AI Fortune Telling: 100 Million Won in Sales, The Reason Behind a Common Item

Ellen developed and sold a service where AI provides fortune readings and users can chat with AI based on those readings. This service achieved sales of over 100 million won in its first month after launch.

However, sales halved in just two months, though it continues to generate steady revenue, reaching a cumulative 1.3 billion won to date.

Ellen realized the importance of item selection through a past business failure of 5 billion won. She diagnosed that 9 out of 10 early-stage entrepreneurs often fail trying to pioneer new items not yet in the market.

Focusing on markets where people are already spending money and purchasing power is confirmed is a way to reduce risk.

According to Ellen's experience, dietary supplements for diabetes were a market item with clear demand, as existing competitors were already generating hundreds of millions of won in sales. In contrast, melatonin products were approached with the vague assumption that the entire nation was a target, but actual advertising response was low.

The melatonin business struggled with declining sales despite over six months of A/B testing and improvement efforts, a result of relying solely on speculation for demand.

Conversely, a cosmetic product launched after confirming clear demand successfully stabilized monthly sales, illustrating the importance of validating market demand.

The fortune-telling app market, with annual sales in the tens of billions of won, can be validated by competitor ad spending and search volume. Ellen planned the AI fortune-telling service after confirming the market size, with some fortune-telling apps generating tens of billions, even over 100 billion won, in annual sales.

She analyzed competitors' ad spending and Naver search volume to verify market potential. Thanks to this thorough market analysis, she achieved 100 million won in sales in the first month.

It is crucial to select items with assured buyers; entering existing markets increases the probability of success more than novel ideas.

Competitor analysis can utilize Facebook Ad Library to check competitor service ad portfolios, identifying what ads they are running.

Additionally, use Naver Advertiser Center's keyword tool to look up search volume for specific brands. For example, if mobile and PC search volumes combined exceed 30,000 per month, it can be estimated that the competitor is spending at least several hundreds of millions of won on advertising.

Ellen advised that if a competitor service is generating sales with a certain quality of advertising, one can assume that similar efforts could secure a certain level of sales.

Even after selecting an item, it is common to waste over a year trying to create a perfect product. People tend to delay product launch by obsessing over adding features or refining screen designs.

However, even a perfectly crafted product can have unpredictable customer reactions. Many cases miss the launch window by constantly deliberating whether customers will like the product or if it will sell.

If you focus solely on development without confirming if customers are willing to pay, you could waste over a year. This is not just a waste of development time but can lead to greater risks.

For example, if sales are only 30,000 won for every 1 million won in advertising, it's wise to shut down the business immediately. If only 1-2 out of 10 product features resonate with customers, a strategy of discarding the other features and focusing on the popular ones is necessary.

The more effort put into a product over a year, the harder it is to give up, even if there's no market response. Excessive attachment to a product leads to the sunk cost fallacy, making one cling to the business.

This reluctance to admit failure and attempt to revive it can lead to another year wasted. Consequently, losing a total of two years on a low-probability business is the biggest risk.

The advice to launch quickly does not mean neglecting quality. It means not to obsess over perfection when you don't know if the product will sell.

A product can always be modified and improved after customers are acquired. Whether customers will purchase a product cannot be resolved by pondering in front of a computer; it must be confirmed by launching the product in the market and selling it directly.

Marketing is not an issue to consider after a product is complete; it must be clearly defined from the product planning stage. A systematic approach is needed to design the entire process from customer awareness to usage and payment.

The customer journey—from awareness, free trial experience, satisfaction and desire, to the call-to-action button for payment—must be concretely mapped out. Simply adding features haphazardly and then mimicking copy won't generate sales.

Advertising is simply paying money to bring customers to a website; the product itself must have the competitiveness to drive conversions. If a high-conversion revenue structure is in place, where 5 out of 100 visitors make a purchase, advertising spending can be increased to expand sales.

In Ellen's case, after confirming 1 million won in sales with 200,000-300,000 won in advertising, she increased ad spend to achieve 100 million won in sales. This demonstrates that advertising investment in a validated product is effective.

A specific scenario should be established: 20 sign-ups from 100 visitors, 15 free trials, 5 continuing for a week, and 2 final payments. For example, pre-set a scenario where out of 100 visitors, 20 sign up, 15 use the free trial, 5 continue the service for a week, and 2 make a final payment.

Defining conversion rate targets for each step in advance allows for sales forecasting and management. This goes beyond simple design improvements, leading to contemplation of messages and storytelling to increase conversion rates at specific stages.

Such structural design enhances sales predictability and provides clear direction on where improvements are needed.

For increased ad spending to directly translate into sales, a selling structure must be in place beforehand. Without pre-building such a structure, significant marketing difficulties will arise after launch.

One should not solely focus on feature development before validating hypotheses with data. Creating a selling structure by designing the customer's purchase journey from before launch is essential for achieving sales.

He consistently achieved over 100 million won in monthly sales with 12 items annually. According to his experience, creating situations where customers themselves feel a need is far more powerful than simply persuading them.

One must be able to imagine the thoughts and feelings customers experience at every moment, from seeing an ad and becoming aware of a product to clicking the payment button. This connects to defining UI/UX not as mere design but as the entire customer experience.

The core principle of successful business lies in deeply understanding and designing the customer journey.

To achieve sales performance, one must capture and utilize various psychological states of customers, such as boredom or doubt.

In an era where it's easier to create products using AI, success once achieved is not guaranteed to last. Ellen shared that even items she made successful saw advertising efficiency halve in just two months, indicating significant market variables.

Therefore, one must not dwell on a single successful item but continuously discover new items and adapt to changes. The success of the AI fortune-telling service and its subsequent decline in sales illustrate this market environment.

Among Google Android apps, only 3% of the total are subject to fees, meaning they generate revenue. Furthermore, according to a survey by the international payment processor Stripe, only 3% of 8,800 startups achieve monthly sales of 10,000 dollars.

Ellen advised that to become one of the top 3% that generate good revenue, one must select validated items that people are already spending money on and prioritize the customer experience.

Ellen built a system where AI fully manages advertising budgets ranging from 50 million to 100 million won per month. AI performs the entire process of ad operations, including competitor analysis, copywriting, ad creative production, and bid adjustment.

Ellen's employees all utilize Claude Code to systemize their work. This maximizes ad operational efficiency and minimizes human input, showcasing active use of AI in practical work.