How to Build a Video Flywheel That Sells (10 Steps)

Key Takeaways

• A video flywheel that sells starts with your sales call transcripts. Buyers ask real questions there, and each question becomes one video.
• Keep the videos largely unedited. Cut the very beginning, the very end, and the mistakes. A real, unpolished answer builds more trust than a polished one.
• Packaging is what gets the video found: a custom thumbnail, an SRT transcript, chapter markers, and end-screen suggestions. YouTube gets pushed into AI Overviews more every month, and if you are not in those results you are invisible to the buyer.
• Check the analytics after about 48 hours. If people watch past 30 seconds and past one minute, they are getting value. The whole game is stretching that out.
• The end state is a library that answers every buyer question. I call it content bingo. When buyers get about 80% of the way there, they are willing to take a sales call.

What is a video flywheel that sells?

It is a repeatable loop that turns the questions your buyers already ask into video answers, then uses the performance data to decide what to make next. The goal is inbound leads. Instead of chasing buyers, you answer the questions they are asking while they self-educate, and they come to you.

The reason it works is simple. Video is very helpful for buyers as they research on their own, and YouTube gets pushed into AI Overviews at an increasing rate. If you do not show up in those results, you are out of the buyer's evaluation set and you will not sell. Basically, you are invisible.

There are 10 steps. Here is each one.

This fits inside a bigger picture. If you want the full view first, here's how to show up in AI search results with B2B video.

Step 1: pull buyer questions from your sales transcripts

Start with the transcripts from your buyers. These are sales calls where they are asking real questions, in their own words. That is the raw material for every video that follows.

Step 2: build the script and graphics

Take those questions and build some graphics on a platform like Miro, which is what I use. If you want a scripted opening to hook them, write that out. Then create the graphics you will show as you walk the buyer through the answer.

Step 3: record the video

Record your answer. The graphics you built in Step 2 are what you point at while you talk.

Step 4: edit lightly

Take out maybe just the beginning and the very end, and any mistakes. Otherwise keep it largely unedited. It seems more real and authentic that way, and the whole point is to build trust.

"Keep it largely unedited so it seems more real and authentic, because we're trying to build trust." Dane Frederiksen

Step 5: package the video for YouTube and AI search

This is where the video gets found. Packaging means:

• Make the thumbnail.
• Optimize the YouTube page with the transcript in SRT file format, which is better for AI engines crawling the video.
• Add chapter markers.
• Put suggestions on the end screen.

Building the flywheel is the easy part. The hard part is knowing where you're already winning and where you're invisible. That's exactly what our free AI Visibility Snapshot shows you, before you spend a dollar.

Step 6: post the video

Post it and put all of the packaging together on the YouTube page.

Step 7: read the analytics after 48 hours

After about 48 hours you start to see data about how the video performed. The signal I look for is watch time. If people are watching for over 30 seconds, and over one minute, they are getting some value.

Then the game is to stretch that out. Make people click on your thumbnails, click on your titles and descriptions, and watch longer, by answering the questions better. That is the whole game.

Step 8: prioritize what to make next

Use what the analytics taught you. Are there unanswered questions you have not done a good job of answering yet? Those are the next videos.

Ideally, after answering every question in individual videos, you end up with a library that answers all of them and gets buyers to about 80%, where they are willing to take a sales call. That is the end state. I call it content bingo: fill up all the questions.

Step 9: iterate

Keep doing this over and over. Every time you get analysis, you get data, and things begin to improve. You get better at this.

Step 10: repeat

Ideally the buyers stop asking the questions you have answered, and those questions come up less often on sales calls. You might get new questions as the market changes. Get those from the transcripts and repeat the process.

You will be able to measure sales leads, and you will know it is working, because you will see people watching the videos and getting their questions answered.

Most teams don't get this right on the first try. Here's how we approached it for Twitch. Curious what it'd look like for you? Let's talk it through.

Frequently Asked Questions

Where do the video topics come from?

From your sales call transcripts. Buyers ask real questions on those calls, and each question becomes one video answer.

How much should you edit a B2B video?

Very little. Cut the beginning, the end, and any mistakes, and leave the rest. A largely unedited video seems more real and builds more trust.

What does packaging a YouTube video include?

A thumbnail, the transcript uploaded in SRT format so AI engines can crawl it, chapter markers, and end-screen suggestions.

How do you know if a video is working?

Check the analytics after about 48 hours. If viewers watch past 30 seconds and past one minute, they are getting value. Then work to stretch that watch time out by answering the question better.

How many videos do you need before buyers take a sales call?

Enough to answer all of their questions. When buyers get about 80% of their questions answered, they are typically willing to take a sales call.

Full Transcript

This is how to build a video creation flywheel system that sells. There's 10 steps and I'm going to walk you through each one now. The goal is to drive inbound leads by answering your buyer's questions with video content. The reason this works is because video is very helpful for buyers as they self-educate and YouTube gets pushed into AI Overviews at an increasing rate. If you don't show up in those results, you're out of the buyer's evaluation set and you won't sell. Basically, you're invisible.

Step one, we're going to start with our transcripts from our buyers. These are sales calls where they're asking real questions. Then we're going to take those and we're going to build some graphics on a platform like Miro, which is what I use. And then you're going to, anything you want to script in the beginning to hook them, you can write that out and you're going to create some graphics of what to show as you walk them through and answer their questions.

Step three, we record the video. Step four, edit the video, taking out maybe just the beginning and the very end. Any mistakes, but keep it largely unedited so it seems more real and authentic because we're trying to build trust.

So then packaging, we're going to make thumbnails. We're going to optimize the YouTube page with transcripts in an SRT file format, which is better for AI engines crawling. We're going to put chapter markers. We're going to do things like put suggestions at the end screen at the end. We're going to post the video and put that all together.

We're going to then do some analytics after 48 hours or so. We're going to start to see some data about how it performed. And if we look and we see, okay, people are watching for over 30 seconds, over one minute, that means they're getting some value. And then the game is to just stretch that out, make people click on your thumbnails and click on your descriptions and titles and make them watch longer by answering the questions better. That's the whole game.

Then we're going to prioritize based on what we learned in our analytics. Are there unanswered questions that we haven't done a good job of answering just yet? Then we can go and make that content for. Ideally, after answering all their questions in individual videos, at some point we'll have a library that answers all the questions and gets them to 80% where they're willing to take a sales call. So this is the end state we're approaching is to fill up all those questions. I call it content bingo.

So then we just iterate and we keep doing this over and over again as we repeat. And every time we get some analysis, we get some data, things begin to improve. So we get better at this. And ideally the buyers will stop asking the questions that were answered and they won't come up as often. You might get new questions as the market changes, but we're going to get those questions from the transcripts and just repeat the process. And then you will be able to measure sales leads and you'll know that it's working because you're going to see people are watching this stuff, getting their questions answered. I know this is complicated. If you want some help, just reach out. This is what I do and how I can help you.