An EMARKETER analysis published today contains a remarkable statistic: social media captures nearly 30% of all U.S. advertising dollars while accounting for only 12.5% of adults’ daily media time. And they expect that imbalance to grow. [1]
How can businesses justify such an imbalance, and yet keep spending?
One big reason, because it happened to us, is they are seduced by the metrics. Social platforms give marketers exquisitely detailed reports showing impressions, clicks, conversions—and especially ROAS (Return on Ad Spend).
There’s just one problem: ROAS isn’t revenue. And paid media doesn’t operate in a vacuum: it often gets credit for demand that was created elsewhere—through reputation, word of mouth, earned media, search, email, or prior brand exposure.
We encountered this disconnect firsthand while helping scale SommSelect from $0 to roughly $20 million in ARR. We were significant buyers of paid social advertising and our campaign reports routinely showed ROAS approaching 600%. Who wouldn’t keep buying advertising that reportedly returns six dollars for every dollar spent?
Except something wasn’t adding up. Our actual revenues weren’t increasing at anything near the rate suggested by the advertising dashboards.
Eventually, we conducted the simplest and most revealing attribution test imaginable: we turned the ads off. For an entire month. And revenue was essentially unaffected. What did go down was our advertising expense, along with the substantial cost of continually producing fresh creative to feed the advertising machine.
The lesson wasn’t that advertising doesn’t work. It was that we had confused attributed revenue with incremental revenue. We began asking a different question: how much demand were the ads actually creating, versus merely taking credit for demand our other marketing had already created?
Social platforms operate enormous attribution ecosystems that often claim credit for customers who may have purchased anyway—either because they had previously visited the website, or were already familiar with the brand, or received an email (we sent daily email offers), or searched Google, or encountered myriad other marketing touchpoints. We trusted the metrics that the ad platforms provided… until we didn’t.
And there is solid legal evidence to be suspicious, as well as good reason to independently verify platform-generated numbers with your own numbers. After repeated measurement problems at major platforms, advertisers and the Association of National Advertisers pushed social networks toward independent Media Rating Council audits, precisely because platforms were operating as “walled gardens” in which the seller of the media was also effectively grading its own performance.
Facebook agreed to a $40 million settlement in litigation alleging that it had inflated video-viewing metrics. Earlier, Facebook acknowledged that one average viewing-time metric had been overstated by roughly 60% to 80%.[2] In a separate case, advertisers alleged that Meta’s “Potential Reach” metric overstated audiences by counting accounts—including duplicate and fake accounts—rather than actual people. In 2024, a federal appeals court allowed the advertiser damages class to proceed.[3]
LinkedIn has had similar problems. In 2020, the company disclosed two measurement issues that affected more than 418,000 advertisers. A subsequent advertiser lawsuit alleging overcharges was settled for $6.625 million in 2024, with LinkedIn agreeing to independent auditing of certain advertising metrics while denying wrongdoing.[4][5]
None of this means marketers should abandon paid social. It means the metrics you use to decide your spend shouldn’t come exclusively from the company selling you the advertising.
Lead with Earned; Scale with Paid
Ultimately, we learned at SommSelect that our best marketing mix included a solid base of organic earned media with some paid ad spend on top. And more importantly, in that sequence: we built credibility first with earned media, then bought amplification with ads.
Results from Edelman’s 2026 brand research recommends exactly that—“lead with earned, scale with paid.” The research found that 88% of consumers consider trust the most important or critical factor in choosing a brand.[6] Advertising is very good at buying distribution. Earned media, customer advocacy and organic discovery are what give that distribution something credible to talk about and amplify.
And most importantly, make your marketing mix decisions based on your own metrics and not solely upon reported numbers from the ad platforms. Sometimes the most valuable business metric is the one your ad platform can’t calculate.
Sources:
- EMARKETER, “Social media’s audience share continues to be at odds with how marketers spend on the channel,” August 24, 2026.
- The Guardian, “Facebook inflated video viewing times for two years,” September 23, 2016; Law360, “Facebook Cuts $40M Deal To End Suit Over Video Ad Metrics,” October 7, 2019.
- Reuters, “Meta Platforms must face advertisers’ class action, US appeals court says,” March 21, 2024; U.S. Court of Appeals for the Ninth Circuit, DZ Reserve v. Meta Platforms.
- LinkedIn, “We discovered two measurement issues. Here’s how we’re making it right,” November 12, 2020.
- Reuters, “Microsoft’s LinkedIn settles advertisers’ lawsuit over alleged overcharges,” July 26, 2024.
- Edelman Trust Institute, 2026 Edelman Trust Barometer Special Report: Brand Growth in an Insular World, June 22, 2026.