If people already search for what you sell, try Google Search first: you pay more per click (a US$5.42 median in 2026, against US$0.60 to US$1.80 on Meta) but you are paying for someone who asked. If they don’t search for it yet, try Meta first, because Facebook and Instagram put you in front of people who were not looking and a lead costs less (US$27.39 against US$66.69 on Google Search). Either way, set up conversion tracking before the first dollar, because both platforms’ automatic bidding learns only from the results you report to it.
How we researched this. We read 40 sources on 12 October 2026. Two are the platforms’ own reports to investors: Alphabet’s quarterly filing with the SEC and Meta’s second-quarter results. Seven are benchmark studies built on thousands of accounts, from WordStream, LocaliQ, Triple Whale and Skai. Twenty are Google’s and Meta’s own help pages, which we quote rather than paraphrase, because they change. The rest are a survey of business owners, audience and market data for Australia and the United States, the ACCC, and two practitioners, each named with what they sell. Every figure is listed, with its link and date, at the end.
What is the real difference between them?
Where the person is when they see you. On Google Search your ad can show when someone types words related to your keywords. The person has a question and you are one of the answers. On Facebook and Instagram nobody typed anything. An auction decides which ad to show a person scrolling their feed, based on the audience you described and how likely Meta thinks they are to act.
That one difference drives everything else. Google sells you demand that already exists; Meta sells you attention you then have to turn into demand. So the search ad has to answer a question in a line, and the Meta ad has to stop a thumb. They are different skills. One is choosing the right words and the right page; the other is making images and video, often, because people tire of the same ad. Practitioners put it the same way from opposite sides. Ameet Khabra, whose agency runs both, says Google wins when people know what they want, and Meta wins when you want to influence what they want.
What does a click, a lead and a sale cost on each?
Here are the medians from the 2026 studies, paired so that each Google bar and its Meta bar come from the same study and the same method.
A click
A lead (service businesses)
A sale (online stores)
1,000 views (online stores)
Sources: LocaliQ and WordStream (US service businesses), Triple Whale (online stores worldwide). Full lines at the end of this guide.
Read it top to bottom and the verdict flips. For a service business collecting leads, Meta is far cheaper per click and per lead in the WordStream and LocaliQ data. For an online store selling products, Google wins on the number that pays the bills: Triple Whale’s 21,000 stores paid US$28.14 per sale on Google against US$38.99 on Meta, and got a return on ad spend of 3.27 against 1.88. Yet those same stores put 66.88% of their budget into Meta and 22.50% into Google. The reason is volume. There are only so many people searching for a product today, and Meta can find more buyers than search can.
Two warnings before you use any of these numbers. First, a Meta lead and a Google lead are not the same thing. A Meta lead is often a form filled in without leaving the app, by someone who was not looking for you a minute ago; a Google lead usually came to your website or called after searching. Neither study measures how many leads became customers. Second, these are US medians across very different industries. A Google lead for a lawyer cost US$131.63 in the LocaliQ data, and one for a dentist on Meta cost US$61.56 in WordStream’s. Your industry matters more than the platform average.
Why are some costs going up and others down?
Both companies told investors their prices rose in the second quarter of 2026: Google’s cost per click by 3% and Meta’s average price per ad by 12%. Yet Skai measured paid social clicks getting 18% cheaper in the same quarter, and WordStream’s Meta traffic campaigns fell from US$0.70 to US$0.60 a click. Both are true. On Meta, each thousand views got dearer, but people clicked more often (Skai saw click-through rates rise 33%), so each click cost less. On Google, more people clicked ads (paid clicks up 13%) and advertisers spent more, and Alphabet names device mix, country mix and advertiser spending as the drivers of its cost per click.
For a small advertiser the useful reading is simpler. Neither platform is getting cheap. Google’s cost per lead fell for the first time in five years in LocaliQ’s 2026 data, after Meta’s rose 20.94% the year before. Prices move every year, so test with your own numbers rather than last year’s article.
Which one suits which kind of business?
The practitioners agree on the pattern, even when they sell different things:
- Urgent services people search for (a lawyer, a clinic, a vet): Google first. The person needs you today and types it.
- Local services people plan (a dentist, a builder, a beauty salon): Google first for the searches that exist, then Meta to be remembered by the people who will need you later.
- Products people research (furniture, electronics, equipment): Google, including Shopping, because the buyer compares before paying.
- Cheap, visual or impulse products (clothes, gifts, jewellery): Meta first. Nobody searches for a gift they have not seen yet.
- Something new that nobody searches for: Meta first, because there is no search to buy.
- B2B with a long sales cycle: Google for the named problem, with Meta or LinkedIn to stay in mind while they decide.
PPC.io, which sells Google Ads management, gives the simplest test: if people already search for what you sell, run Google first; if they don’t, run Meta first. Its bias points to Google, and the test still holds. Before you decide, look your service up in Google’s free Keyword Planner. If almost nobody searches for it where you work, the question answers itself.
How each platform works, side by side
The table below is from the platforms’ own help pages. It matters because the rules decide how much money a test needs before it tells you anything.
| Google Search | Meta (Facebook, Instagram) | Source | |
|---|---|---|---|
| What shows your ad | Someone searching words you chose | An auction for a person in your audience, scrolling | Google Ads Help; Meta Help Center |
| What you usually pay for | A click | Views by default; clicks on some objectives | Google Ads Help; Meta Help Center |
| Spend on a busy day | Up to 2× the daily budget; no more than 30.4× a month | Up to 175% of the daily budget; no more than 7× a week | Google Ads Help; Meta Help Center |
| Minimum budget | None stated on the budget page | Varies by country and objective; 5× your cost-per-result goal | Google Ads Help; Meta Help Center |
| Learning before it settles | A few conversion cycles (1-2 typically) | About 50 results in the week after the last big edit | Google Ads Help; Meta Help Center |
| Tracking it learns from | Conversion actions, enhanced conversions, consent mode | Meta Pixel plus the Conversions API | Google Ads Help; Meta Help Center |
| Restricted categories | Healthcare, among others; advertiser verification | Housing, finance, employment, politics | Google policies; Meta Help Center |
The line that catches most people is the learning one. Meta says an ad set usually stops “learning” after about 50 results in a week, and that results are less stable and usually cost more until then. At WordStream’s median cost per lead, 50 leads a week is a four-figure weekly budget, and every significant edit starts the count again. With a small budget, a lead campaign may never leave learning, so the platform keeps guessing. Google’s automated bidding has a similar need: it calibrates over a few conversion cycles, and Google suggests judging a Target CPA campaign on 30 days with at least 30 conversions. A small budget is not hopeless on either platform, but it changes what you test: fewer campaigns, one goal, and no editing every day.
What has to be in place before you spend?
Tracking, and on both platforms it is the same idea: tell the platform when a visit became a call, a form or a sale, so its bidding can find more of the same. On Google that means conversion actions, ideally with enhanced conversions, and consent mode so that the cookie choice on your site is respected. On Meta it means the pixel plus the Conversions API, which sends results from your server and is less affected by ad blockers. Two rules are easy to miss. A health business must not send health information to Meta through the pixel, and Meta restricts the audience options for housing, finance and job ads in the United States, Canada and parts of Europe.
The cost of skipping this is measurable. In WordStream’s study of 15,666 Google Ads accounts, 29% recorded no conversions at all over 90 days, and the average account spent about US$3,127.38 a month. Some of those accounts converted nothing; others converted and never recorded it. From inside the account you cannot tell which.
Then the page they land on. A search ad that sends someone to a home page that doesn’t mention what they searched for wastes the click you paid the most for. And in Australia, everything the ad and the page say has to be true. The ACCC is clear that claims must be accurate and based on reasonable grounds whether or not you meant to mislead, and that fake or arranged reviews are against the law.
Is it different in Australia?
The audience is there for both. Google handled 87.9% of searches in Australia in September 2026 (Bing 9.17%), and Facebook’s ad tools reached 17.7 million people here in late 2025. The population is 27.0 million. Australian businesses spent A$8 billion on search advertising in 2025; all online advertising came to A$18.4 billion. What is missing is a public Australian benchmark for what a click or a lead costs on either platform. Statista’s data shows Australia had the highest Google search cost per click in Asia-Pacific in December 2024; the figure itself is paywalled. Treat the US medians above as a shape, not a quote, and expect search clicks here to cost more, not less.
One survey is worth knowing. LocaliQ asked 300-plus business owners, mostly in North America, and 56% were running social ads against 45% running search ads; 83% said referrals were still their best source of customers. Ads add to word of mouth; they rarely replace it.
What I’d do
I’d start with the search. Type what you sell into Google the way a customer would, and look at what comes up. If there are ads and a map, people search for it and Google is where your money should go first, to the one service that pays best, on a page written for that search. If nothing comes up, don’t pay Google to show you to nobody; take that budget to Meta, show people your work, and make it easy to message you. Either way, I wouldn’t spend a dollar before the tracking is in and tested with a real enquiry. And I’d give it a month of results before judging it, not a week of clicks.
Juan Pablo Tagliaferro, founder, HagoDigital
Questions people ask next
Is Facebook advertising cheaper than Google?
Per click and per lead, usually yes: US$1.80 a click and US$27.39 a lead on Meta against US$5.42 and US$66.69 on Google Search in the 2026 US medians. Per sale for online stores, no: US$28.14 on Google against US$38.99 on Meta. Cheaper is the wrong question; what matters is what a customer from each one is worth to you.
Can I run both at the same time?
Yes, and most businesses end up there. With a small budget, splitting it in two can leave neither side with enough results to learn from, so start with one, get it measured, then add the other.
What about Instagram?
Instagram ads are bought in the same Meta account and auction as Facebook ads, so “Facebook ads” here means both. In Australia, Instagram’s ad reach was 15.2 million people in late 2025; Facebook’s was 17.7 million.
How much should I spend to test?
Neither platform publishes one number. Meta’s rule is a daily budget of at least five times the cost of the result you want; Google tells you to start small and check daily. Set the test by results, not dollars: enough to buy the 50 results Meta wants in a week, or 30 conversions over 30 days on Google, at your industry’s cost.
Do I need a website first?
For Google Search, yes: the ad sends people to a page, and that page has to answer the search. Meta can collect leads in a form inside the app, but you still need somewhere to send people who want to check you out before they reply.
About this guide. HagoDigital runs Google and Meta ads for clients, so we have an interest in this question; that is why every figure here comes from someone else and is listed below, including what each source sells. Our own prices live on the pricing page and are not repeated here, so this guide stays true when they change. If you want us to look at your ads, this is what we do.