I Sighed Hearing an Internal AI Case Study: Reducing Financial Product CPA from 200k to 20k in One Week

@MOLTS_2030
JAPANESE2 months ago · Jun 02, 2026
128K
171
22
1
234

TL;DR

An insightful look into a case study where CPA dropped from 200,000 to 20,000 yen. It highlights that while AI accelerates execution, the core value lies in human strategic insight and hypothesis generation.

====

On Thursday, June 11, 2026, from 15:00-16:00, we will be holding a webinar on the theme of "How to build an AI-based marketing project that creates 'double the results in half the time' than before"! Please join us if you can!

https://moltsinc.co.jp/webinar/ai-marketing-2x-half-time/

======

There is a saying that has often been repeated within our company.

"You should automate routine tasks. But for non-routine tasks like marketing, while you should use AI aggressively, you must also train your own fundamental skills."

To be honest, I only half-understood what this meant for a long time.

I've been using AI aggressively for the past six months

Claude Code, Codex. Catching up on the latest tools. Automating routine tasks, generating reports, and systematizing content production.

Tasks that used to take several hours now finish in 30 minutes.

The number of creatives and the quality of output have clearly improved. A sense of confidence that I was "mastering AI" was gradually building up.

An internal professional reduced the CPA of a financial service to one-tenth in one week

I heard a story from a senior colleague about a client support project.

It was a financial matching service. The CPA (Cost Per Acquisition) exceeded 200,000 yen. Looking at the competitors, every company was making similar appeals. Same conversion points, same words, speaking to the same target audience.

My senior colleague judged that "there is no way to win if we keep fighting like this."

What they did next was observe the target. They didn't do interviews. They looked at digital data and past lost-deal data. Instead of asking "why didn't they inquire?", they dug into "why did they stop even though they were interested once?"

Then, they found a certain common sensation that the target audience—high-income earners—had inevitably experienced in their lives.

Everyone knows it. Everyone has felt it. But there wasn't a single company touching on that in their marketing.

They verbalized that sensation and made it the core of the appeal.

They used AI to verify if there were similar cases in other industries and confirm the probability before moving. LPs, creatives, and conversion content were all tested speedily using AI.

As a result, the CPA dropped sharply in one week.

By further evolving the appeal, the CVR (Conversion Rate) rose even more, and the CPA dropped from 200,000 yen to 20,000 yen. Ultimately, they got so many orders that the sales team couldn't keep up.

I cannot replicate this

Hearing the story made me sigh.

It wasn't "I can do this if I use AI." It was "This is impossible for me."

Reading the market. Deconstructing the competitor's structure. Digging up the sensations that the target has accumulated in life. Identifying the psychological barriers to conversion. Bundling all that into a single appeal axis, element, or conversion content.

AI only starts to play a role after that. It's used to increase the accuracy of hypotheses, to build systems faster, and to create creatives.

The AI my senior used, I can use too. I can create LPs, systems, and appeals by myself.

But I cannot produce the same performance from scratch at the same speed as my senior.

AI replaces routine tasks. But it's you who improves performance

What my boss said finally clicked.

"You should automate routine tasks with AI."

This is correct. AI certainly handles repetitive tasks faster and more accurately than humans. What I've been doing for the past six months definitely worked in this area.

"But for marketing, train your own fundamental skills."

AI is a tool that expands a person's performance. But what is being expanded is "one's own performance."

If the base is low, that low quality is just multiplied.

The reason I can't replicate the results my senior achieved in a week, even if I use AI, is right there. The precision of my thinking hasn't reached that level yet.

What is truly necessary in the AI era is not just the skill to master AI. It's creating "your own thinking" that is worthy of being handed over to AI in the first place.

I was made to realize this from the bottom of my heart. I'm still lacking in that. So I felt strongly that I must train.

執筆—THE MOLTS煨陆途

One-click save

Use YouMind for AI deep reading of viral articles

Save the source, ask focused questions, summarize the argument, and turn a viral article into reusable notes in one AI workspace.

Explore YouMind
For creators

Turn your Markdown into a clean 𝕏 article

When you publish your own long-form writing, images, tables, and code blocks make 𝕏 formatting painful. YouMind turns a full Markdown draft into a clean, ready-to-post 𝕏 article.

Try Markdown to 𝕏

More patterns to decode

Recent viral articles

Explore more viral articles