10 min read7 viewsinvesting

Subtract What You Can't Do: Takahiro Kato on His Decade at Recruit and the Seed of Jimoty

Dawn at the foot of an apartment stairwell, a sofa resting on the bottom step, the stairs receding upward into shadow with light falling from the top landing

Investor's Sunday, 15 November 2026, with Jimoty president Takahiro Kato: from selling ads to one-room craft schools, to turning around a Car Sensor that trailed 7-to-3, to picking a business where sixty million people each pay ¥500. Listener notes, educational, not investment advice.

  • careers
  • platform business models
  • Jimoty
  • Recruit
  • user participation
Contents
  1. I assumed someone who gets this far could already see the finish line
  2. He had no map; he had a question he could not put down
  3. Where I land: sixty million people at ¥500 each, and what has to hold
  4. Further reading
  5. One thing to take with you

Dawn at the foot of an apartment stairwell, a sofa resting on the bottom step, the stairs receding upward into shadow with light falling from the top landing

The Master said: “Do not be troubled that others fail to know you; be troubled by what you cannot yet do.”

—— Confucius, Analects, “Xian Wen” (Spring and Autumn period; my own translation)

The 15 November 2026 episode of Investor’s Sunday hosts Takahiro Kato, president of Jimoty, and spends the whole hour on his 2001–2011 decade at Recruit. Kato says he never worked backward from a thing he wanted to build; what pushed him was a different discomfort — he did not want to meet the version of himself who wanted to solve something and could not — so he kept taking on whatever he had not done yet, and only after ten years could he see what he wanted to make. When he took over product for Car Sensor, the magazine trailed its rival Goo-net 7-to-3 in share, and the turnaround took him over three years; in 2011 he picked the side of the market where sixty million people each pay ¥500, on the arithmetic that it needs ten million monthly users before it works at all. This is part one — the IPO and this year’s tender offer come next week.

I assumed someone who gets this far could already see the finish line

I have a bad habit when I read a founder’s history: I read it backward and turn every stage into groundwork. “Three years in sales, two in internet marketing, then product, then editor-in-chief, then a platform for second-hand goods” — my head fills in a straight line, as if he knew in year one what year ten was for and spent the gap collecting parts.

On the left, a single one-way line runs through four job titles in order and then on to starting a company; on the right, the same four titles sit in a ring with every arrow pointing back to one question at the center.

Underneath that reading sits a more specific belief about my own working life: find the thing you want to do first, then stack the expertise that serves it, and every change of team or subject scatters what you have stacked. I have measured myself with that ruler, and friends’ job moves too. Its only marking is “is there a through-line”, and anyone who cannot name one is wasting time.

I also assumed that someone who ends up running a company must have wanted to run one at some point along the way. All three of those came out of this episode changed.

He had no map; he had a question he could not put down

Kato joined Recruit in 2001 selling ads for Keiko to Manabu, a magazine of lessons and classes that no longer exists. He did cold calls, which meant the smallest accounts on the list: pottery schools, calligraphy schools, one room at a time, monthly ad spend in the tens of thousands of yen.

He describes the knot in his head at the time precisely. Sales productivity is money per hour of your attention, so the rational target is a national English-conversation chain with a big budget. A pottery school has no intention of expanding and will never buy much advertising, and he knew while he was doing it that walking a street door to door was poor use of a day. Yet when he stepped inside those rooms and looked at what they taught and what that location meant to the people nearby, his read was that the school mattered to its district and someone should carry word of it outward — which is the job the magazine was supposed to be doing. The productivity of advertising and the role of a medium that gets information to the people who need it would not reconcile. He says his numbers were so poor the company had no idea what to do with him.

A two-axis scatter plot with ad budget on the horizontal axis and meaning to local residents on the vertical axis: the pottery class sits in the top-left corner and the cram-school chain in the bottom-right, so the productivity yardstick only measures the right half.

I listened to that part twice. What Jimoty does ten years later is make the people who cannot afford advertising visible. The seed was planted in his worst sales year, and at the time it registered as a contradiction, not an answer.

One more detail from that stretch that I love: once the list ran out, he was sent to pitch building tenants who ran no school at all on whether they would like to start one. He calls it a nuisance his employer inflicted on the city — see a foreigner on the street, wonder about an English class. He says sales is still a third-rate skill of his, that he has no talent for it, and the only thing he kept was the nerve to knock.

In 2003 he was moved to a unit doing internet marketing across all of Recruit’s media. The company then ran dozens of properties, the larger ones each pulling revenue on the order of hundreds of millions of dollars, and the argument for one horizontal team buying traffic for all of them was winning. He read the transfer as a demotion. Google’s ad system then took off, he became Recruit’s counterpart to Google and Yahoo Japan, and within a year he was steering an advertising budget on the order of a hundred million dollars, with Recruit and Amazon trading places as Google’s largest spender in Japan. People told him he was the person in the country who understood this best, and asked whether he wanted to move over and make a career of it.

He says the interest would not come, and that is the moment he learned he has no appetite for going deep on one specialty — staying there, he would become a boring person. Two years in he lobbied his way back to the business side, to product and editorial for From A and Town Work.

The mechanism that job taught him is cleaner than most business books. For a part-time job medium, the ad price is set by the client’s own arithmetic: how many hires can this listing bring me, therefore what is it worth. So product and editorial compete on how many applicants they can deliver to which client — in print that is page size and layout, online it is result ordering, screen construction and how the search filters are built. Those choices decide whether a user ever reaches a given listing.

What he liked was the speed of the feedback. Recruit had a saying: a rep on a weekly grows faster than a rep on a monthly, because the closes come four times as often and you take four times the at-bats. Online, a screen you change today and a bidding rule you touch today show their result tomorrow, so you can stay in the batter’s box indefinitely. He notes this reads as long hours, and that his own problem was the reverse — a day without an at-bat bothered him, so he ran experiments every day. What stays with me is that the thing he loved sat in the frequency of feedback rather than in the wins; whether a person can keep at something for years usually turns on that.

Then Car Sensor, where he was head of product and editor-in-chief at once. The used-car medium trailed Goo-net 7-to-3 when he arrived, and he explains why that number matters. In a two-horse market, 5-5 or 6-4 is still a contest; 7-3 dismantles your sales force, because a client tells your rep that for the same money the other listing brings twice the walk-ins, the rep has no answer, and back at the office the hunt for a culprit begins — product arrives from above, so blaming product is the easy move.

Three timelines compare how many at-bats fit into the same span of time: the monthly has a single dot, the weekly has four, and the web has a dense row of dots.

His first trip was to the Nagoya office to introduce himself. Nagoya is his rival’s home ground; seventy or eighty reps were sitting there and he was shouted at before he finished his greeting. In Kansai it went further and a drinking session turned into a scuffle. His reading: whoever shouts loudest wants to win most. So he asked those people to take him to their clients and started from rebuilding the client’s trust in the medium, alongside them. The ones who later backed him hardest, he says, were that same group.

Four horizontal share bars splitting left and right: the 50/50 and 60/40 bars still look close to even, but at 70/30 our own segment is visibly much shorter and is flagged as a warning, and the bottom bar shows the 70/30 split flipped back in our favour after a little over three years.

For the next three-plus years he rebuilt product and brand end to end. The method was to go sit in used-car dealerships — wash cars, use their back-office screens, watch the whole floor process, see how staff greet someone walking in and how they work out which medium sent that person. His conclusion was that the user-facing search side needed changing and so did the management console the client touches every day, a full rewrite of both. Three years on, the share flipped.

He treats that stretch as his dividing line: before it he was training how to solve a problem, after it he began training why this problem and which problem is the real one. And what he says he is actually good at is pulling businesses that look unrelated apart into functions, seeing the shared structure underneath, then pushing back out to the specific. Ten years in, that set of skills plus the question he could not put down added up to a sentence: this one is mine to solve.

Where I land: sixty million people at ¥500 each, and what has to hold

If you also worry that your history has no through-line, here is how I see it now. Kato’s accumulation does not sit in the titles; it sits in the question he kept returning to. Sales, media buying, product, editor-in-chief — the thread across all four is who pays and who gets seen. Same question, a new seat each time. So I changed my ruler: failing to name a career through-line is fine, and naming one question you circle back to every few years is the through-line.

The same cut is usable on investing. When he left Recruit in 2011 he had not thought about being an executive, about equity, about listing. His account is that he had learned what Recruit had to teach, and what he wanted next was the mirror image of his own strength: Recruit trained him to serve the big spenders well and build media of high importance to them, and he wanted the kind where each person contributes little and the pooled value is large. He cites two: Wikipedia, which nobody pays for and which became the best encyclopedia in human history, and Tabelog, which runs on a different footing from Recruit’s own Hot Pepper because the medium is constituted by the people writing reviews — user participation is the content.

He names the two routes: client-agent, and customer-agent. Between a service where a hundred thousand customers each pay ¥1,000,000 and one where sixty million people each pay ¥500, he wants the second, because he reads its cultural staying power as higher. He says he asks himself whether the service will still be here in a hundred years.

Two blocks compared side by side: the left one is narrow and tall, standing for a few clients paying a high price each, and the right one is wide and short, standing for a huge number of users each paying a little.

When I look at holdings I get taken in by lines that elegant, so the condition has to go back in. The threshold is in his own words: monetisation needs ten million monthly users, and below that there is no cosy middle state — fall short of scale and nothing converts. He walks through both models on air: few clients at high unit price, or a mass of users each contributing a trickle, with the second requiring platform-level headcount first. He offers Craigslist as the forerunner, saying it began as a mailing list in 1995, became a website in 1997, runs on about fifty employees, earns margins he calls absurd, has kept the same layout for years and has settled into local life. Those figures are his account on the show and I have not checked them, but even discounted, the leverage explains why he was willing to bet on scale.

The limit worth keeping sits in the same passage. Kato says Craigslist holds users at a distance and works as an information board, while Jimoty has to sit closer to how people actually live, because the era moved. Translated into business terms, headcount will not stay as light as Craigslist’s, and the unit economics have less room than the fifty-person version.

A curve hugs the ground with almost no height until monthly actives reach 10 million, rising steeply only past that dashed line, with the whole area to its left marked as buying you nothing.

One more thing this episode never touches: it is Recruit from start to finish. Jimoty’s competition, how Japan’s second-hand market changed afterwards, the numbers after listing — none of it, all of it next week. So “cultural staying power equals durability” appears here as his reason for choosing the problem; it has not been tested. I filed it under questions worth asking, not under criteria.

Further reading

  • Investor’s Sunday, 15 November 2026, with Takahiro Kato; Jimoty’s listing and this year’s tender offer continue next week
  • Jimoty’s own disclosures (annual securities report, results presentations), for what monthly users and monetisation actually look like
  • Craigslist itself, against Kato’s line about a layout unchanged for years
  • Recruit Holdings’ business presentations, for the scale and revenue sources of the properties he describes

One thing to take with you

One idea: subtracting what you cannot do is easier to start than finding what you want to do. What you want is mostly visible only in hindsight, while “if I needed this next week, I could not do it” is a list you can write today.

Here is a thing I tried. Take a sheet of paper and write three things you could not do if next week required them, specific enough to act on — I cannot read what my family’s insurance policy actually covers, I cannot explain my job once in language my elders follow, I have never booked my own health check. Then pick the smallest one and do it once today, badly is fine.

This article is an educational discussion of investment method. It is not advice to buy or sell any individual security, offers no target prices, and does not analyze any current holding. Investing carries risk; make your own decisions or consult a qualified professional.

Comments

Loading comments…

Sign in with Google before posting. Only your name and profile picture are shown.