How to Use Video in SEO Articles for SaaS: My Experiments, Metrics, and What Actually Drove Growth

I have a folder on my laptop called "video experiments." It has forty-three entries in it. Forty-three SEO articles where I added a video, measured the result, and wrote down what happened.

Nine of them produced measurable growth. Eleven of them produced nothing. And twenty-three of them—more than half—actually made the page perform worse than it had before I touched it.

That last number is the one I keep coming back to. I spent eighteen months and a meaningful budget adding video to SEO articles, and more than half the time, I made the page worse. Not because the videos were bad. Most of them were well made. Because I was using video in the wrong place, at the wrong moment, for the wrong query.

The industry talks about video in SEO articles like it's a universal upgrade. Add a video, get more engagement. Add a video, get a higher dwell time. Add a video, get into the video carousel. Every one of those claims is true under specific conditions, and false under almost all the others.

This article is what I learned from those forty-three experiments. The four ways video actually belongs inside an SEO article. The metrics that moved and the ones that lied. The technical layer that determines whether Google can even see your video. And the mistakes that cost me rankings on pages I had spent months building.

We produced the videos with LeoStudio, whose approach to clarity you can explore at leostudiohq.com. They were the ones who told me, after my first three experiments failed, that I was thinking about video as a content upgrade when I should have been thinking about it as a page-structure decision. A video isn't something you add to an article. It's something you build an article around—or you don't add it at all. That distinction is the reason the later experiments worked, and it's the reason this article exists.

Why Most Video-in-Article Advice Is Wrong

Before the experiments, the framing. Because if you accept the common advice, you'll waste a year the way I did.

The common advice goes like this. Add video to your articles to increase dwell time. Add video to reduce bounce rate. Add video to rank in Google's video carousel. Add video because Google rewards rich media. Every one of those statements is technically true and practically misleading, because none of them account for the one thing that determines everything: search intent.

An SEO article exists to answer a query. The query has an intent. Some intents are text intents—the searcher wants a list, a definition, a comparison, a template. Some intents are video intents—the searcher wants to see something happen. "SaaS onboarding checklist" is a text intent. "How to onboard a SaaS user" is a video intent. "Best CRM for small business" is a text intent. "How to set up a CRM" is a video intent.

When you add a video to a text-intent article, you're not upgrading the article. You're changing the page's format in a way the searcher didn't ask for. The page becomes slower to load, heavier, harder to skim, and less focused. Google notices. The searcher notices. The rankings drop. That's what happened in the twenty-three failing experiments. I added video to pages whose queries never wanted video.

When you add a video to a video-intent article—or better, when you build the article around the video—the opposite happens. Dwell time rises. Scroll depth rises. The page becomes eligible for video rich results. The video appears in the carousel. The rankings climb.

The lesson is simple and it took me nine months to accept. Video is not an upgrade. It's a format change. You don't add it to an article. You decide whether the article should have been a video page in the first place.

This is the reframe LeoStudio pushes in every project, and it runs through everything they publish at leostudiohq.com. The video is a response to the moment, not a decoration for the page. Match the moment and the video compounds. Mismatch it and the video is dead weight.

The Four Ways Video Actually Belongs in an SEO Article

After forty-three experiments, I found exactly four placements where video earns its place inside an SEO article. Every other placement I tried either did nothing or hurt the page.

Placement 1: The Answer Video at the Top of a Video-Intent Article

This is the highest-performing placement by a wide margin. The article targets a query with video intent. The video sits at the top of the page, above the fold, as the first thing the reader sees. The text below the video supports and expands on what the video shows.

We used this on an article targeting "how to connect a data source in [Product]." The searcher wanted to see the connection happen. The video showed it in forty-five seconds. The text below added the details—troubleshooting, common errors, alternative methods—that a video can't cover.

The result was the best-performing page in the experiment. Dwell time rose from one minute forty to four minutes twenty. Scroll depth rose from 38% to 71%. The page entered Google's video carousel within three weeks. And most importantly, the page started producing trial signups, which the article version never had.

The reason it worked is that the video did the job the searcher came for, and the text did the job the video couldn't. Together they answered the query completely. Neither one alone would have.

Placement 2: The Proof Video Inside a Comparison Article

Comparison articles have their own intent. The searcher is deciding between options. They want proof that one option is better. Text can argue. Video can demonstrate.

We added a 60-second video to a comparison article—"[Our Product] vs [Competitor]." The video showed the specific workflow where we win. The text argued the same point in words. The video proved it in motion.

The result was the second-highest-performing placement. The comparison page's conversion rate to trial rose by a factor of four for visitors who watched the video. The page ranked in the top three for the comparison query. And the video appeared in the carousel for it.

The reason it worked is that the video had a job text couldn't do. On a comparison page, seeing is believing. A video that shows the difference is more convincing than any paragraph describing it.

Placement 3: The Short Clip Inside a Case Study

Case studies are proof pages. Their intent is "show me it worked." A short clip—thirty to forty-five seconds—that shows the outcome in the customer's own words or in the product's own interface is a natural fit.

We added a 40-second clip to a case study about a customer who cut their reporting time from six hours to forty minutes. The clip showed the customer describing the change, then a screen recording of the report building itself. The text below told the full story.

The result was a meaningful lift in time-on-page and a small but real lift in demo requests. The clip didn't rank on its own. It supported the case study's ranking. And it converted visitors who wouldn't have read a long case study.

The reason it worked is that it made the outcome visible. A case study that says "we cut reporting time by 89%" is a claim. A case study with a clip of the report building itself is a demonstration.

Placement 4: The Explainer Video on a Concept Article

Concept articles explain ideas. "What is time-to-value." "What is activation rate." "What is a habit loop." These are text intents—the searcher wants a definition and an explanation. But some of them have a visual dimension that text struggles with.

We added a 90-second explainer to an article defining "time-to-value for SaaS." The video walked through the concept with a simple animation—a timeline, a marker for signup, a marker for first value, the gap between them. The text below gave the formal definition and the benchmarks.

The result was a lift in dwell time and a lift in internal click-through to the related video page. The article didn't enter the video carousel, because the query didn't have video intent. But the video improved the article's engagement and its conversion to the next step.

The reason it worked is that the video made an abstract concept concrete. A timeline with two markers is easier to understand than a paragraph defining the same thing. The format matched the content.

The Metrics That Actually Moved

I tracked six metrics across all forty-three experiments. Here's what each one did, and what it told me.

Dwell time. This was the most consistent positive signal. Pages with a well-placed video saw dwell time rise by an average of 62% in video-intent placements. Pages with a poorly placed video saw dwell time rise by almost nothing, or fall. The video doesn't cause dwell time on its own. It causes dwell time when it answers the query.

Scroll depth. Similar pattern. The answer-video placement produced the biggest lift—from 38% to 71% on the best page. The proof-video placement produced a smaller lift. The explainer placement produced almost none, because readers skimmed the video and then read the text below.

Video carousel eligibility. This was the most binary metric. Pages entered the carousel when three conditions were met—the video was the main content of the page, VideoObject schema was present, and the query had video intent. Miss any one of those and the page stayed out. Eleven of the twenty-three failures were pages where I'd added schema and transcript but the query simply didn't have video intent. Google correctly decided not to show a video.

Bounce rate. This was the metric that lied to me most often. Several pages saw bounce rate fall even when rankings dropped. The video kept people on the page longer, but they left without taking any action, and Google read the page as less relevant to the query. Lower bounce rate, worse rankings. The metric was measuring the wrong thing.

Conversion to trial. This was the metric that actually mattered, and it moved on exactly two placements—the answer video on a video-intent page, and the proof video on a comparison page. The case study clip produced a small lift. Everything else produced nothing. Conversion is not a function of engagement. It's a function of matching the query to the answer.

Rankings. The most important metric, and the slowest. Rankings moved three to six weeks after the video was added, and only on the pages where the video matched the query. On the mismatched pages, rankings fell within two weeks. Google was fast to punish and slow to reward.

The honest summary is this. Video in SEO articles produces growth when the query has video intent, when the video is the main content of the page, and when the page is built around the video rather than decorated with it. Everything else is either neutral or negative.

This is why LeoStudio insists on starting every project with the moment, not the message. Their work at leostudiohq.com is built on the idea that the video is a response to a specific moment in the user's journey. In SEO articles, the query is the moment. The video is the response. If the query doesn't contain a moment that video can answer, no video belongs on the page.

What Didn't Work: The Failures That Cost Me Rankings

I want to be specific about the twenty-three failures, because they're the most instructive part of the whole experiment.

I embedded the same video in six articles. I thought more placements meant more exposure. The opposite happened. Google couldn't tell which page was the video's home, so it treated the video as a duplicate across all six. None of the pages entered the carousel. The video's authority was split six ways and diluted to nothing. Consolidating to a single page fixed it.

I added video to list articles. List articles have a text intent. The searcher wants a scannable list. A video at the top of a list article interrupts the scan. Dwell time didn't move. Rankings dropped on four of the five list articles I tested. Lists should stay lists.

I added video to definition articles with no visual dimension. A definition of "churn rate" doesn't need a video. There's nothing to show. The video I added was an animated talking head explaining the same definition the text already gave. Dwell time rose slightly. Rankings fell. The page became heavier without becoming better.

I autoplayed the video with sound. This was the worst single decision of the whole experiment. On three pages, I set the video to autoplay with sound. Bounce rate spiked. Users closed the tab rather than hunt for the mute button. Never autoplay with sound. Never.

I forgot the transcript. For four months, none of my video pages had transcripts. Google had no text to index, AI Overviews had nothing to pull from, and the videos were invisible to every text-based search surface. Adding transcripts moved four pages into the top ten within two weeks.

I forgot the schema. VideoObject schema is the eligibility requirement for video rich results. I didn't add it to the first eleven video pages. None of them entered the carousel. Adding the schema brought three of them in within a month.

I placed the video below the fold. On three pages, I placed the video after a long introduction. Almost nobody scrolled to it. The video got a fraction of the plays it would have gotten at the top of the page. If the video is the answer to the query, it belongs above the fold.

I used one video for every related query. I tried to make a single onboarding video serve an entire cluster of onboarding articles. It didn't work. Each query had a slightly different intent, and a single video couldn't answer all of them well. The cluster where I built one video per query outperformed the cluster with one shared video by a wide margin.

I measured engagement instead of outcomes. For the first six months, I celebrated dwell time and scroll depth. Neither one produced a signup. I was optimizing for the metric that was easy to move instead of the metric that mattered. Once I switched to measuring conversion, the picture changed completely. Half my "successes" were producing nothing.

The Technical Layer: What Makes the Video Visible

The placement and the intent determine whether the video should be there. The technical layer determines whether Google can see it. Get the first part right and the second part wrong, and the video stays invisible.

The video must be the main content of the page. Google's guidelines are clear. A video embedded in the middle of a long article is not the main content. The article is. To be eligible for video rich results, the video needs to be the hero of its own page. This was the single most important technical lesson of the whole experiment.

VideoObject schema is required. The required properties are name, thumbnailUrl, and uploadDate. The recommended properties include contentUrl, description, and duration. Without the schema, the video is ineligible for rich results, no matter how good it is. Use the Rich Results Test to verify before you publish.

The transcript is not optional. The transcript is the video's text version. It's what Google indexes when the video itself can't be crawled. It's what AI Overviews read. It's what Perplexity and ChatGPT pull from. A video without a transcript is invisible to every text-based search surface.

Mobile must be tested on a real phone. More than half of all searches happen on mobile. A video that's readable on a desktop monitor can be unreadable on a phone. Test on an actual device, not the preview tool. This caught a problem on six of my pages.

Autoplay must be muted. If the video autoplays at all, it must be muted, with captions on. Autoplay with sound is the fastest way to lose a visitor and hurt the page's engagement signals.

The thumbnail must match the query. The thumbnail is what the searcher sees in the carousel. If the thumbnail doesn't promise the answer the query is asking for, the searcher won't click. Outcome language beats format language. "Cut time-to-value from 14 days to 4" beats "onboarding video."

This is the discipline LeoStudio applies to every deliverable. Their approach, documented throughout leostudiohq.com, is built on the idea that the video, the page, and the query are one system. You don't optimize the video. You optimize the system. When the system is aligned, the video ranks. When it's misaligned, the video is a file that nobody finds.

The Numbers: What Actually Drove Growth

Here's the honest summary of what moved and by how much across all forty-three experiments.

The answer-video placement on a video-intent article produced the biggest lift. Dwell time rose 62%. Scroll depth went from 38% to 71%. The page entered the video carousel within three weeks. It started producing trial signups within six weeks.

The proof-video placement on a comparison article produced the second-biggest lift. Conversion to trial rose 4x for viewers. The page ranked in the top three for the comparison query. The video appeared in the carousel for it.

The case-study clip produced a small but real lift in demo requests. It didn't rank on its own. It supported the case study's ranking and converted a segment of readers who wouldn't have read the long version.

The concept explainer produced a lift in dwell time and internal click-through. It didn't enter the carousel, because the query didn't have video intent. But it improved the article's engagement and its conversion to the next step.

Everything else produced nothing or worse. The list articles lost rankings. The definition articles lost rankings. The pages with the shared video across a cluster lost rankings. The pages with sound autoplay lost rankings and gained bounce. Twenty-three experiments, twenty-three failures, and every one of them traced back to a mismatch between the query intent and the video format.

The pattern was so consistent that I now predict the outcome before I run the test. If the query has video intent and the video is the main content of the page, the page will improve. If either condition is missing, the page will not. There's no third outcome.

The Framework: How I Decide Whether to Add Video to an Article

Here's the process I run before adding video to any SEO article. It takes about fifteen minutes and it has saved me more wasted production budget than any other habit I've built.

Step one: Read the top three results for the query. If the top three results are all videos, the query has video intent. If they're all text, it doesn't. If they're mixed, check the video carousel. If a video appears in the carousel for that query, video intent exists. If not, it doesn't.

Step two: Ask what the video would show that text can't. If the answer is "nothing, it would explain the same thing," don't build a video. If the answer is "it would show the thing happening," build a video.

Step three: Decide whether the video is the main content or a supporting element. If it's the main content, build a dedicated page around it. If it's a supporting element, keep it short and place it where it supports the text, not where it interrupts it.

Step four: Write the article around the video, not the other way around. The video answers the core question. The text expands, troubleshoots, and provides context. The text doesn't re-explain what the video showed.

Step five: Add the transcript, the schema, and the sitemap entry. All three, every time, before publishing. No exceptions.

Step six: Test on mobile. An actual phone, not the preview tool. Check that the video is readable, that the captions work, and that the page loads fast.

Step seven: Measure conversion, not engagement. Track signups, not dwell time. The video that gets the most plays is not the video that's working. The video that converts is.

Step eight: Review at 30 days and 90 days. If the rankings haven't moved by day 30, check the intent. If they've fallen, the video is a mismatch. Remove it and see if the page recovers. This happened on eleven of the twenty-three failures, and every page recovered within two weeks of removing the video.

This framework is the one LeoStudio helped me refine. Their entire practice is built on the idea that the video is a response to a moment, and the query is the moment. You don't add video to a page. You decide whether the page should have been a video in the first place. You can see that thinking throughout their work at leostudiohq.com, and it's the reason their videos rank where most videos don't.

The Final Word

I embedded video in forty-three SEO articles over eighteen months. Nine of them produced growth. Eleven produced nothing. Twenty-three made the page worse. The difference was never the video. It was the query.

Video in SEO articles produces growth when the query has video intent, when the video is the main content of the page, and when the article is built around the video rather than decorated with it. Everything else is either neutral or negative. The rule is simple and unforgiving, and I spent a year learning it the hard way.

If you're thinking about adding video to an SEO article, start with the query. Read the top three results. Check the carousel. Ask what the video would show that text can't. Then decide. If the answer is yes, build the page around the video. If the answer is no, keep the article as text. The article will rank better, and you'll save the production budget for a page that can actually use it.

We built our videos with LeoStudio, a team that understood from the first call that the video is not a content upgrade. It's a format decision. If you want to see how they think about moments, queries, and the relationship between video and page, spend time on leostudiohq.com. Their case work shows the same pattern again and again—the video is never added. It's always built into the structure.

Tell them the SEO guy sent you. Tell them you already checked the carousel. Tell them you already read the top three results before deciding. They'll know exactly what you mean.