Achieve Personal Loans Landing Page · Achieve. com, 2025
Redesigning the entry point for a product people already wanted.
Users were arriving ready to explore a loan and leaving without starting. One CTA change, one line of microcopy, and a stripped-back IA moved the needle across the entire funnel.
Timeline: May to August 2025
Role: Lead Product Designer, 0 to 1
Team: 1 PMs, 1 UX Writer, 3 Engineers
Tools: Figma, Figma Dev Mode, Figma Make
Impact: +$638k incremental marketing margin/year. Lending first brand initiative: +$2M annualized total
01
THE PROBLEM
The APL page had strong intent traffic.
Users were arriving with a clear goal: to explore a personal loan, check a rate, and understand their options. But downstream conversion lagged. Something in the experience was creating hesitation at the exact moment users should have felt most confident moving forward.
The page was speaking to someone in distress. The users arriving were optimizing.
Baseline 90,804 sessions/mo · 38.6% Direct-to-MP · 14.1% Lead/Session
02
Before redesigning anything, I needed to understand what was actually stopping users from clicking. Three signals stood out.
RESEARCH
Process anxiety, not product rejection. Users weren't leaving because they didn't want the product. They were leaving because "Get Started" implied a commitment they weren't ready to make. Would it affect their credit? What was the next step? The CTA was answering the wrong question.
Trust signals arrived too late. Trustpilot ratings and third-party badges were on the page, but below the fold. Users were deciding whether to proceed before they'd seen any reason to trust us.
The homepage test told me what not to do. A parallel A/B test showed that more content created paralysis for financially anxious users. I applied that learning from day one. APL would be stripped to essentials.
03
This project was an early test of what AI-integrated practice looks like in production.
HOW AI SHAPED THE PROCESS
I used AI to synthesize competitive patterns and surface trust-signal gaps faster than manual analysis could, compressing days of research into hours. Figma Make let me explore multiple hero architectures and IA structures in parallel rather than sequentially, which created space for deeper synthesis earlier in the process.
For microcopy, AI helped me generate and pressure-test multiple directions quickly, critical in a compliance-sensitive flow where every word carries legal weight. The line "It's quick, free, and won't affect your credit score" went through eight iterations before it landed.
04
Four changes. One principle: earn trust before asking for commitment.
FOUR DECISIONS
"Get Started" to "See my rate." Apply implies commitment. See my rate implies exploration. For a user with financial anxiety, that distinction is the difference between clicking and leaving. This isn't a copy decision; it's a trust architecture decision.
One line that drove more conversions than any visual decision. "It's quick, free, and won't affect your credit score." - directly below the CTA. Every word addresses a specific fear surfaced in research: time, cost, and credit impact.
Trust signals moved into the hero. Trustpilot 4.8 | 11,000+ reviews and LendingTree Top 3 badge repositioned above the CTA. Credibility at the decision moment, not after it.
Radical simplification. Three value props. A three-step process section. No quiz, no solutions sidebar. More content creates paralysis for a financially anxious user. That was the first principle.
05
The hardest conversation on this project wasn't about design; it was about compliance.
STAKEHOLDER ALIGNMENT
In a regulated financial environment, every trust signal, rate reference, and eligibility statement requires legal review. The instinct is to bring compliance in at the end, after the designs are final. I brought them in at the start.
That meant sharing early wireframes and copy directions before they were refined, which felt uncomfortable but saved three weeks of revision cycles later. More importantly, it shaped the design. The specific language around credit impact, the sequencing of disclosure statements, and the framing of third-party badges, all of it was better because legal had a voice in the room early rather than a veto at the end.
Cross-functional alignment isn't a process step. It's a design decision.
06
The variation didn't just attract more users. It attracted better-fit users.
RESULTS
Higher intent coming in meant better close rates downstream. That's what intent alignment looks like in production. Contributing to $2M in annualized margin for the Lending First initiative.
+15.5%
Lead Rate
+29.2%
Enrollment Rate
+25.4%
Close Rate
+75.5%
Margin / Session
+$638k
Incremental margin / yr
07
WHAT I LEARNED
What the Achieve Personal Loans page taught me
The gains were driven primarily by mobile. Desktop performance was softer than the control. In hindsight, I'd run a desktop-specific layout iteration earlier; the two surfaces have meaningfully different anxiety profiles for a user considering a financial product.
The bigger lesson: stated intent in concept testing and actual behavior in production are not the same thing. Design to the behavior, not the stated preference.