The most common question about advertising is also the one most poorly answered online. Search for what Facebook ads cost in the US and you will find a single average cost per click presented as though it means something, usually pulled from a dataset mixing insurance, e-commerce, dentistry and restaurants together. That number is arithmetically correct and practically useless, because a click on a mass tort keyword and a click on a taco truck keyword are not participating in the same auction. What follows is a more honest framing: what actually determines your costs in the US market, why the published averages mislead, and which numbers are worth building a budget around.
You Are Not Paying a Price, You Are Winning an Auction
Both Google and Meta price advertising through a real-time auction, which means there is no rate card and no fixed cost. What you pay is a function of how many other advertisers want the same person at the same moment, and how relevant the platform judges your ad to be. Two businesses in the same city advertising the same service can pay materially different amounts on the same day.
Competition in your specific auction
The dominant factor by a wide margin. This is why legal, insurance and financial keywords are expensive and why niche B2B terms can be cheap despite being valuable. It has little to do with your industry in the abstract and everything to do with who else is bidding.
Ad relevance and engagement
Platforms discount advertisers whose ads people actually engage with, because engaged users keep using the platform. A genuinely good ad costs less to deliver than a poor one targeting the same audience. This is the one cost lever entirely within your control.
Audience value to other advertisers
Reaching an audience that many advertisers want — high income, in-market for a major purchase, in a large metro — costs more than reaching one nobody is competing for.
What you asked the platform to optimize for
Optimizing for purchases costs more per result than optimizing for link clicks, because the platform is doing harder work to find people who will complete the action. Comparing costs across different objectives is meaningless.
Geography inside the US
Major metros cost considerably more than smaller markets. National campaigns average this out and hide it, which is why national-level benchmarks are so misleading for local businesses.
💡 The most useful thing to understand: the US is among the most expensive advertising markets in the world, because it has the highest concentration of advertisers competing for high-purchasing-power consumers. If you are comparing US costs to numbers you saw for another market, the gap is real and is not a sign you are doing something wrong.
Why Industry Averages Break Down
Published benchmarks group businesses by industry label, and industry labels hide the variables that actually determine cost. Three examples of how the same "industry average" covers wildly different realities.
- 1Legal spans an enormous range — Personal injury and mass tort keywords are among the most expensive in all of search. Estate planning and traffic defense are a fraction of that. A single "legal industry average CPC" describes neither.
- 2Home services swing with the weather — An HVAC keyword during the first heat wave and the same keyword in October are effectively different products. An annual average describes a month that never happens.
- 3E-commerce depends entirely on product price — A brand selling $30 items and one selling $3,000 items have completely different tolerable costs per click, and averaging them produces a number neither can use.
- 4Local and national are different auctions — A benchmark that mixes national brands with neighborhood businesses reflects the national brands, because they spend more and therefore weight the average.
- 5The useful benchmark is your own — After thirty days you will have real data from your own market, your own creative and your own auction. That number is worth more than any published figure.
The Seasonal Pattern Is Predictable and Expensive
US ad costs are not stable across the year. The fourth quarter in particular reshapes pricing for every advertiser, including those who have nothing to do with retail.
November and December — the most expensive weeks of the year
Retail advertisers flood the auction ahead of the holidays and drive costs up across the board. A service business with no seasonal element still pays more in Q4 simply because it is bidding against retail budgets.
January — costs fall, intent rises
Retail spend withdraws sharply after the holidays while consumer intent in categories like fitness, education, home improvement and professional services is at its annual peak. Frequently the best value window of the year.
Late summer — a genuine lull
Attention and competition both dip in much of the US during the summer holiday period. Cheaper reach, though conversion rates often soften alongside it.
Election cycles
Political advertising consumes enormous inventory in US election years, particularly in contested states during the final months. If you advertise in a battleground state, budget for higher costs in that window.
Minimum Viable Budget Is a Real Constraint
There is a floor below which campaigns do not work properly, and it has nothing to do with getting fewer results. Both platforms need a certain volume of conversion events to learn who to show your ads to. Below that threshold the algorithm never stabilizes and performance stays erratic regardless of how good the campaign is.
- 1Meta needs roughly fifty conversion events per week per ad set to exit the learning phase — Work backwards from your cost per result to see what that implies for your budget. If your target action is too expensive to hit that volume, optimize for a cheaper upstream action instead.
- 2Spreading a small budget across many campaigns is the classic mistake — Five underfunded ad sets all stuck in learning perform worse than one properly funded ad set. Consolidate before you expand.
- 3Expensive categories have higher floors — Legal, insurance and competitive home services genuinely cannot be tested meaningfully on a small budget in a major metro. That is a real constraint, not a failure of effort.
- 4Give it at least thirty days — Weekly performance in a low-volume account is mostly noise. Judging a campaign after five days leads to changes that reset learning and make everything worse.
Track These Instead of Cost Per Click
Cost per click is the metric everyone asks about and one of the least informative available. A campaign with expensive clicks and a strong conversion rate beats a campaign with cheap clicks that go nowhere, every time.
Cost per qualified lead
Not cost per form fill. Strip out wrong-area, wrong-service and spam submissions first. Many businesses discover a meaningful share of their "leads" were never prospects.
Lead-to-customer conversion rate
Owned by your sales or intake process, not your ad account, and usually the largest single lever on profitability. Improving it changes what you can afford to bid.
Cost per acquisition against customer value
The only figure that determines whether advertising is working. A high cost per acquisition is fine if the customer is worth many multiples of it.
Repeat and lifetime value
Businesses with strong repeat custom can afford to lose money on first purchase. Those judging on first-order profit alone systematically underbid and lose auctions they could win.
Response time
Rarely tracked and frequently decisive. In most lead-generation categories, speed of first contact affects conversion more than anything happening inside the ad account.
Common Mistakes
- 1Planning a budget from a published industry average — It describes a blend of businesses that do not resemble yours.
- 2Comparing costs across different objectives — Cost per link click and cost per purchase are not comparable numbers.
- 3Cutting budget in January because Q4 was expensive — January is frequently the cheapest high-intent window of the year.
- 4Testing an expensive category on a token budget — Below the learning threshold you learn nothing, and conclude the channel does not work.
- 5Optimizing for cheap clicks — Cheap traffic that does not convert is more expensive than costly traffic that does.
Stop Guessing What Your Campaigns Should Cost
Adyft builds the audience, writes the ad copy, sets the budget, and launches your Google, Facebook and Instagram campaigns — then tracks what each one actually costs you. Used by businesses in 163 countries, from single locations to multi-market groups and agencies. Plans from $49/month, 14-day free trial.
Start Free Trial