What's happening to Facebook Ads right now?

Summary

Ben Heath argues that Meta advertising is not “dead,” but has become more expensive and competitive. Rising costs reflect greater advertiser demand, limited room to expand ad impressions, inflation, and improvements in advertiser and platform effectiveness.

In his view, many advertisers are still getting results. If they were not, spending would fall and CPMs would decline. The challenge is to improve execution—not assume another platform offers a cheaper escape.

Economics and measurement

• CPMs are rising, but Heath cautions against judging campaigns by CPM alone. More important measures include return on ad spend (ROAS) and cost per conversion, considered alongside customer value and conversion quality.
• Higher CPMs can be justified if ads reach more valuable prospects or generate conversions more efficiently. A low CPM is not automatically good, and a high CPM is not automatically bad.
• Meta’s ad auction reflects advertiser demand versus impression supply. Heath expects costs to keep rising, but believes there is a ceiling: if advertising stops being viable, advertisers will reduce spending or leave.
• He argues that the average advertiser may be only slightly profitable at market equilibrium. Advertisers who perform better than average can gain a stronger advantage.
• Heath sees few obvious alternatives. Costs are also rising across other major advertising platforms, so businesses need to learn to compete in the channels where their customers’ attention is concentrated.

Practical playbook for competing

Heath recommends avoiding decisions based on a handful of conversions or a single poor day. Campaigns need enough conversion volume to provide useful evidence, and lower-budget campaigns may take much longer to produce it. He also encourages advertisers to treat tests objectively: a failed experiment can still provide useful information, and yesterday’s results should not dictate long-term decisions.

Rather than overfocusing on low-impact details—such as the number of ads in an ad set—advertisers should prioritize larger opportunities:

• Test partnership ads with suitable creators.
• Use value rules to guide Meta toward outcomes the business values.
• Improve bidding and optimization choices.

Other recommendations include:

• Reset acquisition-cost expectations. The best advertisers do not necessarily have the lowest cost per customer; they may accept higher acquisition costs because customers are worth more over time.
• Clarify what makes the business meaningfully different. Ads that imitate competitors are unlikely to work at scale if the offer gives customers no compelling reason to choose it.
• Do the difficult work competitors avoid. This might mean learning new AI tools for creative production, rebuilding an offer, testing price points over time, or recruiting and testing creators.
• Balance short-term performance with long-term learning. Testing may temporarily weaken immediate results, but it can uncover better creative, offers, or pricing that support future growth. The right balance between testing and scaling depends on the business.
• Tailor advice to the business. A tactic that works for e-commerce may not suit lead generation. Assess whether a strategy fits the company’s model, economics, and circumstances before implementing it.
• Keep learning. Platform features and best practices change quickly, and tactics that once worked can become outdated if advertisers and agencies stop adapting.

Metrics, examples, and timelines

• Heath says he has spent more than $300 million on Facebook and Instagram ads over 13 years.
• He cites advertisers seeing a $100 CPM in California in a lucrative niche and a $30 cost per purchase. He argues that these figures should be evaluated against the business’s economics rather than treated as automatically excessive.
• He contrasts Starbucks’ roughly $7 average order value with an estimated customer lifetime value of more than $14,000 to illustrate why a business may rationally tolerate a high acquisition cost.
• Heath says his Meta Ads Academy community reached about 1,000 members in just under eight weeks and ranked among the top 80 measured communities by recurring revenue at the time of recording. He attributes its differentiation to his personal brand, track record, case studies, and proof of results.
• In one example, a single ad recorded in about two minutes could work for Heath because of his established credibility, but might not work for a business without comparable differentiation.
• He recommends allowing for a potentially six-month price-testing process to understand how different price points affect conversion and profitability.
• Illustrative testing and scaling allocations include 20% testing / 80% scaling or 50% / 50%. Heath says the appropriate balance depends on performance and ambition.
• He describes creator testing as potentially turning roughly 3× ROAS into 8× ROAS when the right creator is found, while emphasizing the effort and iteration involved.

Presenters and sources

• Presenter: Ben Heath.
• Sponsor: HubSpot, which provided a free pack of 13 AI prompts for Meta ad creative. Heath says the prompts can help generate creative angles, UGC-style scripts, partnership-ad briefs, and signals of creative fatigue.
• Other references and examples: Meta/Facebook and Instagram, Google, TikTok, LinkedIn, Starbucks, Tim Ferriss, and Meta’s Andromeda update.