Before the dusty shelves,
25 years on

1-1. Consumer Psychology Learned from Mart Shelves

Shelves never lie.

Which product makes a consumer stop. Reach out. Put it back down. The truth of marketing lives inside those three seconds.

Consumers don't buy products. They buy context.

There is a reason the produce section is at the entrance of every mart. It makes you experience freshness first. It activates the brain's trust circuit. After that, the path leads to processed foods, frozen goods, and household items. The feeling of "this mart can be trusted" is established first. Then the purchase happens.

The single most common mistake SME manufacturers make: "Our product is better — why isn't it selling?" Quality is not sufficient for purchase. No matter how good the product, if the context for consumer recognition, trust, and choice is not built, it dies on the shelf.

A mart is not a space for selling products. It is a space for designing context. Whoever understands that design controls distribution.

1-2. What Manufacturers Miss, What Distributors Know

Manufacturers and distributors see the same product in different languages.

The manufacturer says: "How good are these ingredients." The distributor thinks: "Which shelf does this product need to be on to sell." The language that touches the consumer directly is always the distributor's.

This is where Lee Dong-yeol of Korea Tech coined his concept of "Distribution R&D" — reverse-engineering products from real distribution-floor data. "Developers fixate on the product itself. It's hard for them to focus on consumers." The result of using consumer insights from the distribution front to back-engineer R&D: KAHI revenue of ₩250 billion in 2021.

What manufacturers miss comes down to one thing:

Consumers don't choose a product. They confirm what they've already decided.

Over 70% of purchase decisions happen before the consumer ever stands in front of the shelf. Brand awareness, word of mouth, SNS exposure. Designing that process is marketing. The party that holds the most of those designs is the distributor.

Distributors have data. They know which product sells in which weather, at what time of day, to which customer segment. This information asymmetry — where one side holds far more information — is what shapes the structure of B2B distribution.

1-3. The Structural Inefficiency of Domestic B2B Distribution

Let's start with numbers.

₩2,000 trillion (approx. $1.5 trillion USD). Annual domestic B2B transaction volume. Of this, the digitization rate is under 8%. The remaining 92% is conducted via phone calls, business cards, and trade fairs.

3 to 6 months. The average time it takes a small manufacturer to find a new trade partner. A sales rep's personal network is the company's sales power.

Why does this inefficiency persist? Three structural reasons:

First, information asymmetry. Manufacturers don't know market average prices. Distributors don't know a manufacturer's actual production CAPA (capacity). Negotiations happen in mutual ignorance, creating a structurally weaker party.

Second, trust costs. How do you trust a manufacturer you've never dealt with? There is no standardized process from sample request to contract. Every transaction rebuilds trust from zero. These costs compound into trillions of won in missed opportunity.

Third, channel fragmentation. Home shopping, online, offline, and overseas subsidiaries are run separately. An integrated sourcing strategy is impossible.

This is not a technology problem. It is a structural one. Structural problems can only be solved by platforms.

1-4. Why Distribution Must Change Now

Three things are changing simultaneously, right now.

Democratization of AI. Demand forecasting, matching algorithms, and risk scoring were once the exclusive domain of large corporations. Today, small platforms have access to the same tools. The entry barrier has fallen.

Accelerating consumption trends. SNS virality has compressed product life cycles from three years to three months. A 6-month sourcing process is already behind the market.

The arrival of Millennial and Gen Z buyers. The generation making B2B purchasing decisions has changed. They find partners through platform searches, not trade fair business cards. Platforms — not suppliers — are now the standard of trust.

The reason distribution must change isn't moral imperative. It's survival.

The answer to that question I asked 25 years ago has finally arrived — in the era of data and AI.

2-1. From Import Distribution to Own Brand — Korea Tech's Turning Point

Lee Dong-yeol started in 1998 at Sewoon Shopping Arcade — Seoul's legendary multi-story electronics and hardware complex. He chased rural traveling markets that rotated on 3-day and 5-day cycles. He set up stalls, demonstrated glass-cleaning gadgets with flair, and secured exclusive distribution rights. He eventually broke into home shopping.

After founding Korea Tech, he hit three consecutive exclusive import successes: ReFa, the beauty roller endorsed by actress Lee Young-ae; Sixpad, the EMS fitness device; and Pao, the facial muscle trainer.

But he felt a fundamental deficiency. "No matter how well I sell, in the end it's someone else's product." When a licensing contract ended, he had no guarantee of business continuity.

That deficiency drove him to build a manufacturing-distribution integrated company. He didn't just build a factory — he brought the intuition honed on distribution floors into the very first stage of product development. Not "make something that will sell well," but "find what the market truly wants, then realize it through technology." This was the birth of Distribution R&D — a methodology for reverse-engineering products from distribution-floor data.

2-2. The Secret Behind KAHI's ₩250 Billion — Reverse-Engineering Marketing

2021. A single pink stick detonated the Korean beauty market. KAHI Multi-Balm.

The result: ₩250 billion (approx. $190M USD) in revenue within one year of launch. Cumulative sales of 10 million units. Not just a hit product — a phenomenon.

The starting point was consumer data. Lee Dong-yeol identified one clear insight: "Women want wrinkle care, but complicated skincare routines feel like a burden." He focused on solving that contradiction.

No mess on the hands. Swipe it on anywhere, anytime. Existing stick-type cosmetics were either too hard or applied poorly. He solved the problem technically with Jeju fermented oil as the core ingredient.

But what came next was decisive. Before launch, he designed the marketing context first. He deployed drama PPL (product placement), home shopping demonstrations, and influencer joint purchases simultaneously. He planted the perception: "KAHI handles wrinkle care in one step."

It didn't sell because the product was good. A selling structure was built first, and then the product was placed inside it.

2-3. The Three Stages of Distribution R&D — Discovery, Re-Creation, Channel Fit

Lee Dong-yeol's Distribution R&D model has three stages.

Stage 1: Discovery (Sourcing). On the distribution floor, look for items where you sense "this could work." The criterion is not technological superiority — it is whether the product can solve a consumer's unmet need.

Stage 2: Re-Creation. The discovered source technology is brought to the in-house research lab and redesigned. You don't say "it's good" — you acquire clinical data. That data is then translated into marketing language. This is the evidence-building stage.

Stage 3: Channel Fit. Channels are assigned to match the product's characteristics. Products requiring demonstration go to home shopping. Products requiring explanation go to detailed online pages. Products requiring experience go to physical stores. KAHI deployed all three channels simultaneously — a true omni-channel strategy.

2-4. The Roller-Coaster P&L — The Lesson of Hit-Dependency Risk

Where there is light, there is shadow.

Korea Tech's financial statements are called a "roller coaster." In numbers:

• 2018: ₩100 billion
• 2019: ₩30 billion (steep decline)
• 2021: ₩250 billion (KAHI explosion)

When a hit product lands, margins and revenue explode. When the follow-up hit doesn't come, fixed costs become poison. Single-brand, single-product dependency always carries inherent instability.

This is where the necessity of a platform model emerges. If Distribution R&D is fishing for a jackpot, a platform is the business of creating fishing grounds where fish naturally gather.

Watching Lee Dong-yeol's success and his risk simultaneously — this was the decisive reason I conceived MARTMART. Not one-time jackpots, but a system that connects thousands of transactions happening every single day.

3-1. Compressing 3–6 Months of Partner Discovery into 72 Hours

Traditional B2B sourcing moves at this speed: finding a new manufacturer and reaching a contract takes an average of 3 to 6 months. Wait for trade fair season. Exchange business cards. Receive samples. Conduct factory inspections. Negotiate terms. All analog, all slow.

MARTMART declared 72-hour matching.

A distributor enters their requirements — category, unit price, MOQ (minimum order quantity), certifications. AI searches the manufacturer database for the optimal partner. A matching list arrives within 72 hours.

This isn't simple efficiency improvement. It means gaining the ability to respond to trends in real time. When tanghuru (a wildly viral candied fruit-skewer snack) was sweeping Korea, launching a product 3 months later was already too late. Finding a manufacturer in 2 days and getting on shelves in 2 weeks — that's what generates revenue.

The 72-hour match operates in two stages. At 48 hours, escalation kicks in automatically: the supplier receives a notification that "if you don't respond within 24 hours, you will be automatically replaced." At 72 hours exceeded, an alternative supplier is connected automatically. Unresponsive suppliers are passed over by the system. Buyers don't wait.

3-2. The Network Effect Built by 3,710 Companies

The greatest challenge of any platform business is the cold start problem — no one comes to an empty marketplace with no sellers and no buyers.

MARTMART confronted this head-on.

Built over 13 years. Since founding Saram-gwa-gori (People & Links) in 2013, we directly organized and hosted 29 distribution exchange meetings. The result: 2,163 manufacturers + 1,547 distributors = 3,710 companies in total DB.

In April 2026, 5,000 Southeast Asia (SEA) seller databases are added. The total becomes a global network of 8,710+ companies.

This data is a moat — a competitive barrier that rivals cannot easily replicate. Naver and Coupang have vast B2C data. But they lack B2B-depth data like factory production CAPA and OEM availability.

The network effect has just begun. More manufacturers attract more distributors. More distributors attract more manufacturers seeking to list. As data accumulates, matching accuracy improves. Switching costs rise. The platform's winning formula.

3-3. PSS v4.0 — Reading Products in Numbers

MARTMART developed PSS (Product Sellability Score) v4.0 — a proprietary scoring system that quantifies "how well can this product sell."

PSS v4.0 is composed of 5 components:

1. Marketability — demand intensity in the relevant category
2. Profitability — margin potential relative to supply price
3. Competition — competitive intensity within the same category (lower is better)
4. Data Quality — completeness of product information
5. Trend — rate of change in 7-day view count

Grades are assigned in 4 tiers:

Grade	Threshold	Meaning
S	80+ points	Priority sourcing target
A	65+ points	Recommended for matching
B	50+ points	Conditional review
C	Below 50	Needs re-evaluation

Currently, 2,638 products have been scored.

Why does this matter? Distributors no longer need to rely on gut feel when selecting products. A-grade and above are the products the data says "have a high probability of selling." Manufacturers can see their product's objective competitiveness in numbers — for the very first time.

PSS is MARTMART's core intellectual property with 2 patents filed.

3-4. How the AI Matching Algorithm Works — Demand Reverse Matching

MARTMART is not a simple keyword search engine. The core is Demand Reverse Matching.

Most platforms work like this: manufacturers post "come buy our goods." MARTMART works in reverse. When a distributor posts "we need this," the AI finds potential manufacturers and proposes them.

This is implemented through the expose/seek tag system. Products are tagged as either expose (available to supply) or seek (demand exists). When a buyer attaches a seek tag, the AI matches manufacturers with expose tags in reverse direction.

The AI matching formula:

Match Score = PSS × 0.35 + CTS × 0.30 + BAS × 0.15 + Response Rate × 0.20

• PSS: Product Sellability Score
• CTS (Company Trust Score): composite of response rate, completion rate, ratings, and account age
• BAS (Buyer Activity Score): level of buyer platform activity
• Response Rate: speed and rate of response to recent trade requests

Supplier trust is quantified through 25 badge types. Badges are automatically granted and revoked based on business registration verification, transaction history, response speed, and certification documents. Trust is read in numbers.

A transaction risk index — combining credit ratings, past transaction history, and claims data — also provides advance warnings. The "fear of not getting paid" that haunts B2B transactions is lifted by the system.

3-5. Before and After — A Real-World Sourcing Case Study

Case of Retailer A, running a food mart in Gyeonggi Province. (Illustrative example)

Before

• To find a new kimchi supplier, went through 3 layers of wholesalers
• Accumulated middleman margins — no price competitiveness
• Even with quality complaints, could not communicate directly with the manufacturer
• 20 hours per month spent on sourcing alone

After

• Entered conditions on MARTMART: "Domestic Chinese cabbage, 10 kg, 500 boxes/month, under ₩30,000/unit"
• Within 72 hours, matched with 3 kimchi factories in Haenam, South Jeolla Province
• Direct trade, no middlemen — unit price reduced 15%
• Freshness improved via factory-direct delivery
• Sourcing time cut to 30 minutes

Retailer A reinvested the saved time and money into store marketing. Sales up 30%.

4-1. B2B Is 5× Slower and 10× Bigger than B2C

B2C runs on emotion. If it looks good, you buy it. Impulse works.

B2B is different. Logic comes first. Then numbers. Finally, accountability.

The average approval chain runs 5 levels deep. The account manager loves it, but the team leader blocks it. The team leader approves it, but finance blocks it. That is why B2B is 5× slower.

In exchange, transaction scale is 10× to 100× larger. Once you break through, that single deal can become one year's revenue.

The core of B2B marketing is not exposure. It is building trust. Trust means the conviction: "I won't get hurt doing business with you."

"Our product is beautiful" is B2C language. "We can reduce your risk" is B2B language.

Buyers don't purchase good products. They purchase products that won't fail them.

4-2. The MD's Proposal Checklist — What They Actually Look At

An MD (Merchandiser — the buyer responsible for selecting products for a retailer) receives dozens of proposals a day. Reading time: 30 seconds.

Proposals that pass in 30 seconds share one thing: they have numbers. The 5 items an MD actually checks:

1. MOQ (Minimum Order Quantity): How much to start. Write a number.

2. Lead Time: How many days until delivery after order placement. Write in weeks.

3. Unit Price Structure: What is the unit price at 100, 500, and 1,000 units? Show as a table.

4. Certification Documents: KC, FDA, Halal, Vegan. Write only what you actually hold.

5. USP (Unique Selling Point): One line that differentiates you from competitors. Just one.

"Premium high-end natural sensibility" is not a USP. From the MD's perspective, it is empty space.

Remove emotional adjectives. Fill that space with numbers.

4-3. Faster Sourcing Beats Trends — The 3-Month Preemption Strategy

First place creates the trend. Second place chases it fast. Third place holds the inventory.

If you're not first, use the second-place strategy.

On average, it takes 3 months for a product trending on overseas TikTok to reach Korean mart shelves. That time gap is the opportunity. Source products trending overseas now to align with domestic launch timing 3 months later. This is the 3-month preemption strategy.

MARTMART's predictive matching reads these signals for you. "Source this product now" is not just a recommendation. It is investment intelligence that pulls future revenue forward to today.

Distribution Is People

We talked about AI. We talked about data. We talked about algorithms.

All of it is true. But there is only one conclusion.

Distribution is people.

A handshake comes before a contract. Trust comes before price. The conviction that this person won't get hurt — that's what creates a transaction.

Technology reduces costs. It reduces search time. It reduces mistakes. But technology cannot create trust.

Right now, MARTMART runs 24 cron jobs (automated tasks) 24 hours a day. 4 databases operate separately by role. 25 badge types are automatically attached and removed. All of this has one purpose: helping people trust each other faster.

Technology accelerates the speed of trust. It does not replace trust itself.

Where Does the Next Distribution Revolution Come From?

Three directions.

1. Finance + Distribution Convergence
The biggest risk in distribution is money — you deliver the goods, but payment doesn't come. Escrow pre-settlement solves this. You receive 80% of the transaction value upfront. The remainder arrives after delivery confirmation. The seller's cash flow changes entirely.

2. K-Food and K-Beauty Expansion into Southeast Asia
We are adding 5,000 SEA (Southeast Asia) company databases. Connecting overseas buyers who want K-products with domestic sellers. This is cross-border matching.

3. Global MAMA
A service that directly connects overseas distributors and manufacturers. The platform lowers language and trust barriers. Domestic sellers find overseas buyers; overseas sellers find domestic buyers.

The next distribution revolution lies outside national borders.

How to Start with MARTMART

Change is frightening. I understand. I have heard that sentence thousands of times over 25 years.

But not changing is more dangerous. Experience taught me that, too.

The way to start is simple.

Step 1. Register for free. Zero cost.

Step 2. Onboarding begins immediately after registration. Right now, the 72-hour matching system is already running.

Step 3. Your partner arrives in 72 hours — connected from a database of 3,710 companies matched to your conditions.

No complicated procedures. No sales calls. No contract pressure.

"Start for free. Judge after 72 hours."

I started working in Yongsan at the age of 16. Writing paper receipts and tax invoices, eating instant noodle soup over the counter — soup splashing onto paper invoices I still had to hand over...

The tools have changed. Excel became a database. Fax became an API. Business cards became algorithms.

But the essence hasn't changed.

Helping good products find their rightful owners.

What I did in the corridors of Seonil Shopping Complex, I now do through a platform. Only the method has changed.

Right now, MARTMART has 3,710 companies. The matching system runs 24 hours a day. The 72-hour guarantee holds today.

There is no human team yet. Instead, there is an AI agent team. PSS v4.0 scores the sellability of 2,638 products. The AI matching engine connects manufacturers and distributors within 72 hours. CTS reads supplier trust on a scale of 0 to 1,000. Seek/Expose reverse matching finds buyer demand first, then sends it to suppliers. 25 cron jobs attach and remove badges, send reminders, and close transactions. These agents run the matching system 24 hours a day, without rest.

This is possible because the era of agentic AI has arrived. A system that is alone, yet not alone. That is MARTMART.

Finding trade partners — the search is over.
It took 25 years to earn the right to say that sentence.