Meta Ads spent the most on 35-44 men, the highest CPC in the set. Cost per lead was still $33
We pulled the lifetime age and gender breakdown for every AUD account in the set and pooled it. The segment that received the most money was men aged 35 to 44: $107,366, 13.6% of the pooled spend, more than any of the other eleven age-gender cells. It was also the segment with the highest cost per click in the table, $1.21, and a click-through rate of 1.0%, the joint-lowest among the four segments that took the most budget. On every column an Ads Manager breakdown colours red, it was the worst buy in the set.
Its cost per lead was $33. That is joint third-cheapest of the twelve segments, and within three dollars of the cheapest.
This post is about that gap: why the columns most people read said one thing, why the column that matters said another, and why Meta's delivery system was right to put the money where it did. Part 15 takes the same table and asks a different question, about what cheap clicks are actually worth.
The table
Twelve segments, pooled across the AUD accounts, unknown-gender rows removed. Leads are each account's own lead event, so the definition is mixed across accounts; the ratios still hold within a segment because the same accounts contribute to every row.
Segment | Spend | Share | CPM | CTR | CPC | Leads | Lead rate per click | CPL |
|---|---|---|---|---|---|---|---|---|
35-44 male | $107,366 | 13.6% | 11.75 | 1.0% | 1.21 | 3,214 | 3.6% | $33 |
45-54 female | $96,908 | 12.3% | 15.13 | 2.0% | 0.74 | 2,986 | 2.3% | $32 |
55-64 female | $86,779 | 11.0% | 13.20 | 1.9% | 0.68 | 2,597 | 2.0% | $33 |
45-54 male | $80,872 | 10.3% | 10.98 | 1.0% | 1.08 | 2,420 | 3.2% | $33 |
65+ female | $69,591 | 8.8% | 7.26 | 1.2% | 0.59 | 2,338 | 2.0% | $30 |
35-44 female | $69,184 | 8.8% | 11.61 | 1.5% | 0.76 | 1,849 | 2.0% | $37 |
65+ male | $68,141 | 8.7% | 10.20 | 1.6% | 0.64 | 1,617 | 1.5% | $42 |
25-34 male | $65,661 | 8.3% | 7.70 | 0.9% | 0.89 | 1,743 | 2.4% | $38 |
55-64 male | $65,353 | 8.3% | 9.72 | 1.1% | 0.87 | 1,769 | 2.4% | $37 |
25-34 female | $41,329 | 5.3% | 8.31 | 1.2% | 0.67 | 931 | 1.5% | $44 |
18-24 male | $13,549 | 1.7% | 4.64 | 0.7% | 0.64 | 306 | 1.5% | $44 |
18-24 female | $10,753 | 1.4% | 5.72 | 0.9% | 0.61 | 315 | 1.8% | $34 |
Read the top row across. The impressions were expensive (CPM 11.75, fourth-highest), few of them were clicked (1.0%), and each click cost more than in any other segment ($1.21, against $0.59 at the cheap end). Then look at the last three columns. The segment produced 3,214 leads, the most of any cell, because 3.6% of its clicks became a lead, the highest lead rate in the table by a clear margin. The next best is 45-54 men at 3.2%; no female segment is above 2.3%.
So the segment that cost the most per click needed the fewest clicks per lead, and the two effects nearly cancelled. $33 per lead is the same figure as 55-64 women and 45-54 men, and three dollars off the 65+ women who lead the table at $30.
Why this happens
Meta's auction prices a click by what other advertisers will pay for that person, and advertisers pay more for people who convert. A 35-44 man in Australia is competed for by every trade, finance, insurance and B2B advertiser in the country. That competition is what a $1.21 CPC measures. It is not a verdict on whether the segment is good for you; it is a reading of how good other people think it is.
The click-through rate reads the same way. A person who sees a lot of ads clicks a smaller share of them. The 35-44 male segment has the fourth-highest CPM and one of the lowest CTRs, which is consistent with a heavily advertised audience that has learned to scroll past. Their 3.6% lead rate says that when they do click, they meant it.
The lower rows show the mirror image. 18-24 men had the cheapest impressions in the set (CPM 4.64), a CPC of $0.64, and a 1.5% lead rate, which lands them at $44 per lead, the joint-worst. Nobody was bidding against the accounts for those impressions, and the reason nobody was bidding is in the last column.
There is a Meta mechanism underneath. Under lowest-cost bidding with a conversion or lead optimisation goal, the delivery system estimates each person's probability of firing the event and bids in proportion. It does not see CTR as a target. It will happily pay a high CPC for a low-CTR audience if the estimated lead rate justifies it, and the table says that is what it did: the three segments with the highest CPCs (35-44 men, 45-54 men, 25-34 men) have the three highest male lead rates, with 55-64 men tied on 2.4%.
What it does to decisions
The breakdown view in Ads Manager shows spend, impressions, CTR, CPC and cost per result side by side, with conditional colouring available on any column. The columns most advertisers sort by are the ones that update fastest and have the most data behind them, which means CTR and CPC. On those columns, 35-44 men look like a segment to exclude. Here is what the same twelve segments look like ranked by each metric.
Segment | Rank by CTR (best first) | Rank by CPC (cheapest first) | Rank by lead rate (best first) | Rank by CPL (cheapest first) |
|---|---|---|---|---|
35-44 male | 8= | 12 | 1 | 3= |
45-54 female | 1 | 7 | 5 | 2 |
55-64 female | 2 | 6 | 6= | 3= |
45-54 male | 8= | 11 | 2 | 3= |
65+ female | 5= | 1 | 6= | 1 |
35-44 female | 4 | 8 | 6= | 7= |
65+ male | 3 | 3= | 10= | 10 |
25-34 male | 10= | 10 | 3= | 9 |
55-64 male | 7 | 9 | 3= | 7= |
25-34 female | 5= | 5 | 10= | 11= |
18-24 male | 12 | 3= | 10= | 11= |
18-24 female | 10= | 2 | 9 | 6 |
The first two columns are what the dashboard rewards. The last column is what the business pays. They disagree in both directions: 65+ men rank third on CTR and third on CPC and tenth on cost per lead; 35-44 men rank twelfth on CPC and third on cost per lead. Only 65+ women look good on every column, and they are the exception that makes the rule seem reasonable.
An advertiser who acted on the click columns would have done one of three things, and each of them costs money.
Action taken on click metrics | What happens to cost per lead |
|---|---|
Exclude 35-44 and 45-54 men (worst CPC, low CTR) | Removes the two highest-lead-rate segments, 5,634 leads at $33 each, and pushes spend elsewhere, where the other ten segments range from $30 to $44 |
Shift budget toward 65+ men and 18-24 (high CTR, cheap CPC) | Buys leads at $42 and $44 instead of $33 |
Build a female-only ad set because women "click more" | Sends more spend to the 25-34 female segment at $44, alongside the strong 45-54 and 55-64 rows |
The 80 female-only ad sets we counted in part 13 against 7 male-only ones are consistent with the third of those. Women in this set do click more, by roughly double in the 45-64 age bands, and the 25-34 female segment is the second-worst cost per lead in the table.
The allocation Meta actually made put its five biggest allocations (35-44 men, 45-54 women, 55-64 women, 45-54 men, 65+ women) into exactly the five cheapest-CPL segments, and put the least money into 18-24, where men cost $44 per lead and women $34. On the evidence here, the system was reading the right column.
Where it varies
The pull for this post is pooled, so we cannot split it by band or account without a second pull, and we have not done one. Two things about the pooling matter for how far to trust it.
First, the lead definitions differ by account: some count an instant-form submission, some a pixel event, some a phone call. Because every account contributes to every age-gender row, that mixing does not favour one segment over another, but it does mean $33 is a pooled figure and not something any one account should expect to see.
Second, the volume is high enough that no segment is below the floor. The smallest cells are 18-24 men (306 leads on $13,549) and 18-24 women (315 leads on $10,753), both well clear of the 30-lead and $500 thresholds we use in this series. The 65+ rows are the ones to be a little careful with for a different reason: Meta stores 65+ as a single bucket, so it pools people from 65 to whatever age they stop using Facebook, and the 65+ female segment's $30 cost per lead is a good number rather than a well-understood one.
Within the table, the spread by gender is worth a look on its own.
Male segments | Female segments | |
|---|---|---|
Highest CTR | 1.6% (65+) | 2.0% (45-54) |
Lowest CTR | 0.7% (18-24) | 0.9% (18-24) |
Highest CPC | $1.21 (35-44) | $0.76 (35-44) |
Lowest CPC | $0.64 (65+, 18-24) | $0.59 (65+) |
Highest lead rate | 3.6% (35-44) | 2.3% (45-54) |
Lowest lead rate | 1.5% (65+, 18-24) | 1.5% (25-34) |
Cheapest CPL | $33 (35-44, 45-54) | $30 (65+) |
Dearest CPL | $44 (18-24) | $44 (25-34) |
Men in this set cost more to click and convert more often per click; women cost less to click and convert less often. The two land in the same place on cost per lead, $33 to $44 either way. A dashboard sorted by CPC says the accounts should be buying women. A dashboard sorted by cost per lead says it does not much matter, and 35-44 is the strongest age band for men and 45-64 for women.
What to check in your account
Ads Manager, Campaigns or Ad sets view, set the date range to at least 90 days, then Breakdown, By delivery, Age and gender. This gives you the twelve-row table above for your own account.
Add the columns CTR (link click-through rate), CPC (cost per link click), Results, and Cost per result. Make sure the result is your lead event and not a proxy; part 1 covers how far the default Leads column can drift from the real count.
Add a custom metric for lead rate per click: Results divided by Link clicks. This is the column Meta does not show by default and it is the one that explains the others.
Sort by CPC descending. For the top three segments, read the lead rate and cost per result before deciding anything. If the expensive click is buying a cheap lead, leave it alone.
Sort by CTR descending. For the top three, do the same check. A high CTR at a low lead rate is a segment that clicks and does not convert, and it costs you in the column that arrives as an invoice.
If you have any age or gender exclusions live, find out when they were set and what column justified them. If the answer is CTR or CPC, re-run the check with cost per result.
Check the volume behind each cell before acting. Below 30 leads in a segment the cost-per-lead figure is noise, and at that point the click metrics are all you have, which is a reason to wait rather than a reason to trust them.
Our Facebook Ads management work starts most audits with this breakdown, because it is the quickest way to find out whether an account is optimising to the metric it reports or the one it needs.
The principle
Every column in a breakdown is an answer to a different question. CTR says whether people notice. CPC says what other advertisers think the person is worth. Lead rate says whether the click meant anything. Cost per lead is the only one that includes all three, and it is the only one the business is charged in. Meta's delivery system optimises to the last of those and mostly gets it right. The mistake in most accounts is not the algorithm's allocation; it is a human reading the first two columns and correcting it.
