Why Are Your Meta Ads Costing More and Converting Less?

If your Meta ads are costing more and converting less, you're not imagining it. Almost every client review we ran this year started with the same question. The usual answer is that Meta has become more expensive and everyone is bidding. That's partly true, but it's the least useful explanation, because it leaves you nothing to fix. So we went back through three years of our own Meta data to find out what actually changed.

How much do Meta ads cost in Australia?

Meta ads cost depends on three things: what you pay for impressions, how many people click, and how many of them buy. Across the accounts in our audit, the median CPM (cost per thousand impressions) was $11.65 in 2024, $13.73 in 2025 and $15.44 in 2026 to date. Blended e-commerce cost per purchase went from $31.78 in 2024 to $37.91 in 2025 and $38.44 in 2026 to date.

For outside reference, Meta's own results showed the average price per ad up 12% year on year in the second quarter of 2026, across all its apps worldwide. In the US, WordStream's 2025 benchmarks put the average Facebook cost per lead at $27.66, against $70.11 on Google. Australian CPM trackers are thin, so treat any single number with caution.

Those blended numbers hide a lot, because one large account can move the average. The more useful view is what happened inside the same accounts over time.

The three numbers behind your Meta ads cost

Your cost per purchase is your CPM divided by your click-through rate and your conversion rate. In plain terms: what you pay to be seen, how many people click, and how many of those people buy. If any one of them moves the wrong way, your cost per purchase rises, even if the other two hold.

That's why "Meta is expensive" is the wrong starting point. The first job is to find which of the three moved. A rising CPM points to the auction, your audience or your placements. A falling click-through rate points to your creative. A falling conversion rate points to what happens after the click: the landing page, the offer, the checkout, or your tracking.

What our audit found: it's mostly not the price

We looked at 38 Meta ad accounts with at least $5,000 of spend, from September 2023 to 28 September 2026. That's $7.4M of spend across e-commerce, lead generation and not-for-profit accounts. Every account's change in cost was split into its three parts.

• Costs rose, but not everywhere. 19 of 29 account-year comparisons from 2024 onwards got more expensive per result, covering 68% of spend.
• On a constant panel of seven e-commerce brands, cost per purchase went from $26.00 (2024) to $35.44 (2025) to $41.69 (2026 to date), and ROAS fell from 4.89 to 3.37.
• Price barely moved. Across the same accounts, the median CPM rose 3.6% in 2025 and fell 3.6% in 2026. CPM fell for 15 of 25 accounts this year. On the seven-brand panel, CPM explained only 7% of the 2025 cost rise.
• Click-through drove 2025. Panel click-through fell 17%, which accounted for about 60% of that year's cost rise.
• Conversion drove 2026. Conversion rate was the main driver in 9 of the 19 accounts where cost rose.

Like-for-like, June to August 2026 against June to August 2025, e-commerce cost per purchase rose from $34.85 to $42.43 and ROAS fell from 2.99 to 2.58. Not every brand moved the same way, and the exceptions are where the lessons are.

Four reasons Meta ads convert less

1. Fewer of the people who click are buying

This was the biggest single driver. For one gifting brand, CPM in Australia was flat ($13.31, then $13.29) and order value held at around $200. But conversion rate fell from 8.32% to 5.04% to 3.48%, and cost per purchase went from $22 to $53. Nothing changed in the auction. Something changed after the click. Our data can't separate the cause, so check the landing page, the offer, the checkout experience and your tracking, in that order.

2. The creative wore out

For a women's fashion brand, CPM fell 19% between 2024 and 2025, but click-through dropped from 0.74% to 0.49%, so cost per purchase still rose 46%. Meta's own analysis found that by the fourth exposure to the same ad, the chance of converting drops by about 45%. And since Meta's Andromeda update, its delivery system reads the creative itself to decide who sees an ad. Your creative now does much of the job your targeting used to do, so tired creative costs more than it used to.

3. Spending more on the same people

In our June to August comparison, spend rose 84% year on year and median frequency went from 3.90 to 4.21. For one pet food brand, frequency reached about 10 within a single quarter. More budget into the same audience mostly means reminding the same people, at a higher price.

4. Some of the numbers are misleading

The same pet food brand's click-through rate looked up 51% in 2026. But Audience Network, which took 4.6% of spend, delivered 53.9% of link clicks and converted at 0.36%, against 8.07% for the rest of the account. Excluding it, click-through actually fell 12.6%. Its largest campaign, a trial offer, spent $328,465 at a cost per purchase of $32.23, which looks efficient, but returned $0.82 per dollar. Measurement also shifted this year: Meta changed what its reporting API returns for view-through attribution from January 2026, so some conversions didn't disappear, the report changed.

What didn't fix it, and what did

Several common fixes didn't work on their own. One gifting brand ran around 1,300 ads with spend and moved 94% of its 2026 spend into Advantage+ shopping campaigns. It was its most expensive year: cost per purchase up 158% from 2024. Reels share of spend rose in every account we measured (for example 7% to 25%, and 3% to 30%), with no consistent drop in cost.

The clearest success went the other way. A gifting and lifestyle brand saw CPM rise 67%, from $10.83 to $18.04, the biggest price increase in the panel. Its cost per purchase still fell 16%, from $17.87 to $15.08, and ROAS rose from 4.10 to 4.43. It cut its ads from 306 to 156 and put 56% of spend behind its top five. Click-through went from 1.93% to 3.43%, which more than covered the price rise. To be fair, its spend also fell to $48k, and a smaller budget can flatter efficiency. But the lever is clear: better, more distinct creative beat a higher price.

A novelty card brand is the other bright spot. In 2026, CPM fell 28%, click-through rose 25%, and cost per purchase fell 32%, from $10.65 to $7.25.

How to diagnose your own Meta account

1. 1Split your cost per purchase into CPM, click-through rate and conversion rate for each of the last three years. Find the one that moved most.
2. 2Break results down by placement. Make sure cheap clicks from one placement aren't flattering your click-through rate.
3. 3Check frequency by quarter, not by year. If it's climbing, you need new creative or new audiences before you need more budget. Our guide to what UGC does for Meta ads covers how to keep creative fresh.
4. 4Judge offers and campaigns on return, not just cost per purchase. A cheap trial can lose money.
5. 5Check tracking and attribution settings before changing anything else. If the numbers are wrong, every decision built on them is too.

So, why are your Meta ads costing more?

Mostly, it's not the price. In our $7.4M audit, the typical CPM barely moved while cost per purchase rose 60% on the same brands. The causes were fewer people clicking and fewer of them buying. Those are fixable with better creative, a better experience after the click, and numbers you can trust.

That's how we run paid social for clients at Vonnimedia. Every account review starts by splitting the cost into its three parts, so the fix goes where the problem is. See how that plays out in our case studies.