How to open an ad: what the first line is worth in cost per lead

We have run over 18,920 Meta ads, and of all the ways we labeled them, the opening moved cost per lead the most. By opening we mean the first thing the ad says or shows: the first line of copy, the first spoken words, the first frame. Across 2,126 labeled ads, the cheapest opening brought leads at $14.30 and the most expensive at $73.48. Nothing else we labeled had a spread that wide.

That makes the opening the cheapest thing in your ad to change. You do not need a new shoot to test it. You need a new first line.

The eight openings, side by side

Source: Creative Benchmarks, benchmark.sneeze.it (2,126 labeled ads; average $26.62)

Before you rewrite every ad you own, look at how many ads sit behind each bar. Offer only (611 ads) and talking to the viewer (626) are big groups, and their numbers ($26.30 and $27.01) sit right on the $26.62 average. Relatable moment has 178 ads. Bold claim has 99. Everything else is small: authority 63, curiosity 57, contrarian take 41, announcement 40.

So the two ends of the chart, the cheapest and the most expensive, are both built on about 40 ads each. That is enough to tell you where to test first. It is not enough to call either one a law.

The opening is the cheapest thing in your ad to change. You do not need a new shoot. You need a new first line.

What each opening is, in plain words

Contrarian take ($14.30, 41 ads). The ad opens by disagreeing with something the viewer has been told. "You do not need to do X to get Y." It works by making the scroller stop and check whether they have been wrong.

Bold claim ($20.16, 99 ads). A single, specific promise stated flatly. It only works if you can back it up at the front desk, so keep it to something you deliver for most customers.

Curiosity ($23.43, 57 ads). The opening raises a question the ad then answers. The answer has to arrive quickly and be worth the wait.

Offer only ($26.30, 611 ads). The price or deal is the first thing you see. This is the most common opening in our data and it lands right at average.

Talking to the viewer ($27.01, 626 ads). Someone speaks straight to camera, or the copy addresses "you" directly. Also very common, also average.

Authority ($33.63, 63 ads). The opening leans on credentials: years in business, awards, qualifications. Small group, above average cost.

Relatable moment ($37.93, 178 ads). A familiar everyday scene. It feels warm to make and it was one of the more expensive openings in our data.

Announcement ($73.48, 40 ads). "We are open." "Now booking." "Introducing." It costs almost three times the average per lead. The viewer learns that something exists, but not why it matters to them.

Cost per lead and winner rate do not agree here

This is the part that trips people up. Contrarian take was the cheapest opening, but only 5.9% of those ads became winners, meaning Meta chose to spend 10 times the account's median ad on them, and at least $500. Curiosity cost more per lead ($23.43), yet 16.0% of curiosity ads became winners. Bold claim sat in between on both: $20.16 per lead and 14.9% winners.

Here is how to use the two numbers:

• Cost per lead tells you how efficient an ad was with the money it got. Use it to decide which ads deserve more of your budget.
• Winner rate tells you how often Meta's system picked that kind of ad to spend on. Use it to decide which kinds of ad are worth launching more of, because an ad Meta will not spend on cannot bring you leads, however efficient it is on paper.

On that reading, bold claim is the opening that did well on both measures, with a group size (99) that is still small but less fragile than the others. That is where we would start. We go deeper on the disagreement in winner rate vs cost per lead.

The playbook: test openings, keep everything else still

1. Pick one ad you already run

   Choose an ad with an average result in your account. Do not start with your best ad or your worst. You want a clean baseline.

2. Write three new openings for it

   One bold claim, one curiosity, one contrarian take. Keep the rest of the ad identical: same image or video, same offer, same button.

3. Retire any announcement opening

   If an ad opens with "now open", "introducing" or "we are excited to", rewrite the first line to say what the customer gets. Keep the news in the second line if it matters.

4. Launch the versions in the same week

   Same audience, same budget setup, same week, so the only difference is the opening. Launching on different weeks mixes in season and competition.

5. Give each version a fair amount of spend

   Do not call a test after a day. Two in three ads in our data spent under $100 in their whole life, so many ads never get enough money to judge. Decide your minimum spend before you launch.

6. Read both numbers before you decide

   Look at which version Meta spent on and which version brought cheaper leads. If one wins both, scale it. If they split, keep the cheaper one live and write another version of the one Meta preferred.

7. Repeat with the next ad

   One test teaches you about one ad. Three or four tests teach you which openings work for your account.

A worked example

Say you spend $1,500 a month on lead ads and your account's average cost per lead matches our labeled average of $26.62. That buys roughly 56 leads. If an announcement opening performed like our announcement group, at $73.48, the same $1,500 would buy roughly 20. That is arithmetic, not a forecast, and the announcement group is only 40 ads. It shows why the first line is worth checking on every ad you run.

For the full set of labels and how we built them, read what 18,920 Meta ads say about cost per lead. If your ads open with a headline and nothing else, fixing headline-only ads is the next playbook to read.