Own Shopping Mall vs Smart Store: Where Should You Spend 10,000 Won Daily on Advertising?

Soulpapamarketing’s Branding & Marketing Story.

“I know owned malls are good, but my daily ad budget is only 10,000 won. Does that even make sense?”

This is the most common question I get, and it rests on the most fundamental misconception. The question assumes that “the value of 10,000 won a day is measured by the sales that 10,000 won generates.” If that assumption is correct, the answer is clear: 10,000 won a day won’t get you anywhere. But that assumption itself is wrong.

Let me start with the honest side of the counterargument.

Can Meta ads learn with 10,000 won a day?

Meta provides this guidance on when an ad set exits the machine learning phase: the phase ends when the ad set can be delivered consistently, which generally happens when there are about 50 results within a week after the ad set’s last significant edit. If this threshold can’t be met, the delivery column displays “Limited Learning”.

What matters is that these 50 results are applied per ad set, not across the entire account. Meta’s “Limited Learning Information” document explicitly identifies the ad set as the unit of state transition and suggests consolidating ad sets and campaigns as the solution.

10,000 won a day, 70,000 won a week. If your conversion event is Purchases and you need 50 conversions in a week, your cost per conversion must be 1,400 won. Commerce doesn’t produce numbers like that. In other words, a purchase optimization campaign running on 10,000 won a day will most likely get stuck in Limited Learning. If you split the budget across multiple ad sets, things get worse. They all fall short of the learning threshold, and performance comparisons between creatives in that state are unreliable data.

So saying “Run ads on 10,000 won a day” is a lie if you’re claiming it’s based on performance logic. You can disregard any articles making that claim.

Still, there’s a reason I tell you to spend 10,000 won a day anyway. It’s for a different reason.

That 10,000 won isn’t money to buy performance—it’s money to buy assets

Where that same 10,000 won goes determines what’s left after it’s spent.

The key is that last row. What happens to people who browse but don’t buy. If a 10,000 won daily ad brings in twenty people and three buy in a month, the remaining hundreds are losses from a performance perspective. But if those hundreds went through your owned mall, they become pixel events, retargeting audiences, and seed data for lookalike targeting. If they went through the platform, those hundreds belong to the platform’s data.

To be honest though, it takes months before this audience becomes useful as an ad asset. Pixel data is also an asset that depreciates through signal loss and expiration. So at the 10,000 won daily scale, what works first isn’t the pixel but the member database and channel friends. These two don’t depreciate and don’t disappear even if platform policies change.

That’s why campaign design at this stage should focus on building the minimum structure where learning can operate, not on optimizing revenue. Consolidate to one ad set, keep targeting broad instead of narrow, and run tests by swapping creatives within a single set instead of adding more sets. If purchases can’t possibly hit 50 per week, it’s better to move the optimization event up one level. But stop at Add to Cart. If you go down to Link Clicks or Traffic optimization, you’ll attract people who only click cheap links, and those people contaminate your retargeting audience and lookalike seed, destroying the very asset this article is trying to protect. And before you start, nail down that this phase’s goal isn’t revenue but data accumulation. Otherwise, three weeks later you’ll say “ROAS isn’t working” and shut it down.

Doesn’t Smart Store have natural traffic, making it better to start there?

This is the most important part of this article. Ninety percent of people who delay building an owned mall use this sentence as their reason: “Smart Store sends people automatically just for signing up. With an owned mall, you have to build traffic from scratch.”

The first part is true, but the conclusion is wrong. It’s true there’s traffic within the platform. The problem is how that traffic gets distributed.

Naver Shopping search ranking is composed of three axes: relevance, popularity, and trustworthiness. Popularity is what actually determines ranking. Popularity includes click count, sales index (sales performance), review count, bookmark count, and recency. And here’s the critical part—sales index isn’t cumulative total sales but a score calculated from sales volume and sales amount in the recent 2 days, 7 days, and 30 days, and click counts also reflect the recent 7-day data.

What this structure means is one thing:

So saying “Smart Store is advantageous because it has natural traffic” actually means “You’ll have to spend money on your first sales anyway.” You’ll spend that money whether you use an owned mall or a platform. The only difference is what’s left after you spend it, and that’s what the table above shows.

More importantly, what comes after. Since the sales index is based on recent performance, your score vanishes the moment sales stop. One stock-out and sales index doesn’t accumulate, dropping your ranking. Even when you restock, your ranking doesn’t come back until sales start accumulating again. As of August 5, 2026, Naver also started directly reflecting a “mall trustworthiness” score based on Good Service points in search rankings. If shipping is delayed by a day or complaint ratios rise, that affects your ranking too.

So the accurate way to put it is: Platform natural traffic is not an asset—it’s a lease. You pay rent not in money but in recent performance. Miss a month and you’re out. To move back in, you have to pay the security deposit all over again.

And you don’t set those lease terms. On August 13, 2026, Naver announced that it will now optimize keyword-specific ad exposure in real-time based on user response, and their notice explicitly states: “The number and position of ads/general/super-accumulation products are not guaranteed and will be displayed dynamically.” Even the number of spaces where you stand isn’t guaranteed—the platform said it directly.

Finally, the biggest misconception about natural traffic isn’t cost but attribution. Those incoming customers didn’t seek out your brand—they were searching a category and encountered you. Next time, they’ll search Naver again and buy from a different product at the top. Revenue created through natural traffic strengthens Naver’s customer relationship with each repeat purchase, not your brand’s customer relationship.

Top ranking competition eventually means standing in a price comparison line

The real cost of platform dependency isn’t the fee. Sure, transaction fees and sales commissions are deducted at settlement, but those are calculable costs. There’s a different cost that doesn’t get calculated.

The moment you’re displayed in the same shopping section, your brand becomes something comparable. On the screen consumers see, next to your product are other products in the same category lined up by price. Since the interface itself is designed around comparison, even if your brand talks philosophy, consumers’ decisions tend to tilt toward price and review count.

Within the category keyword display window, there are only two ways to win: sell cheaper or buy shopping search ads for placement. Neither is sustainable, and neither builds your brand. When rankings wobble, revenue wobbles. When revenue wobbles, you cut price again. Stay in this structure long enough and your brand becomes raw meat in a butcher shop—judged only by price per gram.

So instead of getting caught up in category keyword rankings, focus on brand name search volume. People who find you through your brand name aren’t coming to compare—they’re coming to buy. And those people don’t get price-compared no matter which channel they come through. You can track this monthly with Naver DataLab and the search ads keyword tool. Just watching this one metric consistently clarifies half of your marketing decisions.

Hidden costs that don’t show on the income statement—channel risk and enterprise value

There’s one more thing. Platform revenue can disappear significantly from a single algorithm change, policy shift, or penalty. Single-channel dependency creates no costs most of the time, then bills you all at once when something goes wrong.

And investors and acquirers understand this structure. Even with the same revenue, a brand with distributed channels and direct member database ownership gets evaluated differently in due diligence. Channel dependency doesn’t show as a line item on the income statement, but it does affect enterprise value.

Owned mall advantage is forged in “deep communication”

Owned mall advantage isn’t fee savings or design freedom. It’s contacting customers directly, repeatedly, with messages that aren’t about selling. This is the essential form of CRM.

Member Acquisition—KakaoSync. KakaoSync is a simple signup solution extending KakaoTalk login. It handles three things in one consent screen: providing user information required for signup (name, email, phone, age range, birthday, etc.), service terms consent, and KakaoTalk channel friend addition consent. So signup and channel friend acquisition happen at once. Place advertising information consent as an optional selection alongside it. This becomes the basis for CRM messages later. Add the autologin feature and reduce dropoff from customers coming through KakaoTalk. Toss in a signup coupon and signup friction drops noticeably—the right amount varies by AOV, so test two or three values and decide.

Re-contact—just don’t send only discount messages. Most brands that acquire KakaoTalk channels then use them only for discounts and event announcements. The result is being blocked. Once blocked, recreating that friendship is harder than acquiring it in the first place. Brand philosophy, product development history, usage guides, information customers actually want in that category—non-selling messages need to be mixed in so selling messages can get through.

Community—KakaoTalk group chat. Don’t just get them to join and stop. Create a room with only your most devoted consumers, give them benefits first, and ask for product improvement feedback. This group delivers more value in product decision quality than in sales contribution. Even with multiple product categories, don’t split from the start. Several small rooms become several dead rooms.

Content—one article weekly. Don’t just put products on your homepage. This is where owned mall and platform diverge most dramatically. And it doesn’t end as sentiment. That article page gets caught in purchase paths. Looking at last-click alone makes this invisible, but when you unfold the path, you see where articles fit. It doesn’t prove causation, but which content repeatedly appears in purchase paths is something only this metric tells you. And once you know that, what to write next is decided.

You might say “I have nothing to write.” Think again. Look inside a little and things to say flood out. A brand with nothing to say isn’t a brand at all.

GA4 and BigQuery—you don’t have to know how to use them now, but you need to connect them today

If the previous section was human assets, this is machine assets. And this item has one quality different from everything else: You can’t backfill it.

GA4 can export data to BigQuery even from the free property. Once connected, every single event gets ingested daily as raw data. The numbers you see in GA4 are already aggregated and processed results, but what lands in BigQuery is raw data before processing. Row by row records who saw what in what order and where they dropped off.

The reason people delay is always the same: “I can’t write SQL.” “I don’t have time to look at this now.” True. You don’t need to look at it now. You don’t need to use it now. This isn’t turned on to use today—it’s turned on for someone to use next year.

Two years from now, whenever you hire an analyst, bring in an agency, or ask AI—the first thing they ask is “Do you have data?” A brand with two years of raw data and one without can do completely different things. With it, you get things like:

• Channel attribution recalculated on full path basis, not last click
• Actual repurchase cycles and LTV by cohort based on signup date
• Which scroll section of a product page causes dropoff
• Reconciling the discrepancy between what ad platforms report and actual checkout conversions
• How second purchase probability differs by first purchase product

None of these are possible with platform reports. Platforms only give you metrics they’ve decided on in forms they’ve decided, not raw data. This is where owned mall and platform gap widens second only to member database.

One practical note: BigQuery charges by query scan volume, not storage. If you deploy queries that scan entire raw data tables without conditions and run them on short cycles, costs spike unexpectedly. Partition your queries by date and create aggregation tables for frequently used metrics. We corrected this late and that’s when our costs normalized.

Should I close Smart Store then?

Ultimately, yes, you should. But not today. These aren’t contradictory—it’s a matter of sequence.

Closing the platform now causes two things simultaneously: cash flow stops, and that cash flow was the funding source for moving to owned mall. Plus the search demand we vacate doesn’t disappear—competitors take it. The reviews we’ve built stay trapped on the platform. That’s why you need a migration plan, not a decision.

Most people misunderstand here. They say “We’ll run both” and three years later they’re still at the same split. Coexistence should be a period, not a state. Without an end date and migration speed, it’s not coexistence—it’s stagnation.

Four Stages of Gradual Migration

Timelines stretch depending on brand circumstances. What matters isn’t the number of months but whether stages are progressing. So watch just two metrics to judge if migration is actually happening.

Many ask about pricing strategy between channels, but I don’t recommend fluctuating prices across channels. You need to first check platform entry terms and channel pricing policies, and more importantly, price-driven movement goes back when price is gone. Create owned mall transition reasons that platforms structurally can’t replicate—points/rewards, bundles, exclusive early access, exclusive services.

Finally, the criterion for when to close. Once owned mall revenue exceeds half total and the majority of repeat purchases start happening at owned mall, the platform’s role shifts from revenue channel to first-touch customer acquisition window. From there, whether to reduce SKUs or maintain them is your brand’s choice, not survival necessity. Reaching that state is the destination of this entire article.

What should our brand do first?—Decision framework by situation

Order for starting with 10,000 won daily

Order matters. If measurement isn’t set up and you burn ad budget, that money disappears without leaving data.

1. Launch owned mall. Whether Cafe24 or Imweb, do it first. Wasting time on tool comparison is the most expensive waste.
2. Measurement first. Install Meta pixel and GA4. This comes before ads. It works at DIY installation level without developers.
3. Simple payment conversion integration. Make sure purchases through Naver Pay and KakaoTalk buttons register as conversions in GA4. Without this, ad data and actual sales keep diverging.
4. KakaoSync + signup coupon + ad consent. Reduce signup friction while securing channel friends and sending authority together.
5. Then ads. One ad set, broad targeting, swap only creatives. KPI for this stage isn’t ROAS but event count accumulating in pixel and new member count.
6. One article weekly. Ads have harder ROI than this. This is where the gap forms.

Can Meta pixel and GA4 be installed without a developer?

“I’m told to install pixels but I don’t have a developer”—this is actually the most common stopping point. The short answer: yes, it’s possible without developers. Here’s the combination we use, based on Cafe24. Imweb has built-in integrations and separate apps; simple payment conversion handling varies by platform, so check before setup.

※ We have no affiliate or compensation relationship with the apps below. The links contain no referral codes, and this is based on actual operational standards. Prices and features may change as of August 2026.

GA Booster doesn’t cost what I’d call cheap at 22,000 won monthly. But the problem it solves is simple payment conversion loss. If purchases through Naver Pay and KakaoTalk buttons don’t register as conversions in GA4, every ad decision happens on flipped numbers. ROAS shows lower than reality so you kill campaigns that work; attribution inverts so you misdirect budget. Direct implementation requires dev work and re-doing it with each platform update. I recommend comparing other apps too, but we use this combo after factoring in support response and setup time. Note: 22,000 won monthly is barely two days of 10,000 won daily ad spend. Ad spend burned in wrong measurement costs way more.

Keepgrow is what threads together the member acquisition and re-contact I described in this article. One-second signup cuts friction, auto-pulls Cafe24 customer data, and auto-sends messages to post-signup non-buyers, birthday customers, and post-purchase N-day cohorts. This is the part where “accumulation” rolls forward without manual touch. It supports both Cafe24 and Imweb, but customer data sync range varies—confirmation needed. Check pricing on the app store and choose the plan.

Drawing out tool selection is the most expensive waste. If the combo above doesn’t appeal, use something else. Just don’t launch ads without measurement.

You don’t need to start big. Just five people a day accumulates to enough for CRM messaging in a year, and birthday coupon repeat buyers start coming. Numbers teach you what accumulation means first.

If you’re a brand, start with an owned mall. Don’t waste time hesitating.

Frequently Asked Questions

Insights from Soulpapa Marketing — Korea’s digital marketing agency.
Original Korean article: https://soulpapa.co.kr/2026/08/13/jasamall-vs-smartstore-ad-budget/