I don't have a client video campaign to show you. I've run plenty of paid media for manufacturers and technical companies, and none of it has been video. So there's no case study here. What I can stand behind without one is the measurement side of B2B video marketing, and specifically the gap between what a "view" means and what it's worth.
That gap is the whole problem with video as a marketing channel. A view on one platform means three seconds of attention. On another it means the video started playing at all, whether or not anyone was still looking. None of that maps cleanly to a lead or a closed deal, which is the only thing I actually care about when I'm asked to justify a channel's budget.
What a "view" actually counts
Every platform sets its own bar for a view, and the bars aren't close to each other. If you're comparing view counts across platforms, or treating a view as a unit of engagement, you're comparing numbers that were never designed to mean the same thing.
YouTube (via Google Ads, skippable in-stream). Google's own documentation for TrueView ads defines a counted view as a case where "A viewer watches 30 seconds or until the end of the video, whichever comes first." A view is also logged if the viewer clicks the call-to-action, a card, or the logo, regardless of watch time. So a 45-second video can register a view from someone who watched two-thirds of it and quit, or from someone who clicked a button in the first second and never watched anything.
Meta (Facebook and Instagram). Meta's Ads Manager tracks several separate metrics, not one "view." Per Meta's own help documentation, a "3-Second video play" is "the number of times your video plays for at least 3 seconds, or for nearly its total length if it's shorter than 3 seconds." Meta also tracks a "2-Second continuous video play," and separately a "ThruPlay," defined as "the number of times your video was played to completion, or for at least 15 seconds." Three different thresholds, three different numbers, and the platform lets you report whichever one looks best.
LinkedIn. LinkedIn's Campaign Manager documentation defines a video view as "two or more continuous seconds of playback while the video is at least 50% on screen," which LinkedIn attributes to the Media Rating Council's cross-industry standard. Two seconds, half the frame visible. That's the bar.
TikTok. TikTok Ads Manager's own metric definitions distinguish a basic "video view," which counts "the number of times your video started to play," from a "6-second video view," defined as the video playing "for at least 6 seconds in an impression session, played in full if it is less than 6 seconds, or received at least 1 engagement within the first 6 seconds." The plain "video view" number, the one most dashboards surface first, requires nothing but a play button being pressed.
Line those up and the range runs from "the play button was tapped" to "thirty seconds of attention, or a click." A view on TikTok and a view on YouTube are not the same unit of anything.
Why that makes views and watch time diagnostics, not KPIs
A diagnostic tells you something is working or broken. A KPI tells you whether the business got what it paid for. View count and watch time are diagnostics. If a video's average watch time falls off a cliff at the ten-second mark, that's useful: it tells you the hook doesn't hold, or the video is in the wrong placement, or the audience is wrong. That's worth knowing and worth fixing.
What it doesn't tell you is whether the video produced a lead, a sales conversation, or revenue. It can't, structurally, because none of the view definitions above reference anything happening outside the platform. A view is a platform-side attention event. A KPI, in my book, is a CRM-side business event: a form fill, a booked call, a closed deal. Those live in different systems, and a video platform has no way to report on the second kind unless you build the bridge yourself.
This is the same distinction I've made about SEO reporting: traffic and rankings are diagnostics, not KPIs, and the fix is the same in both cases. You don't throw out the diagnostic. You stop treating it as the answer to "did this work."
It's also worth naming why video is the worst offender on this particular failure. Views are the easiest number in all of marketing to feel good about. They're large, they update in real time, and every platform puts them front and center in the dashboard specifically because they're the metric most likely to make a campaign look successful regardless of what it actually did. A page of view counts and a chart trending up is a genuinely good-looking report. It's also completely disconnected from whether the phone rang.
What to instrument if you're going to spend money on video
If a B2B company is putting real budget behind video, here's what actually closes the loop back to the CRM, in rough order of how hard each one is to set up.
UTM-tagged links in and around the video, every time. Description links, end cards, pinned comments: tag every one with UTMs specific to that video and that platform placement. This is the minimum bar for knowing which video sent a visitor to the site at all. It's also the step most B2B video efforts skip, because it takes five minutes and produces no visible payoff until someone actually converts.
A dedicated landing page for each video to point to. If every video sends traffic to the homepage, you've thrown away the ability to attribute anything downstream. A dedicated landing page, or at minimum a dedicated URL parameter, is what lets a CRM record eventually get traced back to the video that sourced it.
Form fills and calls tagged with the referring campaign. Whatever CRM is catching the lead needs to record the UTM or landing page source at the point of submission, not just "website" as a generic bucket. If the CRM can't answer "how many of this quarter's opportunities came in through a video-sourced link," the instrumentation isn't done yet, no matter how good the view counts look.
Offline conversion import, if the sales cycle is long. B2B deals close weeks or months after the first touch. Platforms like Meta and LinkedIn support importing closed-deal data back into the ad account so the platform's own optimization can eventually learn from real outcomes instead of proxy engagement. This is more setup than most teams do for a channel they're still testing, which is a legitimate reason to defer it, but it's the step that actually closes the loop.
A clear owner for checking whether any of this happened. Someone has to look at the CRM monthly and ask whether video-sourced leads are turning into anything, the same way you'd check that for a search campaign. Without that review, the dashboard full of view counts is the only report anyone reads, by default, because it's the only one that exists.
None of this is exotic. It's the same measurement discipline paid search and SEO get held to, applied to a channel that usually gets a pass because the numbers on the surface look impressive.
The honest version
Video is a legitimate format. Product demos genuinely shorten sales cycles for complicated products, and a well-made explainer can do work a landing page can't. None of that is in question here.
What's in question is whether the reporting on a video program tells you anything true. If the only numbers in the recap are views, watch time, and engagement rate, you have a diagnostic report dressed up as a results report, and the two read identically until someone asks what it produced in the CRM. Set up the tracking before you spend the budget, so you aren't trying to build it retroactively once someone asks where the leads went.
If you want to see how I hold paid media to this same standard elsewhere, the results page is built on CRM-attributed outcomes, with the platform-reported numbers left out of it. And if video is one channel inside a broader paid media plan you're weighing, the conversation belongs under paid media management, where the thing being decided is the media plan and production is downstream of it.
