LinkedIn video prospecting: the mechanics and the sequence

The mechanics: what LinkedIn actually lets you send

Most guides skip straight to strategy and leave people stuck on the part that actually blocks them, which is that LinkedIn behaves completely differently depending on which device you are holding. The recorder exists on mobile. It does not exist on desktop. That single asymmetry explains almost every confused question on this topic.

On the mobile app, video is a first-class message type. You record inside the thread, it plays inline when the prospect opens the conversation, and it looks like a message rather than a marketing asset. On desktop, there is no record-and-send button at all. You record somewhere else, then either upload the file or paste a hosted link that unfurls into a preview card in the thread.

The connection request is the strictest surface of the three. You get a 300-character note and nothing else. You cannot attach a video, and pasting a raw URL into a note you are sending to someone who has never heard of you is the fastest way to have it ignored. Once the connection is accepted, a 1:1 message to a first-degree connection gives you room to breathe, up to 8,000 characters of text alongside whatever you send.

1. 1

   Open the thread on the LinkedIn mobile app

   It has to be the app. The mobile web view does not expose the recorder, and neither does desktop.

2. 2

   Tap the plus icon beside the message box

   This opens the attachment tray: photo, video, file, GIF, and the camera.

3. 3

   Choose the video option, then record rather than upload

   Recording in-thread produces a native clip that plays inline. Uploading a file from your camera roll works too, but you lose nothing by recording live if the message is short.

4. 4

   Keep it inside 60 seconds

   LinkedIn cuts the recording at 60 seconds. Plan for 20 to 30 so you are never racing the timer, and so the last sentence is not the one that gets clipped.

5. 5

   Add one line of text above the clip, then send

   The text line is what gets read in the notification preview. Treat it as the subject line for the video, because functionally that is what it is.

LinkedIn changes these limits without announcing them. Everything in this table was checked as of the publication date at the bottom of this page. If you are reading it much later, spend thirty seconds confirming the mobile flow before you build a campaign on it.

Native clip or hosted link: the real tradeoff

This is the decision that quietly determines whether your LinkedIn video program is measurable. A native clip is the more organic-feeling object and the less useful one. A hosted link is the more useful object and the more obviously commercial one.

The deciding factor is usually not aesthetics, it is whether you need to know what happened. LinkedIn gives you no view data on a native clip. You will never know whether your video was watched, skipped, or watched twice, which means you cannot tell a script problem from a targeting problem. On a cold first video touch, that blindness is expensive. On a warm thread with someone who already replied to you, it does not matter at all.

• Use a hosted link for the first video touch in a cold sequence, because you need the watch data to know whether the script or the list is at fault.
• Switch to native clips once the prospect has replied at least once. At that point the thread is a conversation and the analytics stop earning their keep.
• Whichever you use, put one line of text above it. A bare video with no context reads as an ad.
• Never send a link in the connection note itself. Save it for after the accept, when you have an actual thread.

If you cannot measure it, you cannot improve the script. That is the argument for the hosted link on cold touches, and the argument against it once someone is already talking to you.

Where video belongs in the sequence, and it is not first

The strongest and least intuitive rule in LinkedIn video prospecting: do not lead with the video. Send text first. Spend the video after the connection has been accepted, or after the prospect has engaged with something you sent. The pages that rank for this query agree on it, and the reasoning holds up.

A personalized video is the most expensive asset in your cadence, whether the cost is your time or a generation credit. Spending it on someone who has not accepted your connection request means most of those videos are never seen by anyone. Worse, a video attached to a first touch from a stranger reads as a sales artifact at exactly the moment the prospect is deciding whether you are a person or a campaign.

There is also a content reason. The video is only good when it can reference something specific, and specificity is easier to earn after one exchange. Gong's analysis of more than 28 million cold emails found that pitching reduces reply rates by as much as 57% (Gong, 2025). A cold video that opens with a pitch inherits that penalty and adds a click on top of it. Sopro's 2026 State of Prospecting research found 57% of sales and marketing decision-makers believe most sales outreach feels impersonal, which is the demand you are trying to meet, and you do not meet it by leading with a produced video about your product.

• Touch 1: connection request, text only, 300 characters, one specific reason it is them.
• Touch 2: a short text DM after the accept. Two sentences. One question. No link.
• Touch 3: the video. This is where it earns its cost, because the thread now exists and the person opted in.
• Touch 4: a text nudge that references one concrete thing from the video, so it is obvious the video was not generic.
• Touch 5 onward: move to email, or close the loop politely. Saleshandy's 2026 analysis of 53.1 million cold emails found 44% of all positive replies came from follow-up messages rather than the first send, so the later steps are not filler.

The exception: a genuinely warm inbound, an event contact, or someone who just engaged with your post. There, the video can come first, because the context is already established.

Four sequence blueprints with branch logic

Tips do not survive contact with a live campaign. Blueprints do. Each of these is a complete play with the branch and fallback logic written out, segmented by the kind of team running it, because a founder sending twelve messages a week and an agency running six sender accounts should not run the same cadence.

Read the arrows as branches. Anything marked as a stop means stop, including the ones that feel premature. The most common failure mode in LinkedIn video prospecting is not sending too few touches, it is continuing to send after the prospect has told you, through silence, what the answer is.

DAY 0   Connection request. No video. No link.
        300-char note: one line on why them, specifically.
          |
          +-- NOT ACCEPTED by day 7 --> withdraw the invite,
          |                             move to email-only,
          |                             retry LinkedIn in 90 days
          |
          +-- ACCEPTED
                |
DAY 0+2h  Text DM. Two sentences. One question. No link.
                |
                +-- REPLIED --> stop the sequence, rep takes over
                |
DAY 3     VIDEO DM, 20 to 30s. Profile-walkthrough pattern.
          Hosted link so you get watch data.
                |
                +-- WATCHED, NO REPLY --> day 6 text nudge that
                |                          references one specific
                |                          thing said in the video
                |
                +-- NOT WATCHED --------> day 6 text nudge on a
                                           different angle. Do not
                                           re-send the same video
                |
DAY 10    Email step to the same person. Subject references
          the LinkedIn thread so it is not a cold restart.
                |
DAY 14    Breakup DM. One line. No ask. No video.

DAY 0   Comment on their recent post. Substantive.
        No DM. No connection request yet.
          |
DAY 2   Connection request referencing that comment.
          |
          +-- NOT ACCEPTED --> stop. Do not chase.
          |                    Founders spend credibility faster
          |                    than SDRs do, and cannot get it back
          |
          +-- ACCEPTED
                |
DAY 3     VIDEO DM, 25 to 35s. Recorded or generated for this
          one person. Name the specific thing you read.
          Founder on camera. No slides. No music.
                |
                +-- REPLIED --> keep talking. Do not pivot to a
                |               booking link on the first reply
                |
                +-- NO REPLY --> day 8 text DM: one sentence,
                                  one question, no repeat of
                                  the pitch
                |
DAY 16    Stop. Move to a quarterly check-in list.

POOL: 3 to 6 sender accounts, each with its own daily
      ceiling, rotating across one shared prospect list.

DAY 0   Sender A: connection request, text only.
DAY 2   Sender A: text DM after the accept.
DAY 5   Sender A: VIDEO DM. One base recording per client,
        per-prospect script on top of it.
DAY 9   Sender A: text nudge.
DAY 12  Email step, sent from the client's domain,
        not the agency's.
DAY 21  Close the LinkedIn thread. Keep the email cadence
        running if it is still producing.

BRANCHES
  Reply on any step ........ pause every other channel
                             immediately, across all senders
  Sender hits daily ceiling  queue the step, do not overflow
                             into a different sender mid-thread
  Sender restricted ........ quarantine that account. Do NOT
                             reassign its live threads to
                             another sender. The prospect will
                             notice the name change
  No accept by day 10 ...... drop from LinkedIn entirely,
                             continue on email only

SEGMENT: closed-lost 6+ months ago, or a thread that
         went quiet after a demo.

DAY 0   Text DM. Name the specific reason it stalled.
        No pitch. No link. No "just checking in".
          |
          +-- REPLIED --> human takes over immediately
          |
DAY 4   VIDEO DM, 20 to 30s. One thing that changed since,
        stated plainly. Reference the old conversation by
        detail, not by date.
          |
          +-- WATCHED -------> day 9 single-question DM
          |
          +-- NOT WATCHED ---> day 9 email carrying the same
                                video link. Different channel,
                                same asset
          |
DAY 18  Stop. Re-enter the segment in two quarters.

LinkedIn scripts run shorter than email scripts

A video sitting in an email can afford 60 to 90 seconds because the reader has already committed to opening a message. A video sitting in a LinkedIn DM cannot. It is competing with a feed, it is usually watched on a phone between other things, and the hard 60-second native cap sets the ceiling anyway. The working band is 15 to 30 seconds and 15 to 40 words.

That constraint is a gift. Vidyard's 2025 Video in Business Benchmark Report found that videos under one minute hold 65% of viewers all the way to the end, the highest completion rate of any length band, while videos over 20 minutes retain only 20%. On LinkedIn you are structurally forced into the good end of that curve.

Forty words does not leave room for a pitch. It leaves room for one observation and one question, which is the correct shape for a first video anyway. Every script below is written to that budget.

[On screen: their LinkedIn profile, scrolling slowly.
 Your face in the corner.]

"Hi {{firstName}}, I'm literally on your profile right
now. Three of the last four things you posted are about
onboarding drop-off. That's the entire reason I'm
reaching out. Worth ten minutes?"

"{{firstName}}, your post about killing the SDR quota
model. I've watched two teams try exactly that and fail
for the same reason. Thirty seconds of context, or
should I just write it out?"

"Thanks for connecting, {{firstName}}. Quick context so
I'm not a stranger in your inbox next week: we work with
{{jobTitle}}s at companies around {{company}}'s size on
one narrow problem. I'll send it in writing so you can
ignore it properly."

"{{firstName}}, saw {{company}} opened the Austin office.
Every team I've watched do that hits the same wall in
month three: the new reps have no local proof to point
at. Two companies solved it in ways worth stealing.
Want the short version?"

"{{firstName}}, I've sent you three things and you've
been polite enough to ignore all of them. I'll stop. If
onboarding drop-off ever moves up your list, I'm here.
Good luck with the quarter."

Read every script out loud with a timer before you use it. Written word counts lie. Spoken pace is what the prospect experiences, and most people run 15% longer than they expect on camera.

Account safety, sending limits, and how many senders

LinkedIn does not publish a single hard number for connection requests or messages, and the ceilings it enforces move with account age, profile completeness, acceptance rate, and how many of your invitations sit unanswered. Anyone quoting you an exact weekly figure is quoting a community estimate, not a documented limit. Treat your own observed ceiling as the real one.

The pattern that gets accounts restricted is not video specifically. It is volume that arrives faster than a human could produce it, combined with a low acceptance rate, which is the signal LinkedIn reads as spray. Video prospecting is actually the safer end of the spectrum on volume, because the whole premise is a smaller, better-researched list. The risk shows up when teams try to run a video program at spray volume and get the worst of both.

On sender count, the honest answer is arithmetic rather than a recommendation. Divide your weekly connection target by the conservative per-account weekly ceiling you have personally observed on a warmed account, then add one for headroom. Adding senders raises capacity. It does not raise reply rate, and a fleet of cold accounts sending identical messages is a faster route to restriction than one warm account sending fewer.

• Warm a new sender account for at least two to three weeks before it sends anything sequenced. Profile completed, some posting, some organic engagement.
• Watch acceptance rate as your health metric, not send volume. A falling acceptance rate is the leading indicator that your targeting drifted, and it is also what LinkedIn is watching.
• Withdraw invitations that have gone unanswered for two to three weeks. A large pile of pending invites suppresses your ceiling.
• Never move a live thread between sender accounts. The prospect sees a different name and the conversation is over.
• Ramp new accounts gradually rather than starting at your target volume on day one. Step changes look automated, because they are.
• If an account gets restricted, quarantine it rather than pushing its workload onto the others. Redistributing the same volume across fewer healthy accounts is how one restriction becomes three.

Sender count is a capacity decision, not a performance decision. If your reply rate is bad, more senders will produce more of the same bad result, faster.

Benchmarks, and the gap nobody fills

Judging a LinkedIn video campaign requires knowing what normal looks like without video, so here is the context, with sources named. Everything below is category data from third parties. None of it is an Outvid result, and none of it should be read as a promise about your list.

• Run your own comparison before you believe anyone's, including ours. Same list, same offer, same week, video on one arm and text on the other.
• Measure watch rate and reply rate separately. A high watch rate with a low reply rate is a call-to-action problem. A low watch rate is a thumbnail, subject-line, or targeting problem.
• Give it enough volume to mean something. Two hundred prospects per arm is the floor before a difference in reply rate is worth acting on.
• Vendor-reported figures, including the HeyReach numbers above, come from campaigns run on that vendor's platform by customers who chose to run video. Treat them as directional, not as a forecast.

Worth stating plainly: our research did not surface a single verified, methodologically transparent head-to-head study comparing video and text reply rates on LinkedIn specifically. The multiples you will see quoted around this topic mostly trace back to vendor marketing that cites other vendor marketing. We would rather tell you the number does not exist than repeat one we cannot stand behind. Use the baselines above as context and your own A/B as evidence.

LinkedIn is one channel, not the channel

Treating LinkedIn as standalone is the most common structural mistake in video prospecting. The prospect does not experience channels, they experience you. A LinkedIn video that goes unwatched and an email that goes unopened are two failed attempts at the same conversation, and running them separately means neither one knows about the other.

The volume argument is straightforward. Gong's 2025 analysis found the average rep sends 344 cold emails to land one meeting, and Saleshandy's 2026 data across 53.1 million cold emails puts the average reply rate at 3.7% with a 21% open rate. Email is a volume channel with thin margins. LinkedIn is a lower-volume, higher-context channel. The video belongs on the side where context is highest and the audience is smallest, which is why the blueprints above spend it on LinkedIn and use email as the persistent follow-up layer.

The practical version: one researched script per prospect, one video asset, and a cadence that carries it wherever the prospect actually is. Reply on any channel pauses every other channel immediately. That last rule sounds obvious and is the one most teams get wrong.

• Reuse the same video across LinkedIn and email rather than producing two. The asset is the expensive part, the distribution is not.
• Change the framing per channel, not the video. LinkedIn gets one line above the clip. Email gets a subject line, a thumbnail, and two sentences of context.
• Sequence the channels so they do not collide. A LinkedIn DM and an email landing within an hour of each other reads as automation, because it is.
• Suppress across all channels on any reply, including replies that are a no. Continuing to send after a decline is what turns a soft no into a blocked account.

Outvid runs this end to end: you record one short base video, it generates a per-prospect script from real research, keeps your actual recorded face on camera, and sends across email, LinkedIn, WhatsApp and Instagram from one sequence with the reply suppression already wired in.