LinkedIn document ads made a lead for $142. Most advertisers skip the format.
The cheapest lead on the channel, from the format the fewest advertisers run.
LinkedIn document ads made a lead for $142 in 2025, against $200 for an image ad, $265 for video and $362 for conversation ads, and turned 11.9% of clicks into leads, the best rate we publish. 51 advertisers ran them; 127 ran image ads. The budget call is a swap, not new money. What we cannot tell you is what those leads were worth: for every format except image ads, one advertiser held too much of the closed-won revenue for the number to mean anything, so the cost-per-customer figure is withheld (why).
Ship one document ad on LinkedIn this week from an asset you already own, 8–12 slides, ungated, mid-funnel, and launch an image ad against it the same day. In metadata.io’s 2025 benchmark, document ads made a lead for $142 against $200 on image (51 advertisers against 127) and turned 11.9% of clicks into leads. Judge the pair on closed-won revenue in your CRM over two full sales cycles, not on lead price, this report’s cost-per-customer number for the format is withheld (why).
2025 · 51 advertisers on LinkedIn document ads · B2B · only groups with at least $50K spent · Methodology · What we withheld, and why
Executive summary
- The economics: across 51 advertisers, document ads made a lead for $142 against $200 on image and $265 on video, and turned 11.9% of clicks into leads.
- The budget call: fund a document-ad line out of image budget this quarter, a swap, not new money, and decide before it launches that closed-won revenue in the CRM is what settles it.
- The limit, stated plainly: this is a finding about what a lead cost and how many clicks turned into leads. The cost-per-customer and payback numbers for each format are withheld (why), so nothing here says a document lead is worth more than an image lead.
Playbook
- The play: ship a document ad from an asset you already own, ungated, mid-funnel, built to be read in the feed.
- The setup: cut it to 8–12 slides with the payoff in the first three, and launch an image ad against it the same day.
- The measurement: closed-won revenue per dollar in the CRM, document against image, over at least twice your average sales cycle, not cost per lead, and not four weeks if your deals take six months.
Which LinkedIn ad format buys the cheapest lead?
Document ads, by a wide margin. On 2025 LinkedIn campaigns set to collect leads, the typical document ad made a lead for $142 across 51 advertisers, 29% below image ads at $200, 46% below video at $265, and 61% below conversation ads at $362, and turned 11.9% of clicks into leads, the best rate we publish. The cheapest format on the channel is also the one fewest people run: 51 advertisers ran documents against 127 running image ads.
Does the cheap lead turn into revenue?
We cannot tell you from this report, and we would rather say so than imply it. On 2026-08-17 every format but one broke our rule that no one advertiser is allowed to swing a number: once the transactional-CRM advertiser was removed, a single account still held more than 40% of the remaining closed-won revenue behind the document, video and conversation numbers. Those numbers exist in our working files and we did not publish them, withheld (why), and the format ranking we had drafted is now entry nine in the kill list, along with what it cost us to drop it.
One format did survive: image ads, at 0.50x, 50 cents back on the dollar in year one, $59,488 per customer, and a 24.3% close rate counting only deals that actually closed or actually died, across 121 advertisers. Across the whole dataset the same numbers read 0.56x and $58,887. One row cannot rank formats, so we print it for context and rank nothing. Read this page as what it is: a finding about lead cost solid enough to act on, with the revenue question left open.
What should you change?
Fund a document-ad line out of the image budget: document ads made a lead for $142 against $200 on image, and turned 11.9% of clicks into leads (51 advertisers and 127). Approve it as a cost test with a CRM readout, not as a revenue promise. The revenue number for this format is withheld.
Approve a head-to-head test inside programs you already run, same channel, same audiences, same offer, no extra budget, and let it run at least twice your average sales cycle. Ask for two numbers back, never one: what a lead cost and closed-won revenue per dollar, document against image. A four-week readout on lead price is exactly how a cheap lead gets mistaken for a good one.
“Documents win everywhere.” This came from LinkedIn, and it is about cost only; other channels do not offer the format. $142 across 51 advertisers is a LinkedIn lead price, not a law of nature, and it says nothing about revenue.
Ship your first document ad this week from an asset you already own, a benchmark deck, playbook or teardown cut to 8–12 slides, run ungated mid-funnel: the format turned 11.9% of clicks into leads and made a lead for $142 across 51 advertisers. Put the payoff in the opening slides, LinkedIn previews them before any form appears.
Launch the image ad against it, same offer, same audience, same day. The readout is what a lead cost and closed-won revenue per dollar in the CRM over a full sales cycle, both numbers, or the test taught you nothing.
“Any PDF will do.” Sales brochures are why document tests die; buyers save teardowns, checklists and playbooks. The $142 lead, and the 11.9% of clicks that became leads across 51 advertisers, were earned by content worth reading in the feed.
What are the exact numbers?
Every LinkedIn ad format with a cost we could publish, and every revenue number we had to withhold. Lead cost and click-to-lead come from the $31.5M of campaigns set to collect leads; payback and cost per customer come from the $29.4M, 154,000-lead, 127-advertiser set we could follow into the CRM. Revenue we can trace back to an ad, inside 12 months.
| Ad format | Advertisers (lead cost · revenue) | What a lead cost | Clicks that became leads | Back on the dollar, year one | Cost per customer |
|---|---|---|---|---|---|
| Document | 51 · 48 | $142 | 11.9% | withheld | withheld |
| Image | 127 · 121 | $200 | — | 0.50x | $59,488 |
| Video | 81 · 50 | $265 | — | withheld | withheld |
| Conversation | 42 · 42 | $362 | — | withheld | withheld |
| All formats together, after the trims | , · 127 | — | — | 0.56x | $58,887 |
Dashes: numbers that live in the interactive explorer rather than being quoted here, or click-to-lead numbers that did not have enough advertisers behind them to stand alone. Withheld: the document, video and conversation revenue numbers broke the 40% one-advertiser limit on 2026-08-17. The reason, and what dropping them cost us, are in the kill list. The image-ad close rate counts only deals that actually closed or actually died, on deals an ad touched, splitting the credit evenly across the ads involved, with deal status as of 2026-08-16: 24.3% against 21.2% across the whole dataset.
How do we know?
Every number above passed the same publication gates before it reached this page. In brief:
- Advertiser counts: the LinkedIn cost numbers we publish rest on 51 advertisers running document ads, 127 image, 81 video and 42 conversation; the one revenue number we publish rests on 121 image advertisers. The advertiser count sits beside every number on this page, and a revenue number needs at least 25 advertisers or 30 wins behind it.
- No one advertiser is allowed to swing a number, on either side: no single advertiser above 40% of a published group’s spend or above 40% of its closed-won revenue. That second half removed the document, video and conversation revenue numbers on 2026-08-17. Withheld, not estimated, reasons in the kill list.
- Spend floor: we publish a number only where at least $50,000 was spent.
- Scope and window: lead cost and click-to-lead come from the $31.5M of lead-collecting campaigns inside $57.6M of analyzed 2025 spend; payback and cost per customer come from the $29.4M, 154,000-lead, 127-advertiser set we could follow through the CRM to revenue that actually closed, over 12 months. Revenue we can trace back to an ad only, not proof the ads caused it, not profit, not lifetime value.
- This is what happened, not a controlled test: every figure describes 2025 as it ran for these advertisers, so treat it as a pattern to check, not a guaranteed cause.
- What else could explain this: who runs the format. The 51 advertisers who picked an unfashionable format may simply be better media buyers, and that edge would travel with them, not with the format. The other candidate is what is being offered: documents usually carry mid-funnel content, which collects leads more cheaply than a demo request whatever format it runs in. Neither can be ruled out here, and the revenue evidence that might have separated them is withheld. Treat the cost gap as a pattern worth testing, not a proven format effect.
Full rules in the Closed-Won Protocol, including the nine claims we refused to publish, with the exact reason each one was killed and the number it cost us.
Cite as: Metadata 2026 B2B Ad Spend Benchmark (n=153 advertisers, $57.6M, 2025), metadata.io/benchmark-report-2026
What do you do different tomorrow morning?
One named change per seat. The columns reorder to match your reading lens; both are always on the page.
Fund the underused format
The one change: carve a document-ad line into every LinkedIn lead program this quarter. Paid for out of image-ad budget, not new money.
- This is a way to buy the same thing cheaper, not a new-channel bet: same channel, same audiences, same offers, a lead at $142 instead of $200 (51 advertisers and 127).
- Because the money moves inside programs you already run, the downside is a four-week test; the upside compounds every quarter the format stays cheap and uncrowded.
- Ask for two numbers back: what a lead cost, and closed-won revenue per dollar in the CRM, document against image, same offer, over at least twice your average sales cycle. Judging on lead price alone is what got the industry here.
What to put IN the document
The one change: ship your first document ad this week from an asset you already own, ungated, mid-funnel, built to be read in the feed.
- Pick something genuinely useful: a benchmark deck, a playbook, a teardown, a checklist, the things buyers save. Sales-brochure PDFs are why most document tests die.
- Cut it to roughly 8–12 slides and put the payoff in the first three, LinkedIn previews the opening slides before any form appears.
- Run it ungated and mid-funnel: let people read it in the feed, with a form on the expand for the full version.
- Launch the image ad against it, same offer, same audience, same day, so the four-week readout is clean.
- Report closed-won revenue per dollar beside what a lead cost, never lead price on its own, and let that settle it.