: TAC of $54,900 in 2024 and $59,926 in 2025, against total cost of revenues of $146,306 and $162,535 for the same two years.

The filing also names the ratio that analysts watch, and the reason it moves: "Our overall TAC as a percentage of our advertising revenues ("TAC rate") has been decreasing primarily due to a revenue mix shift from Google Network properties to Google Search and other properties." It then lists what will keep moving it, verbatim: "changes in device mix; geographic mix; partner agreement terms; partner mix; the percentage of queries channeled through paid access points; product mix; the relative revenue growth rates of advertising revenues from different channels; and revenue share terms."

So if you are reading an article about traffic acquisition cost and the numbers are in the tens of billions, you are reading about a platform paying partners for distribution. If the numbers are in cents, you are reading about a marketer dividing media spend by sessions. Same phrase, two trades. Both uses are legitimate and the confusion between them is why the phrase is worth defining before you use it in a board pack.

Calculating each one without flattering it

The arithmetic is trivial. The honesty is in what goes into the numerator and what counts in the denominator.

• Traffic acquisition cost. Add paid media, affiliate and referral fees, any sponsorship or placement payment, and any revenue share with a distribution partner. Divide by sessions from those sources only. The mistake that flatters it most is dividing paid spend by total sessions including organic and direct, which can halve the number without anything changing.
• User acquisition cost. Divide the campaign spend aimed at sign up by new users from those campaigns. The mistake here is counting users who arrived through a route you did not pay for, and the second mistake is counting a user twice because one person registered on two devices.
• Customer acquisition cost. Divide sales and marketing cost by new paying customers in the same period. Our page on the customer acquisition cost formula works through what belongs in that numerator, and cost per customer separates it from cost per action, cost to serve and retention cost.

A worked illustration, with invented figures chosen to make the gap visible. A month of spend of 20,000 produces 40,000 sessions, 2,000 sign ups and 100 paying customers. Traffic acquisition cost is 0.50 a session. User acquisition cost is 10 a user. Customer acquisition cost is 200 a customer. Nobody in that company is lying when one person says fifty cents and another says two hundred.

Why there is no useful average, and what to use instead

People search for an average user acquisition cost because they want a number to judge their own against. Published averages for this exist and they disagree with each other, for a structural reason rather than a sloppy one: the figure depends on the country, the platform, the vertical, the bidding objective and whether a "user" means an install, a registration or an activated account. An average built across those is an average of different things.

Three comparisons are more useful than any benchmark.

1. Your own trend over four quarters. Direction is a real signal, and it is comparable because the definition is yours and constant.
2. The ratio between your three numbers. If traffic cost is flat and user cost is rising, the landing page or the offer is the problem, not the media buying. If user cost is flat and customer cost is rising, the product's free to paid step is the problem.
3. Payback against gross margin. A customer acquisition cost is affordable or not only against what a customer is worth and how long the money takes to come back.

Where traffic acquisition cost is quietly changing shape

The denominator of traffic acquisition cost is the session, and the session is the part of this that search behaviour is currently altering. An answer delivered inside an assistant or an AI overview resolves a question without producing a visit. That does not raise your cost per visit; it reduces the number of visits a given amount of demand produces, which raises the cost of the visits you do get and leaves the influenced demand uncounted in the only metric you were tracking.

Practically, that means a rising traffic acquisition cost needs one more check before it becomes a media buying decision: is the cost per click up, or is the number of clicks down at the same impression volume? The two have different fixes, and only one of them is a bidding problem.

Questions about acquisition cost calculations

Is traffic acquisition cost the same as cost per click?

No. Cost per click is what a platform charges for one click. Traffic acquisition cost is your total cost of acquiring visits divided by the visits, so it includes affiliate fees, placements and revenue shares that never appear in an ad auction.

What is the difference between user acquisition cost and customer acquisition cost?

The denominator. User acquisition cost counts people who signed up, installed or created a free account. Customer acquisition cost counts people who paid. In any freemium product the second is a multiple of the first.

Does traffic acquisition cost include organic traffic?

Only if you also count the cost of producing it, and most companies that do this count content and tooling cost against organic sessions as a separate line rather than mixing it into the paid number. What it must never do is divide paid spend by all sessions, paid and organic together.

What does TAC mean in Alphabet's accounts?

Payments to distribution partners and Google Network partners. Alphabet's Form 10-K for the year ended 31 December 2025, read 28 October 2026, reports TAC of $59,926 million for 2025 and $54,900 million for 2024, and calls TAC as a percentage of advertising revenues the TAC rate.

Which of the three should a small company track?

Customer acquisition cost, because it is the one that meets gross margin. Track traffic acquisition cost beside it as a diagnostic, since it tells you whether a change came from the cost of attention or from what happened after the click.

The spend that produces no session to divide by

All three of these numbers assume a countable event: a visit, a sign up, a payment. The interesting gap in 2026 is the demand that produces none of them, where a buyer asks an assistant for suppliers in your category, is given three names, and contacts one of them directly. There is no session, no campaign and no cost per anything, and whether you were one of the three names is not a metric any analytics property holds. It is a separate measurement with its own denominator, the number of times the question was asked, and AI Knows Us was built to produce it.