The more products we had, the more people we needed to hire

The first meeting was different. Instead of sitting in a conference room listening to requirements, we visited their office directly and watched how the CEO and staff updated lowest prices in their daily routine. Starting at 9 AM, logging into each platform, alternating between Excel sheets and platform screens, entering prices one by one. "Why do you do it this way?" "What's the criteria for this calculation?" With each question, hidden business rules revealed themselves one by one.

The first technical barrier we faced was login. Each platform required login, and most importantly, the CAPTCHAs that appeared every time were a problem. How could we automate this process that requires manual input? We combined various bypass techniques and automation solutions to build a stable login system. Now the system handles login and authentication processes automatically.

The next step was collecting lowest price data. To check the lowest price for a specific product, we needed the 'unique category value' of each product. Each platform had different ways of identifying products, and the HTML structure containing lowest price information varied. We analyzed browser network packets to reverse-engineer the necessary API endpoints, identifying platform-specific request formats and authentication headers. Based on this, we built a system to collect real-time lowest price data and convert it into structured format.

Price updating was another challenge. We had to accurately analyze the logic for modifying product prices on seller pages. It wasn't just calling APIs - we had to follow each platform's seller page UI flow exactly to send price modification requests. We implemented handling for different authentication methods and request formats across platforms.

But the real difficulty started here. The lowest price calculation logic was not simple. Discount rates, customs, internal promotions, shipping costs, platform fees... all these variables were complexly intertwined. Not only that, there were detailed business rules like "This product should never be sold below 15,000 won minimum", "If current lowest is 20,000 won, set it to 19,500 won", "But if it's within 500 won of minimum, just use minimum price." We stood together with the CEO in front of a whiteboard, checking cases one by one and drawing the complex business logic into a flowchart. "What about this case?" "And that case?" Through repeated questions and answers, we completed the logic. We implemented all these complex rules requiring dozens of branches and exception handling into algorithms. It was the moment when hidden business know-how was systematized into code.

We built it step by step together. We made it web-based so it could be accessed and modified from anywhere. The CEO needed to be able to access it from home or during business trips via phone. We configured the server to automatically set lowest prices, and built a responsively developed back office where everything could be controlled. We wanted to hear "This is easy to use!" rather than "This looks pretty!"

We paid special attention to making the 'most important lowest price algorithm' usable like a formula. The CEO is not a developer. But they're proficient with Excel. So we made it possible to set the lowest price algorithm just like entering Excel formulas. Like "Competitor lowest * 0.95 - 500". I remember what the CEO said when first seeing this screen: "This feels familiar like Excel. I think I can modify this myself."

A system running 24/7 means something could go wrong. So we built a real-time monitoring system. The theme can be switched between dark/light mode, and you can check in real-time when the lowest price setting process ran, which products were updated, and if there were any errors. Logs from 3 AM runs, weekend runs - all transparently visible.

Finally, we secured flexibility in product management. You can see and manage 200+ products at a glance, and set individual rules for each product like 'turn auto-update on/off', 'what's the minimum price', 'sell on this platform only'. Adding new products is just a few clicks. Before, adding a new product meant adding a row to Excel and telling staff "Please check this too". Now, register once in the system and it's automatically managed.