A few years ago, traffic to my language business began to shift. StoryLearning had always been heavy on SEO. Then Google changed how search results were served, and the SEO traffic dropped.
So I worked hard to replace it with YouTube traffic and YouTube leads. That part worked. Eventually YouTube produced more leads than SEO ever had at its peak. Revenue was still down.
Now, why is this?
A Funnel Built for SEO Stopped Matching YouTube Traffic
The funnels I was running had been built for SEO visitors: people who arrived from Google search, read a blog post, opted in to a text lead magnet, and made their way through emails to a text sales page. They had never been rebuilt for viewers arriving from YouTube.
That was the whole problem in StoryLearning’s case.
Not all leads are equal. A lead from one channel can be drastically different in profitability from a lead from another. In that case, more leads and lower revenue wasn’t a traffic problem. It was a funnel mismatch.
Now, I’m not suggesting you build five separate funnels, one for every traffic source. That would multiply complexity without fixing anything. But if your funnel was built in 2015 for readers and your traffic is now video, something has to change.
SEO Leads Are Readers; YouTube Leads Are Viewers
An SEO lead starts with a question typed into Google. Then a blog post. Then a text lead magnet. Then a sequence of emails. Then a long-form sales page. Every step is text, because the person chose text from the start. They were actively seeking an answer. They’re readers by behaviour.
Now picture a YouTube lead. They often didn’t search in the same way, and they’re often not consuming text. They’re watching video, usually on a phone. Ask that person to work through a PDF lead magnet. Some will. Many won’t. Send them long, story-driven emails. Send them a long-form sales page. You tend to lose people at each step.
Some YouTube viewers will happily read every word. It’s just not a given. When the old funnel was built, it treated text consumption as the default. For a video viewer, it often isn’t the default. That single assumption explains a lot of the leakage I saw.
The old funnel assumed the reader. StoryLearning’s new traffic was viewers. That mismatch, on its own, is my explanation for why leads were up and revenue was down. The only variable that changed in that shift was the medium.
Change the Lead Magnet to Match the Medium
Email is still king.
Lead capture doesn’t go away because your traffic is video. You still need a lead magnet. But a text-based lead magnet often isn’t the natural next step for someone who just watched a video.
So, offer a special video instead. Or a free webinar. Or a masterclass. Or something interactive: a widget, a calculator, a tool. Anything other than a long PDF. If they came from video, the next logical step is usually video. It’s their natural habitat. Not a text document they have to read.
That one change targets the first leak I would chase. Before they ever see an email, the medium is more likely to match the traffic.
Send Email a Video Viewer Will Actually Consume
A reader will often work through long-form email. A YouTube or social lead usually won’t. An email that doesn’t get consumed isn’t doing its job. Open rates drift down. The relationship never quite builds. By the time you make an offer, the person hasn’t been primed.
Because of that, the format has to change. Keep emailing.
Shorter emails, then. More visuals. Links out to your best videos. Make the message more direct: problem, solution, pain point. Not long story-based text. This audience doesn’t tend to spend long in an inbox. If you want the email to land, you usually have to hit hard in fewer words.
I’m all for story-based stuff all day long. But, again, if you’re asking people to consume longer stories in text format over email, it’s not a given that that is going to work. Are they going to sit down and read long-form emails? Those same long-form emails that have just worked for so long. Possibly not. Story-based email suits readers. It doesn’t suit every lead. When a lead arrives from social video, a shorter, more direct message is more likely to be consumed. The job of email is to build enough trust that the sales page doesn’t have to do all the work. If the email never gets consumed, that job never happens.
Put the Sales Page on Video Before You Rewrite It
A long-form text sales page still works for a reader. A video viewer won’t reliably sit through it. Long-form pages have their place. They’re not where I would start for someone arriving from YouTube.
So what would you do instead? Well, first of all, you start with a VSL, wouldn’t you? If someone’s coming from a video viewing mindset, give them a video to consume. A video sales letter can be the same copy, delivered on video with a teleprompter. The argument doesn’t have to change. The medium does a lot of the work. Long-form pages traditionally worked because readers would skim for the information they needed. A video viewer often won’t sit and read through it. If people still aren’t consuming the page, retargeting ads can put it back in front of them more often.
Keep the long-form page for readers. Lead with video for viewers once you’ve reached this point. But this is a later step. Test the email funnel first, and leave the sales page unchanged until that test has validated the new pain point. Then worry about rewriting the copy.
Use AI to Find the Gaps in Your Current Funnel
Rebuilding your whole funnel is a big job, right? You’ve got lead magnets, emails, sales pages. You’re going to interpret this as a lot of very hard work, and rightly so. Who are we kidding? This is no small job.
But you don’t have to guess where to start.
Well, connect an LLM to your business data. YouTube comments from the last two years. Support tickets. Customer surveys. Reddit mentions, which I personally avoid like the plague but which are useful data. Email inbox questions. Ask it to compile one report on pain points, motivators, purchase intent, and the exact language people use. And then run a gap analysis: compare that report with what your current funnel actually talks about.
I’ve been doing this recently with entrepreneurs I mentor, and the results are often quite shocking. The funnel focuses on one or two pain points from ten years ago. The data coming back from your current audience points somewhere else entirely. When you see that data laid out in front of you, it just makes no sense to be marketing in the way you’ve been marketing. AI is good at consuming a lot of noisy data and finding the pattern. It can do in minutes what would take a team months. But with anything AI spits out, I don’t want to just take it at face value. Anyway, the next action is to write one new email funnel from that data and test it before you change anything else.
Change Email First, Then Validate Before You Touch the Rest
Once the gap is clear, you can ask AI to write a new email funnel based on it.
Is it perfect? No. Is it going to be full of gaps? Yes. It won’t be as good as copy you wrote by hand. What you lose in polish, you gain in speed.
The important part is alignment, not eloquence.
If your audience has one pain point and you’re writing about another, just closing that gap should give you an immediate change, even with average writing.
Very tempting with AI to start changing everything.
Don’t. Change one thing.
Change the email funnel only. Leave the sales page alone. Leave everything else alone. Then track behaviour. In my business, I use SegMetrics to see what each customer clicks on and engages with after the change. That tells you whether the new pain point is real.
Run that test for three to four weeks. If the result is positive, the pain point is validated. Only then do you rewrite the sales page to match. One step at a time, in that order.
If you’ve been running the same funnel for five years or more, this diagnosis is overdue. The traffic changed. Your funnel probably didn’t.
