Korean Conversion Atlas
Locate conversion evidence from Korean companies and users, identify usable simulation inputs, and distinguish unsupported figures and missing evidence.
Written and reviewed by Eric Han · Co-founder & CTO, reopt
Key takeaways
- Evidence quality varies by domain. Filings provide stage rates, such as Millie's Library free-subscription-to-paid at 34.7–37.3%, but no official industry ad-channel benchmarks were found.
- Vendor ecommerce aggregates report a 2.06% mean and 0.72–5.05% by industry. These are 2021's top 100 stores; for a new store, widen the interval and assess outcomes at the 5th percentile.
- Frequently repeated figures such as Korean ecommerce 1.33%, Powerlink CTR 0.8–2.4%, and SMS opens 98% lack verified sources or formulas.
- We found no source reporting live-commerce conversion of 5–8% that also disclosed its sample. Cumulative Naver buyers divided by views is around 1%.
- Official Naver/Kakao industry benchmarks, friend-invite-to-signup conversion, and photo-upload effects were not found.
Inclusion Rules
This chapter includes measurements of Korean companies, the Korean market, or Korean users. Global and Asia-Pacific averages are excluded; global companies appear only for Korean campaigns or separately reported Korean figures. Korean companies with global samples are labeled accordingly. Cases are divided into Advertising and Commerce, Subscriptions, B2B, Messaging, and Referrals, and Consultations, Finance, and Public Experiments.
Grades follow How to Read the Evidence. Korean case tables place grades at the end of the first cell.
Success stories are sales material
Advertising platforms and tool vendors select successful customers for publication. Use grade D cases only for direction, not numerical inputs. Explicitly null or failed cases are rare and are included prominently when found.
Where Evidence Exists
| Domain | Strongest evidence | Representative figure | Missing evidence |
|---|---|---|---|
| Paid channels | B, Naver transaction-log analysis | Directly managed advertiser CPC mean 992, median 267, units unstated | Official industry CTR/CVR; reported-to-incremental multipliers |
| Store purchase conversion | C, vendor aggregates | Mean 2.06%; industry 0.72–5.05% | Official average; device rates |
| Repeat purchase | B, securities filing | Buyer → repeat buyer 61.2%, cumulative three years | Other companies with matching definitions |
| Live commerce | C, Naver cumulative announcements | Buyers/views about 0.9–1.0%, calculated | Official Naver/Kakao conversion rates |
| Subscriptions | B, securities filing | Free subscription → paid 34.7–37.3% | Same metric for other apps |
| B2B SaaS | B filings; C surveys | Suspensions + cancellations 4.7–11.5%; inquiry → meeting about 9–42.9%, lower bound calculated | Trial/freemium conversion |
| Email / messaging | C, vendor aggregates | Email opens 9.6–19.9%; AlimTalk clicks 11.52% | App-push averages; recent ATT consent |
| Referrals | D, company announcements | Share-referred participation 65%; viral K 0.4 → 0.7, defined differently from this book | Invite → signup; cannibalization |
| Influencers | B, papers measuring word of mouth | 10% more seeded WOM associated with about 0.4% more nonseeded WOM in specialist forums, calculated | Incremental sales |
| Consultations / appointments | B, hospital records and financial authorities | Outpatient no-shows 5–7%; auto-insurance comparison → contract 9.2% | Photo-upload effects; nonhospital form/KakaoTalk attendance |
| App acquisition | C, vendor aggregates | Korean Apple Ads CPI $1.84; tap → install 51% | Google/Meta app-ad rates |
Simulation Inputs
Potential inputs graded C or higher are below. When contexts differ, keep the mean, lower mean × to 1–2, widen the interval, and assess outcomes at the 5th percentile.
| Stage | Input | Evidence and conditions |
|---|---|---|
| Store traffic → purchase | Mean 2.06%; industry 0.72–5.05% | Top 100 stores, 2021; denominator unclear. New stores should lower mean × and assess outcomes at the 5th percentile |
| Live view → purchase | About 1% | Naver cumulative 2020–2021; views denominator; order-of-magnitude only |
| Free subscription → paid | 34.7–37.3% | One content-subscription company; free starters denominator. Its January–August 2022 value of 42.62% falls outside this range |
| Monthly resubscription | 76.2–87.6% | Same company's B2C only; rose from 76.2% in 2020 to 87.6% in first-half 2023 |
| B2B inquiry → meeting | About 9–42.9% | Self-reports from 93 firms collected by a rapid-response tool vendor. About 9% back-calculated from 4.7×. Faster responders at upper end; tool customers 39% |
| Comparison → contract | Auto insurance 9.2%; travel insurance 62.5% | Platform-attributed financial products; before March 2025 version 2.0 |
| Outbound phone connection | About 9–47%; 72% in a prior-SMS case | Public-information calls, some unclear denominators; not directly submitted leads |
Hospital no-show figures mostly concern follow-up patients and are not consultation-booking inputs. Use your own data.
Korean ecommerce order intervals
Estimate traffic from budget and CPC; choose an industry or channel for BIGIN's 2021 purchase rate. KRW 500 CPC is hypothetical because no public benchmark was found. Replace it with a small internal test.
Each input alone moves to its 5th and 95th percentile with others at the mean. Measure the top one first.
- Traffic → purchase8orders ~ 397orders
- Reported ÷ incremental multiplier106orders ~ 227orders
8,000 runs. Stages are independent. If they move together the real interval is wider; if they move in opposite directions it is narrower. The 90% interval holds only rate uncertainty around the expected value; the actual count adds chance variation in counts. Vertical lines mark the 5th percentile, median (darker), and 95th. The last bin collects values above the 99th percentile. Concentration k is an author judgment chosen for mean × k around 1. Do not compare confidence using k alone. The multiplier is an author judgment based on US Meta/Haus evidence. The 2.06% mean pools all traffic; channel figures use last touch.
Values to Avoid
These common Korean-language claims have unverified sources/formulas or are easily misinterpreted.
| Value | Origin | Why not to use it |
|---|---|---|
| Korean ecommerce average 1.33% | Recap of a 2020 presentation | Presented as hearsay, with no source or sample |
| Powerlink CTR 0.8–2.4%, CPC about KRW 530 | Agency article, citing a “2024 Naver Business Platform Report” for its industry table | No such title found in Naver's report list. The same paragraph's 600 million daily impressions and 48 million clicks imply 8%, calculated |
| Powerlink CTR 3.8%, conversion 4.1% | Another article by the same agency, claiming official statistics | No link; conflicts with the preceding values |
| Bizboard CTR 0.5–1.5% | Agency article | No evidence; another article from the same company gives 0.5–1.0% |
| Remember CTR 2–5%+ | Agency article | No source; conflicts with media-kit forecasts of 0.3–2% |
| SMS open rate 98% | Messaging-vendor article | Mixes overseas/internal figures without explaining how SMS opens were measured |
| No-shows of 20% for restaurants, 18% for hospitals, and other industries | Coverage citing Hyundai Research Institute | Original report not found; years/attributions differ. University-hospital records are mostly 5–7%, up to 14–15%, leaving the 18% definition unclear |
| Dental appointment adherence 56% | 2003 paper | On-time arrival among attendees, not no-shows |
| Three-minute versus 30-minute callbacks triple bookings | Agency article | Only described as industry data |
| Five-minute response 4× / 100×; one-minute response +391% | Tool-vendor article | Conflicting claims within the same vendor's articles; no source |
| Proprietary easy-payment conversion 95% versus regular 70% | Payment-vendor article | No sample/period; strong selection among customers with stored payment details |
| App 21%, mobile web 4% | Criteo 2018 | Korean report, but supporting slide covers Northeast Asia |
| Same-day cart purchase 54% | Yozm IT contribution citing SAP 2018 | Original and sample unverified |
| 300 leads for KRW 5 million; lead costs reduced 500% | Remember sales page | First is hypothetical; a 500% reduction is arithmetically impossible |
| Friend invitations cost one-tenth of paid ads | Utility site's ABLY article | No basis for treating it as company-disclosed data |
| Coupang Partners conversion 3% | Calculator site | Called an observed average but used as an unsupported input assumption |
Missing Evidence
Searches found no public values for:
- Korean industry CTR/CVR benchmarks from Naver, Kakao, or Meta, or Korean reported-to-incremental multipliers. Kakao industry comparisons are available only inside advertiser interfaces.
- Official Cafe24/Naver/Coupang purchase conversion, Korean device rates, or cart-abandonment studies that disclose their samples.
- Participation, invite-to-signup, and cannibalization in paid-reward referral programs at Toss, Karrot, Baemin, KakaoBank, and others.
- Trial/freemium conversion at flex, Channel Talk, JANDI, Speak, or Wrtn; role-level conversion; actual Remember CPL.
- Photo-upload effects, nonhospital form/phone/KakaoTalk attendance comparisons, aesthetic-treatment consultation conversion, or Korean randomized SMS appointment-reminder experiments.
- Korean randomized tests of free shipping, page speed, or friend invitations.
The search itself has limits. “Not found” does not mean nonexistent.
- Direct KCI/RISS searches, Korea Chamber of Commerce and Korea Consumer Agency surveys, Naver earnings calls, and some technical blogs such as KREAM, ABLY, Brandi, idus, and Trenbe were not reviewed comprehensively.
- Form-gated materials were not obtained, including full Stibee and Recatch reports and Remember webinar materials.
- Full texts of Jung et al. 2023 and Chu/Joo 2024 could not be opened.
Conversion Atlas by Role
Review cold-email replies, LinkedIn costs, product conversion figures, and gaps in public evidence for CMOs, developers, HR, finance, founders, and other roles.
Paid Channel Baselines
Establish pre-spend baselines using industry CTR, CPC, CVR, and CPL for Google search, Meta, TikTok, and LinkedIn, and identify gaps for Korean channels.