---
title: "Direct Response Copywriting: What 156 Meta Ads Show in 2026"
description: "Direct response copywriting explained with 2026 data: 156 Meta ads by run length, 700 ad lines checked against customer comments, and dated pay figures."
canonical: https://tryadlicio.com/blog/direct-response-copywriting
---

# Direct Response Copywriting: What 156 Meta Ads Show in 2026

Published 2026-10-30. https://tryadlicio.com/blog/direct-response-copywriting

Direct response copywriting is copy that asks the reader to take one action now and measures whether they took it. For this guide we sorted 156 health, beauty and pet ads from Meta's Ad Library by how long they ran, and checked 700 lines of ad copy against 10,087 public customer sentences about the same products.

## What we found

- Long-running ads rarely repeat public customer phrases: 0 of 284 lines, or 1 with a looser check.
- Against bigger Amazon review sets, 1 of 416 ad lines matched, and it was common wording.
- Lasting brand and retailer ads ran a median of 50 words; the very long ones were mostly article-style ads.
- Most lasting ads skip urgency: 53 of 62 had none, though newer ads used it even less.
- Of 101 Reddit comments on AI and copywriting, 2 describe losing copy work to AI.
- Pay estimates dated 2024 to 2026 put direct response copywriters at $63,042 to $79,286 a year.

## Do lasting ads reuse the customer's exact words? We checked 700 lines

An [agency post on review mining](https://www.pigeondigital.com/insight/review-mining-ad-copy-voice-of-customer-workflow-shopify) says "Roughly nine out of ten ad concepts we write start with a sentence we didn't write" and tells readers to "copy and paste verbatim." [Liz Slyman's guide](https://lizslyman.com/direct-response-copywriting-the-ultimate-guide/) quotes Schwartz: "Copy is not written, it is assembled." That may well work. Our narrower question: do the customer's words survive, word for word, into ads that keep running?

### The result: 0 of 284 lines

We split the oldest ads that 10 health, beauty and pet brands still run from their own pages (median age 145 days) into 284 lines and checked each against 3,431 public customer sentences from Reddit, YouTube and Amazon. With a match rule fixed in advance and every hit read by hand, 0 of 284 ad lines repeated a customer phrase from the comments we collected. A looser check, added later, found 1.

| Which ad lines | Lines | Rule fixed in advance | Either check |
|---|---|---|---|
| All eligible lines | 284 | 0 | 1 (0.4%) |
| Lines about a problem, feeling, doubt or result | 140 | 0 | 1 (0.7%) |
| Without the brand that supplied 93 of the lines | 191 | 0 | 0 |
| Without the two comment sets under 50 sentences | 220 | 0 | 1 (0.5%) |
| Against another brand's comments in the same category | 284 | 0 | 0 |

The one match is a common idiom: a probiotic ad's "Die-off can make you feel worse before you feel better" ([ad](https://www.facebook.com/ads/library?id=1474019410446456)) against an [Amazon review](https://www.amazon.com/product-reviews/B01GQSXBMI/) that says "feeling worse before you feel better."

Read it as "no public phrase we found survives verbatim". It says nothing about whether the brands did customer research, because most of what customers write sits where we could not look.

### Even quoted testimonials were not in public comments

11 of the 284 lines are quoted customer testimonials. None matched, so they likely come from sources we could not reach: site reviews, surveys or emails. We tried one: a hair-growth serum's [product page](https://us.champohaircare.com/products/growth-serum) shows a 4.77 rating from 1,334 site reviews, but they load by script and our scraper got 3.

### Bigger review sets, same answer

Two comment sets above were small (29 and 45 sentences), so we reran the check on five brands with 486 to 1,599 Amazon review sentences each. Their ads were at most 68 days old, so this tests review-set size only. 1 of 416 ad lines matched its own product's reviews, and it is category wording ("help keep skin feeling hydrated"). Two matched another brand's reviews. A dog-supplement brand from the first test, rechecked against 12 times more text, matched 0 of 13.

Of the 148 near-matches we rejected, 76 were product names, 41 category phrases ("in their water bowl"), 11 ingredient lists, 7 coincidences or artifacts and 13 echoes of the brand. A creator ad calls a serum "a gentler form of vitamin c suited for sensitive skin" ([ad](https://www.facebook.com/ads/library?id=1117391167612366)); an [Amazon review](https://www.amazon.com/product-reviews/B01M4MCUAF/) says it uses "a gentler form of Vitamin C that doesn't irritate my skin." Where an ad and a review share words, they most likely travelled from the brand to the reviewer.

### What copywriters say they do with the words

We sorted 40 posts and comments in 4 recent Reddit threads on customer research ([voice-of-customer process](https://www.reddit.com/r/copywriting/comments/1sqrqfs/the_voc_research_process_i_run_before_writing_a/), [free sources](https://www.reddit.com/r/copywriting/comments/1t8pk11/a_quick_easy_zerocost_way_to_gather_sources_for/), [research with AI](https://www.reddit.com/r/copywriting/comments/1o5d39o/how_to_do_market_research_with_ai/), [consumer research](https://www.reddit.com/r/AskMarketing/comments/1r4kkfq/are_marketing_teams_actually_doing_real_consumer/)). 16 of 40 name a method: AI tools 11, Reddit 5, forums 4, Amazon reviews 3. On using the words, 2 say keep the exact words (one promotional), 2 say reframe them and 2 warn that some themes test badly. The author of the voice-of-customer (VoC) process post says both:

> "Don't summarize, keep the exact words. The exact words are the whole point."

> "VoC gives you the emotional territory. the creative work is figuring out a fresh way to frame that territory so it doesn't feel like a copy-paste from the reviews page."

Four threads are not a survey, but they fit the ad data: research supplies the territory, the ad supplies the sentence.

### Same pain, different words

Hand-picked illustrations, not matches and not counted:

| Category | Ad line | What customers wrote | Gap |
|---|---|---|---|
| Vaginal-health probiotic | "All of them just kept prescribing antibiotics, not helping me get to the root cause." ([ad](https://www.facebook.com/ads/library?id=1009399674975837)) | "Nothing had ever worked before, antibiotics were ruining me" ([Reddit](https://www.reddit.com/r/Healthyhooha/comments/10var16/anyone_use_happy_v_prebiotic_probiotics_for/)) | Same pain, new words |
| Dog dental powder | "Say goodbye to bad dog breath & yellow teeth with this proven green powder!" ([ad](https://www.facebook.com/ads/library?id=1003576625514683)) | "thinking the powder is going to replace brushing, which it isn't" ([YouTube](https://www.youtube.com/watch?v=xr2wl6t5gYs)) | Objection skipped |
| Mushroom coffee | "game-changing daily immune support and long-lasting energy in every cup" ([ad](https://www.facebook.com/ads/library?id=1432300561861014)) | "I've definitely been less bloated, better sleeps, and more energetic" ([YouTube](https://www.youtube.com/watch?v=ey0WLTj9hrw)) | Benefit left out |

Each ad covers ground customers care about in the brand's own words, and two skip what customers raise. Two of our [27 principles of a winning ad](https://tryadlicio.com/blog/27-principles-of-a-winning-ad) apply: sound like the customer, and answer the objection before it forms.

## What Meta ads that run 90+ days look like in 2026

We sorted 156 US ads, pulled on 2026-09-23, into three groups: long-runners, still active 90+ days after they started (62, median age 145 days); active new entrants with new copy, started 14 days or less earlier (46); and ads that stopped within 14 days, for reasons the data does not show (48). Run length shows that an ad survived, not that it made money.

| Measure | Long-runners (62) | New entrants (46) | Stopped within 14 days (48) |
|---|---|---|---|
| Under 50 words | 22 (35.5%) | 14 (30.4%) | 8 (16.7%) |
| 1,000+ words | 17 (27.4%) | 8 (17.4%) | 8 (16.7%) |
| Median words | 75 | 92.5 | 117 |
| Page name is not the brand | 21 (33.9%) | 20 (43.5%) | 22 (45.8%) |
| Urgency line (hand-checked) | 9 (14.5%) | 1 (2.2%) | 5 (10.4%) |
| Price or discount | 22 (35.5%) | 12 (26.1%) | 11 (22.9%) |
| Guarantee or free trial | 23 (37.1%) | 14 (30.4%) | 13 (27.1%) |

Lasting ads hold more very short and very long copy than the other groups, and name a price, discount or guarantee more often.

### Short brand ads and long advertorials

[Linear's guide](https://lineardesign.com/blog/direct-response-copywriting/) on page 1 says "Copy got shorter, because a feed is not a mailbox," while [Stefan Georgi's video](https://www.youtube.com/watch?v=pZ82pzVoAZo) says a long sales letter "does tens of millions of dollars a year in revenue" (his figure). Our long-runners hold both: brand and retailer pages sit at a median of 50 words (41 ads), like a 15-word Chewy ad for Wuffes joint chews that has run 153 days ([ad](https://www.facebook.com/ads/library?id=3385183138349894)).

The 17 long-runners over 1,000 words are almost all advertorials, long ads written like articles, on pages whose name is not the brand (16 of 17), and 2 advertisers account for 11 of them. The one brand-page exception, from Based Supplies, runs 4,594 words and opens "Please STOP buying hair growth supplements until you check your shampoo." ([ad](https://www.facebook.com/ads/library?id=1229755845899735)). Long copy lasts, but here it is a format a few advertisers run.

### Most lasting ads carry no urgency line

53 of 62 long-runners have none, so urgency is optional; use it only when the deadline is real. Yet long-runners have the highest rate of the three groups (9 of 62, against 1 of 46 and 5 of 48). Four hits are on brand pages, such as Vortexhaircare's "Our biggest special of the year is ending" ([ad](https://www.facebook.com/ads/library?id=1109239824531891)) and three "limited time" offers. Four of the other five are a "sell out" line in a postscript of a long advertorial, such as "[brand] is a small American company and they sell out." ([ad](https://www.facebook.com/ads/library?id=4594018594163268)). For a deadline that is real, see how brands word it in [Black Friday ad examples from the Meta Ad Library](https://tryadlicio.com/ads/black-friday).

Our free [competitor ad finder](https://tryadlicio.com/tools/competitor-ad-finder) groups a brand's repeated copy and sorts its active Meta ads by how long each has run.

## How to use customer research when the words don't transfer

The research still decides what the ad is about:

1. Collect where customers discuss the problem: three- and four-star reviews, where buyers say what almost stopped them (see [turning Amazon reviews into ad angles](https://tryadlicio.com/blog/turning-amazon-reviews-into-ad-angles-for-ecom-brands)), then Reddit threads and independent review videos.
2. Filter out the brand's echo: sponsored, gifted and affiliate videos, and comments that mention an ad. We dropped 43 of 60 videos.
3. Sort what is left into pains, objections, desired outcomes and usage moments, and count each.
4. Find the objection your ads skip, like the brushing question above. Our guide to [extracting customer objections for ad messaging](https://tryadlicio.com/blog/extracting-customer-objections-for-ecom-ad-messaging) shows how to rank them.
5. Write the ad in your own words around that pain and outcome.
6. Test themes, not only lines: "I've seen VoC themes that read great but converted terribly when tested," one commenter warned ([thread](https://www.reddit.com/r/copywriting/comments/1sqrqfs/the_voc_research_process_i_run_before_writing_a/)).

[Adlicio](https://tryadlicio.com/voice-of-customer) is one way to do steps 1 to 3. It pulls verbatim customer comments and reviews from public sources (Reddit, Amazon, YouTube, TikTok, Instagram, Trustpilot, G2, App Store and more) and groups them into pains, objections, desired outcomes and phrases for ad angles.

## Is AI replacing copywriters? What 101 Reddit comments say

Agency job postings give the one hard number: [The Drum reported on 2026-09-09](https://www.thedrum.com/news/as-writing-jobs-go-from-1-6-to-1-30-of-agency-postings-are-copywriters-in-trouble) that writing roles went from "one in six agency vacancies in late 2025" to "one in 30". In the [ProCopywriters Survey 2026](https://www.procopywriters.co.uk/survey-2026/) (573 UK writers), "Almost three-quarters of copywriters now use generative AI tools," yet "64% of copywriters said they'd press it" to erase generative AI. [Forbes](https://www.forbes.com/sites/chelseatobin/2026/05/25/ai-was-going-to-replace-copywriters-instead-anthropic-is-hiring-them/) ("Today, talented human writers and content marketers are in demand") reports hiring, and layoff accounts appear in [Blood in the Machine](https://www.bloodinthemachine.com/p/i-was-forced-to-use-ai-until-the).

We read all top-level comments in the 6 busiest copywriting-and-AI threads ([worth it in 2026](https://www.reddit.com/r/copywriting/comments/1r1i2zg/is_copywriting_worth_it_in_2026/), [replaced with AI](https://www.reddit.com/r/DigitalMarketing/comments/1u4u6a5/my_company_replaced_our_copywriter_with_ai_and_i/), [EntrepreneurRideAlong](https://www.reddit.com/r/EntrepreneurRideAlong/comments/1u27fl4/im_a_copywriter_and_i_started_to_believe_that_one/), [b2bmarketing](https://www.reddit.com/r/b2bmarketing/comments/1vsfobn/why_are_ais_so_bad_at_copywriting_despite_being/), [freelance](https://www.reddit.com/r/copywriting/comments/1u3damx/is_it_still_worth_it_to_pursue_freelance/), [future](https://www.reddit.com/r/copywriting/comments/1pgwa1b/what_is_the_future_of_copywriting_in_2026/)). Of 101, 80 were opinion and 21 first-hand (one has two labels): 13 use AI in their own writing, 2 describe losing copy work to AI, 1 saw copy roles cut, 1 says work held steady (ending in a course pitch), 5 other. Two first-hand accounts:

> "I'm the copywriter who got fired in this scenario lol only took my old company a month and a half after letting me go to put out a call for a freelance writer" ([thread](https://www.reddit.com/r/DigitalMarketing/comments/1u4u6a5/my_company_replaced_our_copywriter_with_ai_and_i/))

> "They had a handful of copy people. They kept one. Now his job is to finesse the Ai copy that the marketing people input into co-pilot." ([thread](https://www.reddit.com/r/DigitalMarketing/comments/1u4u6a5/my_company_replaced_our_copywriter_with_ai_and_i/))

Only 6 comments mention direct response (DR) work (sales letters, video sales letters, ads, funnels), 3 of them first-person and 1 promotional. One writer, 6 to 7 weeks into a sales letter "with Claude, ChatGPT and Perplexity," still did not "feel comfortable at releasing the final version" ([thread](https://www.reddit.com/r/copywriting/comments/1r1i2zg/is_copywriting_worth_it_in_2026/)). Our read: the squeeze shows in agency writing roles, and first-hand evidence about DR work is close to nil.

## What direct response copywriters earn in 2026

The only salary average on Google's first page is $65,310 for 2022, which [Liz Slyman's guide](https://lizslyman.com/direct-response-copywriting-the-ultimate-guide/) credits to ZipRecruiter without a link; the rest are self-reported incomes from 2017 and 2021. Dated sources:

| Source | Figure | Data date | Method and scope |
|---|---|---|---|
| [BLS, writers and authors](https://www.bls.gov/ooh/media-and-communication/writers-and-authors.htm) | Median $76,910; bottom 10% under $44,310; top 10% over $139,870 | May 2025 | Employer survey, all writers; excludes the 65% who are self-employed |
| [PayScale, DR copywriter](https://www.payscale.com/research/US/Job=Direct_Response_Copywriter/Salary) | $79,286 average; 10th to 90th percentile $52k to $107k | 1,268 profiles, updated April 2024 | Self-reported |
| [Salary.com, DR copywriter](https://www.salary.com/research/salary/posting/direct-response-copywriter-salary) | $63,042 average; range $48,402 to $70,715 | September 1, 2026 | Job-posting model |
| [Glassdoor, all copywriters](https://www.glassdoor.com/Salaries/copywriter-salary-SRCH_KO0,10.htm) | $75,004 total pay; middle range $58K to $98K | July 3, 2024 | User-submitted plus model |

The two DR-specific sources put salaried pay at $63,042 to $79,286 a year. Freelance is separate: [Glassdoor's remote freelance copywriter page](https://www.glassdoor.com/Salaries/remote-freelance-copywriter-salary-SRCH_IL.0,6_IS12563_KO7,27.htm) shows $73 an hour (typical range $24 to $146, undated), and [ProCopywriters](https://www.procopywriters.co.uk/survey-2026/) puts the UK median day rate at £440. No source here covers DR project fees or royalties, where DR pay varies most.

## The core elements of direct response copy, briefly

The ranking pages agree on the basics: DR copy asks for one measurable action now, runs in ads, email, landing pages, sales letters, video sales letters and mail, and sells benefits over features.

Guides commonly list five elements: headline, a hook that speaks to the reader's pain or desire, an offer, urgency and a call to action. Of our 62 long-runners, 61 have a headline or strong first line, 51 name the reader's pain or desired outcome, 32 carry an offer term, 9 use urgency and 61 have a call to action button. All five appear together in 7 of 62.

## How we got this data

Pulls ran on 2026-09-23 and 2026-09-24. Ads (primary text) and Google results are US. One coder labelled everything under written rules; every label keeps its source URL.

The customer-words test used the oldest active ads on the brands' own pages: Bask & Lather, Chāmpo, LifeCell, Vibriance, PetLab Co., Wuffes, Dog is Human, Happy V, Silver Fern and RYZE, with Everyday Dose as a comparison brand. The 46 ads, 104 to 606 days old, gave 284 lines after we dropped lines under 4 words and calls to action. Customer sentences came from 15 YouTube review videos (of 60 screened), 198 Reddit comment and post URLs and 36 Amazon top reviews, minus sentences that mention an ad or sponsor: 29 to 1,611 per brand. Some ads sell another product from the same brand.

The fixed rule: a shared run of 4 content words (common words and brand names removed), or a fuzzy similarity of 85 out of 100. The looser check: a shared 4-word run with 2+ content words, or a shared 3-content-word run. A hit counted only with the same specific idea, 3+ content words or a fixed expression, and no brand, ingredient or category term. The matcher found 100 of 100 customer sentences we planted; if 1% of lines repeated a phrase in our comment sets, seeing none in 284 had a 5.8% chance.

The bigger-set check used Amazon reviews collected earlier for TruSkin, Anua, CeraVe, TropiClean and Physician's Choice (112 to 297 reviews each), plus 388 Wuffes reviews. 34 of 51 English ads on the three skincare pages were creator partnership posts.

The run-length study used 4 pulls of about 30 ads each, active and inactive, in the library's order, for 5 searches (dog joint supplement, hair growth serum, dark spot serum, sleep supplement, probiotic): 599 ads, merged into 280 distinct texts, minus 16 off-category or retargeting texts removed by hand. The groups come from 41, 37 and 36 advertisers; 35 of the 48 stopped ads ran 0 to 1 day. "New copy" only sees our pulls. Urgency hits came from a phrase list, each checked by hand, so 9 is a floor. The nearest earlier count, a [July 2026 Ad Library analysis](https://www.markifact.com/blog/competitor-ads-meta-ad-library-analysis) of 40 active ads from eight brands, skipped stopped ads, new entrants, urgency and the page-name split.

The AI threads were posted 2025-12-07 to 2026-08-19, and people who left the field are less likely to post there. In the research threads, 43 of 87 listed comments came back.

## FAQ

### What is direct response copywriting?

Copy that asks for one specific action now, such as a click, call or purchase, and is judged by how many people take it.

### Who is the best direct response copywriter?

We would not rank writers on self-reported sales totals. In our data, 21 of 62 lasting ads run from pages whose name is not the brand, so many lasting DR ads carry no writer's name.

### Is copywriting still worth it in 2026?

Salaried DR pay sits at $63,042 ([Salary.com](https://www.salary.com/research/salary/posting/direct-response-copywriter-salary)) to $79,286 ([PayScale](https://www.payscale.com/research/US/Job=Direct_Response_Copywriter/Salary)). Openings are the worry: writing roles fell to 1 in 30 agency postings ([The Drum](https://www.thedrum.com/news/as-writing-jobs-go-from-1-6-to-1-30-of-agency-postings-are-copywriters-in-trouble)), a smaller share of agency hiring, not zero.

### Is direct response copywriting a scam?

No. It is a method: one ask, measured results. The bad name comes from campaigns that sold weak products with false claims; judge a DR job by the product and its claims.

### Should ad copy use the customer's exact words?

Use their problems, objections and outcomes; the sentence can be yours. Across 700 ad lines we found 2 matches with the product's own public comments, both common phrasing. Our [voice of customer research guide](https://tryadlicio.com/voice-of-customer) covers how to collect and rank them.

Start with your own product: read the comments under its top three independent review videos and one Reddit thread about its problem, list the objections that repeat, and write one ad in your own words that answers the top one. Then test it.