Median Monthly Revenue by Acquisition Channel — Word of Mouth ¥7.5M, Social ¥2.25M, Marketplaces ¥260K

The question “which acquisition channel earns the most?” usually has its causality backwards. To check, we aggregated the 235 published cases that disclose actual monthly revenue by the main acquisition channels recorded in each article’s frontmatter (multiple channels allowed). Figures are as of August 18, 2026.

Median monthly revenue per channel

Between the top, word of mouth and referrals (¥7.5M), and the bottom, marketplaces (¥260K), lies a spread of roughly 29x. The most common channel is social media with 108 cases, at a median of ¥2.25M. Most channels sit above the overall median of ¥1.6M because cases with multiple channels tend to have larger revenue (a case can be counted under several channels).

There is a second regularity in who sits where. The high-median channels — word of mouth, sales, content, all require time or accumulated relationships before they start working. The low-median channels, launch sites, marketplaces, are usable today, with same-day exposure. Immediacy and revenue run in clean inverse correlation across this table. Starting on the quick channels and shifting weight toward the slow ones is the path most cases actually walked.

Channels do not decide the earnings

Reading this table as “do word of mouth and get ¥7.5M” is a mistake. The causality flows mostly the other way.

Word of mouth becomes the main channel for products whose customers bring customers. That is, businesses that already have a strong product. Sales/direct (¥5.79M) sits high because human selling only pays off in high-ticket B2B: the channel did not create the earnings. only businesses with profitable price points can afford that channel. By the same logic, the marketplace figure (¥260K) reflects the scale of what gets listed on BOOTH, app stores, and skill marketplaces, largely low-priced digital goods and personal apps. Where the platform does the customer acquisition for you, businesses too small to invest in acquisition gather. That is a rational choice, not an inferior strategy.

Paid ads (27 cases, ¥4.5M) landing above the middle follows the same reasoning. Advertising is a cash-up-front channel that only businesses with lifetime value reliably above acquisition cost, those whose unit economics already work, can sustain. Being able to run ads at all is proof of maturity. Newsletters (34 cases, ¥3.36M), being an owned list independent of any platform, tend to appear as a secondary channel that turns readers acquired via content or search into an asset.

In short, this table is not a channel report card but a mapping between business models and channels. Once price point and customer segment are fixed, the usable channels are largely fixed too. The practical use of the table is not to “aim for a higher row” but to find the cases in the same row as your own price point and customers, and read what separates the winners within that row.

What the table still tells a practitioner

Still, a few implications hold for choices within the same business model.

One fact stands out: social media (108 cases) is the most used channel yet ranks third from the bottom by median. Nearly half of the 235 cases name social as a main channel, more than any other. Because entry cost is near zero, everyone uses it, small businesses included. “Just post on social” is the hardest channel to stand out in, precisely because of its denominator. The same “social” means completely different things when posting is itself a product demo, as in the social cross-posting tool case, versus when it is mere announcement.

Another pattern: content (¥5.1M) and search (¥4.5M) sit thick above the middle. Acquisition through blogs, documentation, and technical articles starts slow but compounds. This band holds long-horizon cases like the image-editing SaaS that built to $62K MRR on search alone. It is structurally well matched to solo and small operations that can pay time instead of money.

And launch-site traffic (¥1.31M) is a first exposure, not a recurring channel. The low median among cases naming Product Hunt-style launches reflects businesses caught before transitioning from launch-day gusts to a permanent channel. A launch is ignition, not fuel, what these cases moved to next appears repeatedly in the “what didn’t work” sections of their individual articles.

Method and limits

The population is the 235 cases disclosing actual monthly revenue. Undisclosed cases are excluded. Channels are the “main acquisition routes” read from sources at write-up time, not exact revenue shares. One case can count toward several channels, so columns are not independent. Medians for single-digit channels like media coverage/PR (6 cases) should be treated as reference values. And the largest limit is survivorship bias. The table holds only businesses that did well enough to disclose. Those that died on each channel are absent.

Channel and revenue definitions follow the data page’s “How to read the numbers”. Distribution by business type is in the business-type map, and the solo/side-project revenue distribution in this column. Individual cases can be filtered by channel in the case list.