Chief Marketing Officer / Japan
Decisions that cost response and what they bought
- Policies in force
- 1.32m
- Tomoshibi Life, March 2026
- Annual marketing budget
- ¥12.6bn
- Year to March 2026; television, digital, comparison, print and research
- Cost per policy still in force at thirteen months
- ¥71,000
- Digital channels, year to March 2026; ¥96,000 two years earlier
Decisions & consequences
01Saying what the first year pays in advertising for older viewers
In 2021 our relaxed-underwriting medical plan for applicants aged 50 to 85 was Tomoshibi Life's most advertised product. Daytime and early-evening television spots told viewers that most people could join after three health questions. The plan halved hospital benefits in the first policy year and cost about 1.7 times our standard medical plan at the same age. It sold 41,000 policies in the 2020 financial year, and complaints about first-year payments had risen to 214. When our customer research team compared a sample of applications with standard underwriting, it estimated that about 38% of applicants aged 50 to 69 would have qualified for the cheaper plan.
Older customers were paying more than they needed for less than they believed. I proposed three changes: state the first-year reduction aloud in every advertisement, tell viewers in good health that our standard plan costs less, and put every relaxed-plan applicant through the standard plan's questions first. Our head of direct sales opposed the routing in writing, forecasting a fall of a quarter in relaxed-plan sales and a loss of margin. Kawasemi's head of television sales warned that spoken conditions would depress response more than any other change. Our president approved the changes from October 2021, with monthly reporting to the board.
Response fell as forecast. In the 2022 financial year the relaxed plan sold 26,500 policies, and 9,800 applicants who had begun on it bought the standard plan instead, at premiums about 40% lower. New annualised premium finished ¥0.9 billion below plan, and television cost per policy rose from ¥46,000 to ¥58,000.
By the 2023 financial year complaints about first-year payments had fallen to 61, and first-year lapses on the relaxed plan had fallen from 19% to 11%. Standard-first routing now applies to every product with a relaxed version, and the board receives the share of applicants moved to a cheaper plan.
What I would do differently is the order of the work. We corrected the advertising before turning to customers who had already bought. It was 2023 before we wrote to 12,400 relaxed-plan holders aged 50 to 69, offering a move to standard cover if they qualified, and 4,100 moved. They should have heard from us in the same month the advertisement changed.
Leaving comparison rankings we could not label
By 2023 three comparison sites produced most of Tomoshibi Life's comparison business, and the channel accounted for 29% of new policies at about ¥39,000 an application, one of our lowest costs. Two of the sites placed us in recommended rankings that we paid to enter, with a general fee disclosure at the foot of the page. We also used, in advertising and catalogue inserts, a claim that we were first in satisfaction among online medical insurers. It rested on a research firm's survey in which people who were not our customers rated insurers' websites. With Japan about to treat advertising that hid the advertiser as a misleading representation from October 2023, I asked for a review of every placement and claim.
I knew how such rankings were made because I had built one at Hakari Navi. I proposed that we buy only positions labelled as paid at the point of reading, and that we withdraw the satisfaction claim whatever the legal review concluded. Our chief financial officer argued for a phased exit over a year, because the mid-term plan assumed comparison volume. Our head of digital expected competitors to take the positions we left within weeks, which they did. One site agreed to label each paid position. Two declined, and we left them in October 2023.
We gave up our lowest-cost channel. Comparison sites produced 18% of new policies in the 2024 financial year, and total new policies fell 6%, from 183,000 to 172,000. We moved ¥0.6 billion of the comparison budget into search and into explanation pages that answered the questions readers had been taking to rankings.
Thirteen-month persistency for policies from the remaining comparison partner rose from 84% to 90%, and it now pays part of its fee only when a policy reaches its fourth monthly premium. Every public claim of ranking or satisfaction is reviewed by marketing and compliance together, with the survey design attached.
What I would change is how we left. I told both sites by letter with four weeks' notice; a conversation in person, months earlier, would have kept one relationship intact. I also missed that the satisfaction claim appeared in Kawasemi's printed catalogues, where it stayed for another seven weeks.
Paying partners for policies still in force
In the 2023 financial year Tomoshibi Life spent ¥4.9 billion on digital acquisition. Affiliate networks and point sites, which rewarded their users with shopping points for completing an application, were paid about ¥18,000 per application and produced roughly 22,000 applications a year. Our dashboards showed digital as the fastest-growing channel. When our analysts traced those policies forward, 41% of point-site policies lapsed within four months, typically after the points had been credited. Measured on policies still in force at thirteen months, the channel cost more than television.
The stakes were our mid-term plan, which assumed 190,000 new policies a year, and a reporting habit that rewarded the cheapest application. I proposed leaving point sites altogether, paying affiliates only when a policy reached its fourth monthly premium, and reporting cost per policy in force at thirteen months as our main acquisition measure. Our head of growth marketing argued for a lower per-application fee instead, which would have kept the volume. Two affiliate networks ended their arrangements with us. Our chief financial officer asked me to explain to the board that the plan figure for new policies would be missed, and I did.
We gave up about 14,000 policies a year. Digital cost per application rose from ¥31,000 to ¥44,000 in the 2025 financial year, but cost per policy in force at thirteen months fell from ¥96,000 to ¥71,000. Digital spending fell to ¥4.2 billion. With the actuarial team, we showed that the value of new business from digital channels rose even though volume fell.
The board replaced the new-policy figure in the mid-term plan with policies in force at thirteen months, and every channel is now reported on that basis. We also stopped bidding on competitors' brand names in search, which had produced applicants who believed they were buying from someone else.
What I would do differently is to change the sales measures at the same moment. For two quarters our direct sales team was still judged on completed applications. Advisers pressed hard on call-backs to marginal applicants we had deliberately stopped buying, and some of those policies lapsed exactly as the point-site ones had. I should have agreed the measure with the head of direct sales before announcing it.
Declining to advertise against an industry scandal
In the summer of 2019, revelations of improper sales at a large life insurer dominated the Japanese news for months. Many involved older policyholders who had been led to replace existing cover, sometimes paying two premiums at once. Trust in life insurers fell across the industry. I was then general manager of marketing at Tomoshibi Life. Kawasemi's group television team and our head of direct sales proposed a campaign built on the line that no salesperson would ever visit your home. Early tests suggested it would lift response among viewers over 60 by about a fifth.
I argued against it. A contrast campaign would tell viewers that insurance sellers were not to be trusted, at a moment when our own contact centre was selling to the same viewers. It would also imply a promise about our conduct that no insurer can make in advance. I proposed instead a free policy-check line, open to anyone whatever their insurer, staffed by advisers who were not paid on sales and instructed to say when keeping an existing policy was the better choice. The head of direct sales objected that we would be paying advisers to talk people out of buying. Our president, Masayo Tsuchiya, who had taken office that June, approved a trial for six months and then extended it.
The line cost ¥410 million over eighteen months. It took 47,000 calls between October 2019 and March 2021. Advisers recommended no change in 38% of them, and 8,600 callers later bought a Tomoshibi Life policy, many as additional cover rather than replacements. Among viewers aged 60 to 74, the share who said they would consider us rose from 11% to 16% in our tracking survey. We also published a commitment that no one aged 75 or over would be asked to replace an existing policy in a single call.
That commitment later became the basis of our two-call process for older applicants.
What I would do differently is measure what happened to callers we advised to keep their cover. We recorded the advice, not whether they followed it. I also agreed to close it in March 2021 to meet a budget reduction. We reopened it for callers aged 70 and over in 2023, and the gap cost us the habit customers had formed.