CAC inflation is real, and the statistic everyone quotes to prove it was published in 2022. The claim is that customer acquisition cost rose 222% over eight years, and that brands now lose $29 on every new customer they buy. It came from SimplicityDX on 19 July 2022, describing the stretch from 2013 to 2022. No sample size was published. No methodology, no account of how anything was collected, and no list of the brands measured. The company sells social commerce technology. Four years later the figure still circulates as though it described this year.
This page is for anybody being handed a rising-cost story that nobody in the room has checked.
Where the most-quoted CAC inflation figure comes from
One press release, from a company with a product to sell.
SimplicityDX announced on 19 July 2022 that brands were losing a record $29 for each new customer acquired. Acquisition cost, it said, had climbed 222% across the eight years ending in 2022.
Nothing sits underneath that. There is no stated sample, no described method, no geography, and no indication of which brands or categories were counted.
A vendor publishing its own research is normal and often useful. Publishing it without a method is what makes the result impossible to check, and impossible to check is what most of this field runs on.
What four years does to a claim like that
Auction prices do not hold still, so a figure describing 2013 to 2022 cannot describe what you will pay in 2026.
Everything downstream has moved since. Conversion rates, average order values, tracking on iOS, consent rates across Europe, and the mix of platforms people actually buy on.
A second problem is worse, and it is a problem of retelling. A percentage measured up to a fixed year gets re-quoted annually without the end date attached, so the same historic rise appears to be happening again every twelve months.
I have watched that happen with cost per lead figures in my own field. A benchmark from one study gets lifted into a blog post, then into a deck, then into a client conversation, and by the third hop nobody can name the year. Most CAC inflation commentary you will read is that game of telephone with a percentage sign on the end.
The dated evidence for CAC inflation that does exist
Meta publishes a real one every quarter, and hardly anybody cites it.
Its own investor reporting put average price per ad up 12% year on year in both the first and the second quarter of 2026. That figure is audited, dated, and describes an entire platform rather than an unnamed panel.
Twelve percent in a year is a very different picture from the one left by that 222% headline, which covered eight years and stopped in 2022.
Gupta Media’s social CPM tracker gives a second dated series. Its 2025 platform averages put Meta at $8.19 per thousand impressions, Snapchat at $8.60, YouTube at $4.99 and TikTok at $4.82, on a tracker running tens of billions of impressions as a 14-day trailing average.
Both sources name themselves, name the period, and describe what they measured. Neither needs a headline percentage to be useful.
CAC inflation in the auction is not CAC inflation in your account
Media price is one input of three, and it is the only one most people watch.
Your acquisition cost is the price of an impression, multiplied by how many impressions a sale takes, divided by how well the offer converts. CPM inflation of 12% lands as nothing at all when conversion rate improves by the same amount.
My clearest example is a US auto transport client on Meta. That account produced 5,599 leads at $2.68 each and held there while spend scaled, having previously sat above $6.
Nothing about the auction got cheaper while that happened. The account simply got better at converting what it was already buying.
For contrast on the published side, WordStream’s 2026 study of 13,474 US search campaigns put the all-industry average cost per lead at $66.69, covering April 2025 to March 2026. Sample size stated, window stated. That is what a usable figure looks like.
What my own accounts can and cannot prove about CAC inflation
They cannot prove it, and pretending otherwise would be dishonest.
I can show you €14.18 across 1,715 orders from an EU fashion ecommerce client, and $5.05 across more than 63,000 purchases from a US electric mobility client. Both cover the full year 2022. I can also show you a US flag and banner ecommerce client between June and August 2026, at $53.72 per purchase excluding one failed test and $67.24 including it.
Line those up and it looks like a steep rise. It is nothing of the kind.
Different clients, different countries, different categories, different price points, and largely different channels. Two of them leaned heavily on Google Search, while the 2026 account is Meta only. Comparing them describes three businesses rather than a market.
Honest evidence here needs one account, one product and one channel measured over years. Few agencies keep a client long enough to have it.
Q4 is where the increase is undeniable
One stretch of the calendar behaves exactly as the alarming headlines describe.
Gupta Media recorded a Meta cost per thousand of $17.70 on Cyber Monday, 138% above the 2024 annualised average. Thanksgiving week averaged $13.42 against a 2025 Meta platform average of $8.19. Those figures come from its social CPM tracker, published in October 2025.
So the same customer costs roughly twice as much to reach in late November as in October. Seasonal CAC inflation is the one version I can point to in every account I run. Margin usually moves the other way in the same fortnight, because that is when the discounting happens.
Plan the November number in August. A target set on October economics gets missed in the last week of November, and the miss almost always gets blamed on the creative.
What actually pushes your acquisition cost up
Five things, and the auction is only one of them.
Competition raises the clearing price, which you cannot control. Discounting cuts contribution, which changes what you can afford to pay. Gross margin does the same and drifts with supplier pricing. Conversion rate on your own site decides how much traffic you need to buy. Your own scaling raises frequency, which raises cost while the reporting still shows growth.
Across my ecommerce clients gross margin runs between 20% and 50%. A store at the bottom of that range feels every auction increase roughly twice as hard as a store at the top.
Four of those five sit inside your business rather than inside Meta. Which is fortunate, because those four are the ones you can change this quarter. Treating CAC inflation as something done to you from outside quietly writes off most of the picture.
How I answer CAC inflation inside an account
By adding surface area, not by pushing more money through what is already running.
One client went from roughly $10,000 to $50,000 a month with me, and none of that came from raising the Meta budget alone. I added campaign objectives, added awareness campaigns, built custom audiences and lookalikes, and brought in channels the client was not using, including TikTok and Klaviyo.
Klaviyo deserves a sentence of its own. Klaviyo is email rather than an ad platform, so adding it meant adding owned media, where the next send costs almost nothing.
That shifts the blended arithmetic instead of the auction price. A second order produced by email is not exposed to rising media prices at all.
Cutting losers matters as much as adding channels. One ad set in that flag account spent about $1,027 for three purchases at roughly $341 each on a 0.25 return, and it had the highest reach and the most link clicks in the entire account.
What to change this week
Three things, and none of them takes a day.
Check the date on every cost benchmark sitting in your planning documents, then delete any that does not carry one. A statistic with no year attached is a rumour with a decimal point.
Next, build your own trend from your own account, same channel and same product, quarter by quarter. Your measured CAC inflation rate is the only one that actually describes your business.
Then look hard at the two inputs you control before arguing about media prices: conversion rate on the page, and contribution per order. Both move your acquisition cost further than a 12% change in delivery cost will.
Why lead costs keep climbing runs the same argument on the lead generation side. What an online order really costs me at volume gives two dated figures with their sample sizes attached. How the calculation should be put together covers what belongs in the numerator, and how long the money takes to come back turns the same figure into a calendar. If you want the whole account examined rather than one metric, here is how I work. To have your own trend checked before you plan around it, book a teardown.