LinkedIn benchmark v3.0.0 · 26 tests
Know what your post will do before the feed does.
Paste a draft. Foldline scores it against up to twenty-six weighted tests drawn from how the 2026 LinkedIn feed actually ranks — semantic relevance, dwell, saves, suppression — and hands back a modelled reach index, the exact lines costing you distribution, a before-and-after as you fix it, and a corrected version that pastes into the composer clean.
26weighted tests, scoped by format
210characters before the fold cuts
R0to score, no signup
YN
Your name
Headline aligned to this topic
Now
Most posts don't fail because the idea is weak. They fail in line two.
Last quarter I rewrote the openings of 41 client posts and changed nothing else.
Median impressions went from 780 to 2,340.
Three things…see more
619 characters hidden until they tap
LikeCommentRepostSend
The workbench
Paste the draft. Get the verdict.
The score is free and covers every test that applies. The itemised report, the reach breakdown and the one-click rewrites unlock with a credit. Nothing you paste leaves your browser.
The draft
144 words · 829 charactersFirst comment
No link anywhere. Fine — only add one if the post genuinely needs it.
How it’s going out
96score
Clear
Ship it
Both blocking tests pass and the craft rules hold. Post weekday morning and be present for the first 60-90 minutes.
114 of 119 points that apply to a text-only post
1.14×Modelled reach index against a clean baseline post from the same account. Directional, not a guarantee.
SavesHighKeepable - framework or list structure
CommentsHighCloses on an answerable question
DwellHigh144 words, 11 blocks
Unlock the full report
Item-by-item scoring across every applicable test, the reach multiplier breakdown, one-click rewrites and the copy-ready post. R49 for this post.
Credits never expire.
Where the reach goes
1.14×Text-only×0.90
3+/week, 8+ weeks×1.15
Author present in hour one×1.10
Craft & compliance
88 of 119 ptsHow the post is built, and the behaviours the platform suppresses outright.
✓
Hook lands in the first 1-2 linesLI-01 · 10 pts
Opens in 151 characters with tension or specificity.
✓
One clear idea, audience and takeawayLI-02 · 5 pts
One idea, declared and within length.
✓
Original, first-hand insightLI-03 · 12 pts
First-person, specific, and carries a point of view.
!
Format fits the jobLI-04 · 10 pts
This is a framework or step list - documents carry roughly 1.4× the reach of text-only and hold dwell on every swipe.
✓
External link disciplineBlockingLI-05 · 10 pts
No link anywhere: the highest-reach option in the 2026 data.
✓
No engagement baitBlockingLI-06 · 8 pts
Closes on a question a real reader could answer.
✓
Built for the thumbLI-07 · 5 pts
11 short blocks, longest 122 characters.
✓
On one of this account's 2-3 themesLI-08 · 8 pts
Stays on a defined theme, so topical authority compounds.
✓
0-3 relevant hashtags, no stacksLI-09 · 2 pts
2 hashtags - within range.
✓
Right account and posting timeLI-10 · 5 pts
Personal profile, weekday morning.
✓
Present for the first 60-90 minutesLI-11 · 7 pts
Early replies of 15+ words carry far more weight than a short one - and they pull the thread back up.
✓
Within LinkedIn's character limitBlockingLI-26 · 6 pts
829 characters, inside the 3,000 limit.
2026 interest graph
31 of 119 ptsHow the ranking model reads the post, its author, and who it should reach.
✓
Profile matches what you post aboutLI-12 · 6 pts
Headline, experience and topic line up - the model reads author and post together.
✓
Names the reader it is forLI-13 · 5 pts
A specific role, addressed directly - the interest graph needs someone to match it to.
✓
Knowledge and advice, not an updateLI-14 · 6 pts
Teaches something. Advice content is the class that still gets multiples of baseline reach.
✓
Length in the dwell sweet spotLI-15 · 4 pts
829 characters. The 2026 sweet spot is 800-1,000 - long enough to earn dwell, short enough to finish.
✓
Reads human, not model-generatedLI-16 · 5 pts
No obvious model tells - varied rhythm, concrete detail.
✓
Cadence builds topic authorityLI-17 · 5 pts
3+ posts a week for 8+ weeks on the same pillars - the consistency that lifts baseline reach per post.
Fix list, in priority order
0 automatic fixes1
Format fits the jobLI-04 · 10 pts
This is a framework or step list - documents carry roughly 1.4× the reach of text-only and hold dwell on every swipe.
The benchmark
Every draft clears every test that applies to it.
Eleven tests cover craft and the suppression behaviours LinkedIn penalises outright. Six come from the 2026 interest-graph rebuild, where an in-house ranking model reads the post for meaning and matches it to specific readers. Eight more score the asset — image, carousel or video — and apply only to the format you are posting, so a text post is never marked down for lacking a video hook. Three are blocking — a link in the body, engagement bait, or a post over LinkedIn's 3,000-character limit — and failing any one makes the verdict Fix before publishing no matter how well the rest scores. Full benchmark, with weights and evidence grades →
LI-01
Hook lands in the first 1-2 lines
Only ~210 characters show before “see more”
10 ptsOnly ~210 characters show before “see more”
LI-02
One clear idea, audience and takeaway
Split posts get matched to nobody
5 ptsSplit posts get matched to nobody
LI-03
Original, first-hand insight
A real example, result or point of view
12 ptsA real example, result or point of view
LI-04
Format fits the job
Documents and video carry the most reach
10 ptsDocuments and video carry the most reach
LI-05
External link discipline
Blocking — body links cost you distribution
10 ptsBlocking — body links cost you distribution
LI-06
No engagement bait
Blocking — explicitly suppressed phrasing
8 ptsBlocking — explicitly suppressed phrasing
LI-07
Built for the thumb
Short lines, white space, no walls
5 ptsShort lines, white space, no walls
LI-08
On one of this account's 2-3 themes
Topic-hopping resets your authority
8 ptsTopic-hopping resets your authority
LI-09
0-3 relevant hashtags, no stacks
Distribution is semantic now
2 ptsDistribution is semantic now
LI-10
Right account and posting time
People out-reach pages
5 ptsPeople out-reach pages
LI-11
Present for the first 60-90 minutes
Early real replies set the trajectory
7 ptsEarly real replies set the trajectory
LI-12
Profile matches what you post about
Author and post are scored together
6 ptsAuthor and post are scored together
LI-13
Names the reader it is for
The interest graph needs a match
5 ptsThe interest graph needs a match
LI-14
Knowledge and advice, not an update
Announcements are demoted hardest
6 ptsAnnouncements are demoted hardest
LI-15
Length in the dwell sweet spot
800–1,000 characters
4 pts800–1,000 characters
LI-16
Reads human, not model-generated
Em-dashes, flat rhythm, no numbers
5 ptsEm-dashes, flat rhythm, no numbers
LI-17
Cadence builds topic authority
3+ a week, held for 8+ weeks
5 pts3+ a week, held for 8+ weeks
LI-26
Within LinkedIn's character limit
Blocking. Over 3,000 characters cannot be posted at all
6 ptsBlocking. Over 3,000 characters cannot be posted at all
LI-18
Alt text written for the asset
Alt text is the only part of the asset the model reads
4 ptsAlt text is the only part of the asset the model reads
LI-19
The image carries meaning
Decoration costs feed height and earns nothing
6 ptsDecoration costs feed height and earns nothing
LI-20
Aspect ratio owns feed height
Portrait and square own the most screen
4 ptsPortrait and square own the most screen
LI-21
Cover slide carries the hook
The cover is the whole swipe decision
8 ptsThe cover is the whole swipe decision
LI-22
One idea per slide, legible on a phone
One idea per slide, readable at arm's length
6 ptsOne idea per slide, readable at arm's length
LI-23
Hook in the first 3 seconds
Three seconds is the whole audition
8 ptsThree seconds is the whole audition
LI-24
Captions burned into the video
The feed autoplays muted
7 ptsThe feed autoplays muted
LI-25
Video length in the retention band
30–90 seconds
5 pts30–90 seconds
Pricing
Pay per post. No subscription trap.
One credit scores one draft — unlimited re-runs on that draft while you fix it. Prices in ZAR, VAT included.
Single post
R49
one credit
- ✓Full report on every applicable test
- ✓Reach index with the multiplier breakdown
- ✓One-click rewrites and copy-ready output
- ✓Unlimited re-scores on that draft
Ten-pack · most bought
R349
R35 a post · credits never expire
- ✓Everything in Single post, ten times
- ✓Saved drafts and score history
- ✓Carousel slide planner
- ✓Weekly cadence tracker
Agency
R1 490/mo
unlimited posts, up to 10 client profiles
- ✓Unlimited scoring across every client
- ✓Per-client theme locking and voice profile
- ✓White-label PDF report for client approval
- ✓Team seats and shared credit pool
Checkout is not connected yet — the buttons go to the contact form. Try it on this run in the report panel unlocks the full output so you can see exactly what a credit buys.
Questions
The honest answers.
Can you actually predict my reach?
No, and anyone selling you a number is guessing. Foldline gives you a relative index: how this draft compares to a clean baseline post from the same account, based on the format, suppression and quality signals that published 2026 breakdowns agree on directionally. The value is the diagnosis — the specific line costing you the fold, the link halving your distribution — not the decimal place.
What happens if my post is too long?
LinkedIn's hard limit is 3,000 characters and the composer simply refuses anything longer, so test LI-26 treats it as blocking: the verdict becomes Fix before publishing and the note tells you exactly how many characters to cut. Length also affects LI-15 separately, which is about the 800–1,000 character dwell band — that one costs you reach, this one stops you posting at all.
Why does my carousel score out of a different number to my text post?
Because eight of the twenty-six tests only apply to the asset. A text post is scored out of 119 points, a carousel out of 141, a video out of 147 — tests that do not apply are dropped from the denominator as well as the result, so the percentage stays comparable. A text post is never marked down for lacking burned-in captions.
Can Foldline actually see my image or video?
No. It reads text. The asset tests score what you declare — that captions are burned in, that the cover slide carries the claim, that the alt text is written. They are stated intentions, and the report labels them that way. The one thing it does read is your alt text, which is also the only part of the asset the ranking model reads.
Will the copied post keep its formatting on LinkedIn?
That is what the copy button is for. The composer takes plain text only, so before copying, Foldline strips markdown that would show up literally, converts - bullets to real • characters, removes any unresolved [rewrite:] markers left over from a fix, collapses runs of blank lines, deletes invisible non-breaking and zero-width characters, and writes to the clipboard as plain text so no styling rides along. It tells you everything it changed.
Can I see exactly what a fix changed?
Yes. Once unlocked, the report shows your original draft next to the revised one — removed words struck through on the left, added words highlighted on the right. It updates after every fix, and the same view covers the first comment if that changed too.
Why is a link in the post body such a big deal?
Because it sends the reader off-platform, and the feed is built to keep them on it. Published estimates for the penalty range from roughly a fifth to about half of median reach. Foldline treats it as blocking and shows the fix: strip it from the body, signpost “link in the comments”, and park the URL in your own first comment the moment you publish.
Is this affiliated with LinkedIn?
No. Foldline is an independent tool built by DB23 in Cape Town. It is not affiliated with, endorsed by, or connected to LinkedIn Corporation, and it does not connect to your account, read your analytics, or post on your behalf. You paste text; it scores text.
Do you store my drafts?
No. Scoring runs entirely in your browser — nothing is sent to a server. Only your original draft and your baseline score are kept, in your own browser’s local storage, so the before-and-after — including the side-by-side revision — survives a refresh.
Foldline is an independent tool by DB23. Not affiliated with, endorsed by, or connected to LinkedIn Corporation. Reach figures are directional models, not guarantees.