Video prospecting statistics, with every source shown

Our sourcing standard, stated up front

Most pages ranking for video prospecting statistics recycle each other. A vendor blog quotes a second vendor blog, which quotes a webinar slide, which quotes nothing at all. Follow three or four of those chains and you usually land on a dead link or a page that no longer contains the number attributed to it. That is why the same handful of dramatic percentages appear on dozens of sites with no methodology anywhere behind them.

So we set a bar before writing a word, and applied it to every figure on this page.

1. 1

   A named organisation, not a category

   “Studies show” and “industry data suggests” are not sources. Every figure below names the organisation that actually collected or published it.

2. 2

   A year attached to the number

   Outreach benchmarks move. A reply rate from 2019 describes a different inbox than a reply rate from 2026, so the year travels with the figure rather than being dropped to make it look current.

3. 3

   Sample size where the publisher disclosed one

   Where the publisher stated a sample, we carry it: 53.1 million cold emails, 7.5 million cold emails, 96,051 LinkedIn campaigns, 656 revenue professionals, 266 marketers. Where no sample was published, we say the figure is survey-reported and leave it at that rather than inventing a denominator.

4. 4

   A live link we opened ourselves

   Each source URL in the table below was fetched and the figure confirmed on the page. If the page had moved, we linked the new location rather than the URL everyone else still cites.

5. 5

   Exclusion when the chain dead-ends

   If a number could not be traced to a reachable primary source with a stated methodology, it does not appear as fact anywhere on this page. It goes in the excluded list instead, with an explanation. Seven figures ended up there.

One housekeeping note that matters if you are checking our work: the Vidyard URL almost universally cited for video benchmarks, vidyard.com/blog/video-benchmarks/, now returns a 404. The report itself is live at vidyard.com/business-video-benchmarks/, and that is the URL we link. If you are maintaining an older post that cites the blog path, it is worth updating.

Adoption: video outreach is no longer the unusual choice

The most common objection to video prospecting used to be that it was a novelty. The adoption data no longer supports that framing, in either direction: video is widespread enough that it is not exotic, and widespread enough that being on camera is not by itself a differentiator any more.

Vidyard analysed 943,305 videos created by its customers in 2024, 88% more than the prior year, and reported a 241% increase in videos created by individual users — roughly the difference between one video a month and one every ten days. In its Future of Revenue Report (2025), based on 656 sales and marketing professionals surveyed in April 2025, 93% of revenue leaders said they were using video. Wyzowl’s 2026 survey of 266 respondents put business video adoption at 91%, and Wistia found 76% of teams making at least one video a month.

The distribution picture has shifted too. Wistia’s 2026 report has 81% of teams sharing video on LinkedIn, which has overtaken YouTube as the primary B2B video channel. That is the practical case for treating LinkedIn as a first-class video surface rather than an afterthought to email.

The AI line is the one to read carefully. Vidyard reported a 12x increase in adoption of AI Avatars in eight months, and Wyzowl found 63% of video marketers had used AI video tools, up from 51% a year earlier. Those are adoption numbers, not performance numbers — nobody in that data has published what the AI-generated videos actually returned. Treat the trend as real and the outcome as unproven.

• Adoption is near-universal at the team level, which means “we send video” is table stakes, not positioning.
• Individual sending frequency is rising faster than team adoption, so per-rep volume is where the change is happening.
• LinkedIn now outranks YouTube as the B2B distribution channel, per Wistia (2026).
• AI video tooling crossed the majority threshold among video marketers in a single year — with no published performance data attached.

Cold email baselines: why the reputable numbers disagree so wildly

Any claim about video lifting replies is meaningless without a baseline, and this is exactly where most video statistics pages cheat. They pick the lowest cold-email number they can find, put the highest video number next to it, and present the gap as an effect. The three most defensible baselines we found disagree by more than a factor of ten, and the disagreement is methodological rather than one of them being wrong.

• Denominator: Belkins divides by every email sent. A per-contact or per-campaign denominator can move the same underlying performance up several multiples.
• Unit of counting: unique replies per contact is a far stricter count than total replies, which double-counts threads.
• List construction: a managed-service dataset with human curation is not comparable to open platform data where anyone can upload a scraped list.
• Sequence depth: Saleshandy found 44% of all positive replies came from follow-up emails, so a dataset with more follow-up steps will report a higher headline reply rate for identical messaging.

The honest read: quote the baseline whose methodology matches how you measure. If you count unique replies over all sends, Belkins’ 0.45% is your comparison point and any video figure you set against it must be counted the same way. Comparing a per-campaign video number to a per-send email number produces a lift that exists entirely in the arithmetic.

Reply and response lift: what the defensible numbers say

The most-quoted video reply statistics on this SERP are also the least traceable, which is a problem, because the traceable ones are perfectly good. They are just less dramatic.

The cleanest figure we found is SalesLoft data reported by Vidyard (2024): teams adding video to their sequences saw a 16% bump in open rates and a 26% increase in replies. That is a lift on an existing sequence, measured within one platform, and it is the number we would use if we had to pick one. Alongside it, Vidyard (2023) found 63% of sales professionals reporting that video messaging increased their response rates, and that buyers are 50% more responsive to sales engagement platform messages when at least one step includes video. In Vidyard’s earlier survey of 600+ sales leaders (2021), over 70% said custom-recorded video outperformed text-based emails.

Note what those are: self-reported perception from practitioners, not controlled measurement. They tell you what reps believe is happening, which is worth knowing and is not the same thing as an effect size. Vidyard’s own 2025 revenue leader data supplies the useful counterweight — only 40% said they earn more responses from prospects, and only 27% cited shorter deal cycles or faster funnel progression. If video were transformative for everyone, those numbers would not sit at 40% and 27%.

The most actionable finding in this group is not about video at all. Sopro (2026) reports that advanced personalisation lifts response rates to 18%, against 9% for generic emails. That is depth of research roughly doubling response, on the same channel. Gong’s 2025 analysis points the same way from the other side: pitching reduces reply rates by as much as 57%. Relevance is the variable doing the work. Video is a delivery format that makes relevance visible faster — it is not a substitute for having something relevant to say.

Nothing on this page is an Outvid customer result. We have not published outcome data on our own product, and we are not going to imply one exists by placing a category benchmark next to our name. Treat every figure here as directional context for planning, not as a forecast of what your campaign will do.

Length versus completion: the case for staying under a minute

This is the one area where the data is unambiguous and immediately usable. In Vidyard’s 2025 benchmark report, videos under one minute hold viewers to the end at the highest rate of any length band, with 65% of viewers staying engaged all the way through. Videos over twenty minutes retain 20%.

That is a completion curve, not a persuasion curve, and it is worth being precise about what it licenses you to conclude. It says a shorter prospecting video is far more likely to be watched to the end. It does not say a 45-second video books more meetings than a 90-second one — no source we could verify measures that. What it does mean in practice is that if your video runs past a minute, the part of your message you care most about should not be at the end of it.

Combine that with the recall data below and the operational rule writes itself: put the specific, personal, provably-researched line in the first ten seconds, where completion is not yet a factor at all.

• Under 60 seconds: 65% of viewers stay engaged to the end (Vidyard, 2025). This is the prospecting sweet spot.
• Over 20 minutes: 20% stay to the end (Vidyard, 2025) — relevant to demos and recorded walkthroughs, not to a first touch.
• No verified source we found measures reply rate by video duration, so treat any specific “optimal length” claim, including ours, as inference from completion data.

Buyer preference and recall: the demand-side case

The strongest argument for video outreach is not a reply-rate percentage. It is what happens to your message after it is read.

In an EEG-based brain study of B2B professionals conducted by B2B DecisionLabs (2021), nearly 60% of participants remembered zero emails within 48 hours of viewing them. That is the actual baseline problem: not that cold email fails to get opened, but that it evaporates. The same study found 46% of people remembered details from a text email against 59% from a video within an email, and that people were 13% more likely to remember details from video emails overall.

Preference data points the same direction. Wyzowl (2026) found 63% of people would most like to watch a short video to learn about a product or service, against 12% who would prefer a text article. Sopro’s State of Prospecting (2026) supplies the other half of the picture: 57% of sales and marketing decision-makers believe most sales outreach feels impersonal. And Gong, reported by Vidyard (2024), found closed deals involve using webcams 41% more often than lost deals — correlation, clearly, since better-qualified deals also get more camera time, but a consistent one.

Read together, those numbers describe a specific mechanism. Video does not make a bad message good. It makes a good message survive the 48 hours after it is read.

LinkedIn benchmarks

LinkedIn is where the honest answer is thinnest, and we would rather say so than fill the gap with a vendor number. HeyReach’s dataset of 96,051 campaigns (2026) gives a solid channel baseline: a typical 21% connection acceptance rate and a typical 22% reply rate. In the same analysis, personalised video sequences on LinkedIn produced roughly 50% video view rates and an average of 23 booked meetings per 100 messages.

Two caveats you should apply before planning against those figures. First, they come from a vendor analysing campaigns run on its own platform, which is a self-selecting population of teams already investing in LinkedIn outreach. Second, and more importantly, we found no controlled head-to-head study comparing video and text reply rates on LinkedIn with matched lists. The comparison everyone wants does not appear to exist in public. Anyone presenting a precise LinkedIn video-versus-text multiplier is extrapolating, and so would we be.

The 22% baseline reply rate is still the more useful of the two numbers, because it tells you whether your current LinkedIn campaign is underperforming the channel before you conclude that adding video is the fix.

What these numbers do not tell you

Four limits worth holding onto, because they apply to every figure on this page.

• Almost all of it is vendor-published. Vidyard, Wyzowl, Wistia, Sopro, Saleshandy, Belkins and HeyReach all sell into this category. Their data can still be good — published methodology and disclosed sample sizes are what make it checkable — but none of it is independent academic measurement.
• Survey-reported outcomes are perception, not effect size. “63% of reps say video increased responses” is a fact about what reps say. It is not a measured lift.
• Platform data is self-selecting. Teams that adopt a video tool and keep paying for it are not a random sample of sales teams.
• Nothing here is segmented by your industry, deal size or list quality, all of which plausibly matter more than the format of the message. Use these as planning ranges, then measure your own.

How to cite this page

If a figure here is useful to you, please cite the original publisher rather than us — the source and live link are in the table above for exactly that reason. Where you want to reference the compilation, the sourcing standard or the excluded-numbers list, the following works:

Outvid, “Video Prospecting Statistics (2026),” outvid.ai/video-outreach/statistics. Accessed [date].

Videos under one minute hold 65% of viewers to the end (Vidyard, Video in Business Benchmark Report, 2025).

The “216% higher response rate” figure could not be traced to a reachable primary source — see Outvid, “Numbers we could not verify,” outvid.ai/video-outreach/statistics.

Last verified: 27 July 2026. Every source URL was fetched and the figure confirmed on that date. Changelog: 27 July 2026 — first publication; 33 figures verified, 7 excluded; Vidyard benchmark link corrected from the 404ing /blog/video-benchmarks/ path to the live report URL. If you find a figure here that has since moved or changed, tell us and we will fix or pull it.

Every figure, with its source

Each statistic below was checked against the publisher’s live page. Where a number could not be traced to a reachable primary source, we excluded it and listed it separately.

Numbers we could not verify

These figures circulate widely in articles about video outreach. We could not trace any of them to a reachable primary source with a stated methodology, so we do not cite them as fact — and we’d suggest treating them carefully wherever you see them quoted.