How much do Facebook and Instagram ads cost in 2026?

Ask three agencies what Meta ads cost and you will get three numbers, all of them defensible. So instead of handing you one more average, this piece shows where the numbers come from, which of them somebody actually measured, and why 2026 is the first year in a while when it pays to ask about Facebook and Instagram separately.

You ask about price. You are buying attention

Meta does not sell advertising space, it runs an auction. You enter with a budget, and it puts every other advertiser chasing the same person in the same second up against you. The price is not a rate card entry, it gets created from scratch every time.

Five things move it: your industry, the campaign objective, how narrow the audience is, the time of year, and the quality of the creative. That last one carries the most weight and gets the least attention. Meta measures whether people linger on your ad and prices you accordingly. A boring advertiser pays a surcharge that never appears itemised anywhere.

Facebook is getting pricier. And smaller

This is the finding that made the whole article worth rewriting. Tinuiti, which manages over $4 billion in annual ad spend, split Meta in two in its Q2 2026 benchmark, and the two halves point in opposite directions.

Read that first row again. Advertisers put 7% more money into Facebook and got 5% fewer impressions for it. The gap went into price, which jumped 13%. That is growth driven by the price tag.

Meta confirms the direction in its own investor release: Q2 2026 average price per ad up 12% year on year, on 14% more impressions delivered, with advertising revenue of $59.4 billion, up 27%. The auction is filling up and Meta lives off exactly that.

Instagram is holding its price. Because it built new inventory

The second row is the more interesting one. Instagram took 17% more money, delivered 17% more impressions, and the price stayed put. When demand rises in an auction and the price does not move, there is only one explanation: new places to put ads appeared.

Those places are Reels. In the same Tinuiti benchmark they account for 35% of all Instagram ad impressions, up from 21% a year earlier. In twelve months Instagram conjured up a third of its inventory in a format that barely existed before. Supply grew as fast as demand, so the price never moved.

The practical consequence is sharp. The cheap impressions on Meta now sit in vertical video, and anyone who cannot produce it does not get the discount. Your static visual can be beautiful and it still will not fit a Reel.

A personal note. I spent a long time treating vertical video as a teenage fad that had no business selling professional services. The data talked me out of it before anyone on the team did.

The UK and the US are moving the same way

DataReportal, which pulls its figures straight from Meta's own planning tools, shows the same drift on both sides of the Atlantic. Facebook is still the bigger platform and it is still growing, just slowly. Instagram is adding people roughly four times faster.

In the UK, Facebook reaches 38.8 million people, which is 69.9% of adults, while Instagram reaches 35.5 million, or 62.6% of adults. In the US the split is 198 million against 182 million. The gap between the two platforms is closing at about seven percentage points of growth a year.

Czechia is in that table as the outlier worth watching. It is a market where Facebook is actually contracting, down 250 thousand people in a year. Smaller markets tend to show these shifts earlier, so if you operate across Europe, treat it as a preview rather than a curiosity.

Most of the numbers you will find were never measured

Search for Instagram ad costs and you will get dozens of articles with tidy ranges. They read well and they look specific. What they leave out is the only sentence that matters: where the figure came from.

I went through the first page of results and not one of them stated a methodology. No account count, no time period, no spend volume behind the average. Most are recycled benchmarks that somebody rounded and republished. That guesswork then gets cited onward until it hardens into an industry standard nobody ever measured.

Two honest sources can diverge more than two industries do. Gupta Media, working from tens of billions of impressions and fourteen-day trailing averages, put cost per link click at $0.37 in October 2025. LOCALiQ, working from a sample of campaigns, reports $0.70 for traffic campaigns. Both are correct. They are counting different clicks. So before you adopt a benchmark, find out what it measures.

What it actually costs

With all of the above in mind, here are figures from sources that published their methodology.

The spread matters more than the average. Traffic CPC runs from $0.34 in shopping and gifts to $1.22 in finance and insurance. Cost per lead is wilder still: $3.16 for restaurants, $76.71 for dentists. A dentist pays twenty-four times what a restaurant pays for one contact.

Cost per lead also climbed 20% year on year while traffic CPC edged down. Clicks got cheaper and the people behind them got dearer. That is a fairly precise picture of what the auction is doing.

You tune audiences. You pay for being dull

Most planning meetings go into settings: which audience, which optimisation, what budget per ad set. It is pleasant work because it happens at a desk and it looks rigorous. The decisive lever sits elsewhere.

Meta rewards ads that make people slow down. If your creative stops nobody, the system charges you more for the same impression. Weak creative is a fine for being dull, and you pay it without ever seeing it on an invoice.

The second lever is relevance, meaning how precisely the offer fits the person seeing it. The third is measurement, without which you optimise blind and money flows wherever the reporting looks flattering. And with Reels at 35% of Instagram impressions, the ability to produce vertical video at volume is now a lever in its own right.

“How much does it cost” is the wrong question

I made this mistake for years. In meetings I asked about cost per click first, because it is the number closest to hand and it compares neatly. It took me a long time to see that I was measuring diligence and calling it performance.

A click for a penny is expensive when it never buys. A click for a pound is cheap when one in five becomes a customer. Return is what settles it.

Companies that watch CPC alone almost always overpay for cheap clicks that do not convert, then wonder where the budget went. Cheap advertising that does not sell is the most expensive thing you can buy.

Frequently asked questions

Sources: Meta Platforms, Second Quarter 2026 Results (investor release) · Tinuiti, Digital Ads Benchmark Report Q2 2026 (over $4bn in annual ad spend under management) · DataReportal, Digital 2026 for the United Kingdom, United States and Czechia (Meta planning tool data, October 2025) · Gupta Media, Social Media CPM Tracker (tens of billions of impressions, fourteen-day trailing averages, October 2025) · LOCALiQ / WordStream, Facebook Advertising Benchmarks (2025) · and our own client campaign data.

Paying for clicks that never buy? Get in touch. We will look at whether your advertising earns or merely spends, and tell you straight.