I hit 100 users today across 13 Korean scrapers on Apify.
Not 100 signups. Not 100 trial runs. 100 distinct accounts that ran at least one job against Korean data — Naver, Melon, Musinsa, Daangn, Bunjang, and more.
I've been tracking these numbers daily since the scrapers went live in mid-March. Here's what the distribution actually looks like — and what it tells me about who's using Korean data APIs.
The Numbers
| Actor | Total Users | Active (7d) | Total Runs |
|---|---|---|---|
| naver-place-search | 23 | 3 | 1,249 |
| naver-place-reviews | 15 | 4 | 581 |
| naver-blog-search | 14 | 2 | 738 |
| naver-news-scraper | 7 | 1 | 10,942 |
| musinsa-ranking-scraper | 7 | 2 | 63 |
| daangn-market-scraper | 6 | 0 | 51 |
| naver-kin-scraper | 6 | 1 | 100 |
| naver-webtoon-scraper | 6 | 0 | 45 |
| bunjang-market-scraper | 5 | 0 | 42 |
| naver-blog-reviews | 3 | 0 | 606 |
| yes24-book-scraper | 3 | 0 | 39 |
| naver-place-photos | 2 | 0 | 37 |
| melon-chart-scraper | 2 | 1 | 46 |
Total: 100 users, 14,541 runs
Three Things the Distribution Reveals
1. Naver Place owns 40% of my users
naver-place-search (23), naver-place-reviews (15), and naver-place-photos (2) together account for 40 users. That's not 40% of any one scraper — that's 40% of my entire portfolio.
Naver Place is South Korea's dominant local business directory. If you're doing market research, brand monitoring, or competitive analysis for the Korean market, you almost certainly start there. The demand wasn't created by marketing — it existed before I showed up.
2. High volume ≠ high users
naver-news-scraper has 10,942 runs from 7 users. That's ~1,563 runs per user on average.
naver-place-search has 1,249 runs from 23 users. That's ~54 runs per user on average.
These are fundamentally different usage patterns. The news scraper looks like a handful of automation pipelines running on schedule. The place scraper looks like independent researchers doing one-off queries or periodic checks.
Revenue implications: the high-volume users are valuable but fragile. Lose one and you lose hundreds of runs per month. The many-small-users model distributes that risk.
3. "Active in 7 days" reveals the real baseline
The 7-day active column is more honest than total users. Some accounts ran once in March and never came back. The 7-day number shows who actually relies on these scrapers right now.
naver-place-reviews leads at 4 active users despite being second in total users. That's a good sign — recent growth, not just legacy numbers.
What I Didn't Expect
The long tail is longer than I assumed.
I built the portfolio expecting 2-3 scrapers to carry the load. That's partially true (place-search and news dominate run counts). But user-wise, the distribution is flatter. Six scrapers have 5+ users. Three scrapers I thought were niche (daangn, kin, webtoon) each have 6.
Korean internet has more specialized use cases than I modeled. Someone wants webtoon data. Someone else wants secondhand market prices. These aren't the same person, and they're not using the same scraper.
The Show HN Problem
In four days I'm submitting to Show HN. My current working title is something like "Show HN: I built 13 Korean data scrapers, they now run 14,000 times a month."
The user count matters there. 100 users sounds more concrete than "14,000 runs." Runs can be one person with a cron job. Users — even 100 — suggests something more distributed.
But the honest framing is both: high automation (14,000 runs) and real breadth (100 accounts). Neither alone tells the full story.
Next
The 7-day active column is what I'll watch. Not total users — that only goes up. Active users can drop. That's the signal I actually care about.
If you're building on top of Korean data or thinking about scraper monetization, I'm happy to compare notes. The Apify PPE model has some quirks worth knowing about before you commit to it.
14 scrapers, 100 users, 14,541 runs. Day 21 of month 2.
Top comments (0)