Cheap Reach, Expensive Conversions
CPM is often used as the key gatekeeping metric when selecting creators for brand partnerships. So let’s take a look at where CPM and CPA part ways across 2,656 YouTube sponsorships.
Spoiler alert: a great CPM is not a conversion strategy.
EXECUTIVE SUMMARYA Top-Decile CPM Tells You Nothing About CPA
If you buy YouTube influencer inventory on CPM alone, you’re not likely to get great acquisition outcomes. After evaluating over 2,600 YouTube creator brand partnerships, here’s what the data says you actually get when you choose low CPM creators:
across the 266 placements in the top decile of reach efficiency (read: best 10% from a CPM perspective), conversion efficiency spans nearly the entire 0-to-100 range -- from 1.4 to 99.6. This means that the best shows from a CPM perspective are all over the place when it comes to conversion efficiency. Forty-two of those 266 brand partnerships (about 1 in 6) land in the bottom quartile of conversion efficiency. A great CPM is consistent, but the business outcomes vary significantly.
Cheap reach does not buy cheap conversions. Across 2,656 scored placements, the correlation between a brand partnership’s CPM efficiency score and its CPA efficiency score is just 0.19. In other words, a low CPM explains under 4% of the variance in conversion efficiency.
This paper is for performance-minded brand marketers who screen YouTube inventory on CPM -- and it argues for a simple fix: do not screen on CPM alone. Pair it with a CPA floor.
Analysis set: 2,656 YouTube sponsored placements live March 2024 – July 2026 from The Outloud Group, each scored 0–100 on reach efficiency (CPM) and conversion efficiency (CPA). 578.8M cumulative views. All placements anonymized.
SECTION 01THE CPM PROXY PROBLEM
You don't work in influencer marketing if you've never been asked to estimate a CPM. It's the industry's favorite screen: fast to compute, easy to compare, and comfortingly familiar from paid media. The unspoken assumption is that CPM is a proxy for efficient results: a brand partnership delivering cheap reach will also deliver cheap outcomes.
Not so fast.
CPM tells you what you paid for attention. It tells you nothing about whether that attention belongs to people who will ever buy your product. For awareness-first brands, often broad-market retail and e-commerce plays where reach is the outcome, a CPM screen is a reasonable tool. For everyone whose media plan is judged on actual conversions, it's a half-assed measure that feels like rigor. We tested the proxy assumption against 2,656 YouTube sponsored placements from March 2024 through July 2026, each scored 0–100 on two independent axes -- CPM efficiency (how cheaply it bought human-verified views) and CPA efficiency (how cheaply it bought conversions).
We’re Not The Only Ones Saying It
SECTION 02INSIDE THE REACH-ELITE COHORT
Start at the very top of the reach curve. If CPM predicted CPA, the most reach-efficient placements in the book would cluster near the top on conversion too. They don't. Among the 30 placements with CPM scores of 95+, CPA efficiency runs from 11.6 to 99.0 -- and roughly one in three sit below the book's median. Widen to the statistically sturdier top decile (top 10% of content), where we see 266 placements with CPM scores of 84.8 or better -- and the pattern holds: CPA scores span 1.4 to 99.6, 36% fall below the median, and 42 of 266 (1 in 6) land in the bottom CPA quartile.
Here's what that means in dollars. Within the same top-decile-CPM cohort, placements that also ranked top-quartile on conversion paid a 9x lower median CPA. Those that ranked bottom-quartile paid 9x higher at an essentially identical median CPM ($11.46 vs. $11.84). Two buyers, same reach efficiency, 9x apart on what a customer actually cost. The inverse holds, too:
SECTIONS 03 / 04CPM VARIES BY VERTICAL BUT THE STORM ON CPM VS. CPA REMAINS
Some CPM variance is structural: it depends on where you're shopping. Across 10 content verticals with at least 80 brand partnerships analyzed, median CPM efficiency ranges from $34 to $68 -- a 2x spread. Automotive content is the most efficient and Health & Fitness is the most expensive from a CPM perspective. But vertical choice doesn't solve for CPM being a proxy for CPA either because impressions, wherever you buy them, are still only impressions:
Median Efficiency by Vertical: Reach Vs Conversion
Entertainment is one of the most efficient verticals in the book from a CPM perspective. With a median CPM score 67.6, far above the 49.5 overall median. Its median CPA score? 44.0 -- below the conversion median of 47.3. A whole category that looks like a bargain on the reach screen and underperforms on the conversion screen.
Now invert it. Health & Fitness posts the lowest median CPM efficiency of any major vertical (34.0), but it has the highest median CPA efficiency (59.1). What appears to be the most "expensive" reach turns out to deliver the lowest-cost customers. Screen it out on CPM, and you screened out your best converters. This is the danger of screening creators on a CPM basis.
Sections 05 / 06Why CPM Screens Mislead and A BETTER WAY to Buy
CPM rewards low upstream prices and a big view count. Neither has anything to do with whether the audience is in market. It is true that a brand partnership can win on CPM because the creator is undervalued, because the content travels algorithmically, or because the audience is broad and passive. Notably, only the first is unambiguously good news for a performance focused buyer.
Audience intent, offer fit, and a creator's power to move people from watching to doing are invisible to a CPM screen. That's why the correlation between CPM (reach) & CPA (conversion) is nominal: at 0.19, knowing a placement's reach efficiency tells you almost nothing about its conversion efficiency.
To be clear, this is not an argument against CPM. It's an argument against CPM alone. The fix is a pairing framework:
Methods2,689 YouTube sponsored placements live between March 1, 2024 and July 14, 2026 from The Outloud Group's sponsored placements; 2,656 carry both a CPM efficiency score and a CPA efficiency score and form the analysis set (578.8M cumulative views to date). Placements are anonymized; no creator or channel is identified.
CPM and CPA efficiency scores are indexed 0–100 against the book, higher = more efficient. Book-wide medians: CPM score 49.5, CPA score 47.3. Bottom-quartile CPA threshold: score < 25.9. Top-decile CPM threshold: score ≥ 84.8 (n=266). Dollar figures are medians of upstream CPM / upstream CPA within the named cohorts; book-wide medians are $52.88 (CPM) and $750 (CPA).
Pearson r between CPM and CPA scores across the analysis set = 0.186 (r² = 0.035).
Vertical labels normalized for casing and synonyms (e.g., "vehicles" / "Auto"). The 10 normalized verticals with n ≥ 80 scored placements were used for category-level medians.
Third Party Sources1. Brainlabs — "The Metrics Killing Your Programmatic Strategy" — brainlabsdigital.com/the-metrics-killing-your-programmatic-strategy
2. Marketing Dive — "Creators work, but measurement doesn't yet" — marketingdive.com/news/creators-work-but-measurement-doesnt-yet/813952
3. Sagum — "The Benchmarking Trap" (June 2026) — sagum.com/2026/06/20/the-benchmarking-trap
4. Digiday — "OpenAI is taking a (very) cautious approach to the narrative around its ChatGPT ads test" — digiday.com/marketing
5. Exverus — "Paid Media Metrics and Benchmarks: The Complete Playbook" — exverus.com/post/paid-media-metrics-and-benchmarks-the-complete-playbook
6. AdNews — "Why ignoring media's most valuable currency will break advertising" — adnews.com.au/opinion
7. 369 — "Metrics That Mislead When Viewed in Isolation" — 369.ad/blog/metrics-that-mislead-when-viewed-in-isolation
8. MonetizeMore — "Why ad revenues plummeted: GDPR & TCF 2.0" — monetizemore.com/blog/why-ad-revenues-plummeted-gdpr-tcf-2-0
The Outloud Group has been matching brands with creators for over 18 years -- and we don’t stop at surface metrics. If you want your influencer program screened on both axes -- reach efficiency and conversion evidence -- let’s talk.
