The problem
High traffic, zero pathway
The blog was doing its job — ranking, informing, earning trust — and then dead-ending. The only way to act on what you just read was a footer CTA that asked you to leave the article and start over somewhere else.
Research
Two readers, two very different habits
Talking to readers and reviewing behavior data, the same two types kept showing up. Carol reads carefully and wants to know if she qualifies before she commits to anything. Robert skims, and if the next step isn't in front of him, he's gone.
Both wanted the same thing: a quick answer without leaving the page they came for.
"If it's too complicated, I just move on. I'm not going to dig around forever."
The journey
Intent spikes mid-article, not at the end
I mapped the path from landing on an article to talking to an agent. Attention peaks around the section that answers the reader's question — well before the footer. That's the moment the form had to be there.
Mapping the edge cases mattered just as much: no ZIP entered, not eligible, no plans in the area. Each one needed a helpful next step instead of a dead end.
Testing
Four placements, one honest test
Instead of arguing placement in a meeting, I designed four real variants — inline CTA, floating CTA, sticky lead form, and sticky form plus inline CTA — with identical copy, so the test measured position and nothing else.
The hybrid won on both counts: 92% form visibility and a 7.8% lead rate, against 23% and 1.2% for the inline CTA we started with.
- Inline CTABarely discovered — 23% visibility, 1.2% lead rate.
- Floating CTAMore visible, but felt bolted on to the article.
- Sticky lead formA clear improvement at 83% visibility.
- Sticky + inlineBest overall: 92% visibility, 7.8% lead rate.
The design
One screen, three questions
The winning pattern asks for ZIP, age range, and whether you already have a Medicare card — enough to qualify someone, short enough to finish on a phone in a few seconds.
Nothing resets the article. Submitting shows a plain confirmation and hands off to an agent with the answers already filled in, so no one repeats themselves.
Impact
A pattern the whole content library could use
Lead form visibility tripled across the blog, and qualified leads rose 15% now that readers had a direct path from an article to enrollment.
I documented the placement as a reusable component so content and growth could roll it out across articles without waiting on design.
View the Figma prototype ↗What I'd take forward
Position was worth more than any wording change we tried. Next time I'd test placement first with a coded prototype on a couple of live articles, before anyone writes copy or builds components. I'd also tie the form to article topic — someone reading about Part D needs a different first question than someone reading about enrollment dates — and measure which article types actually carry intent.