The parts of a mobile brand that customers actually meet

I design tariffs as statements of what a customer receives, not only what they pay. When Japanese operators lowered prices from 2021, much of the reduction came from removing things: shop assistance, bundled options, contract conditions and sometimes the operator email address. I insisted that our cheaper plan carry a written list of those removals, short enough to sit on the plan page beside the price. I also prefer fewer plans. Thirty-one variants in 2017 became nine for new customers, and I now ask for a reason before any new variant is added. On brand architecture, I argued against a separate low-price brand because it would let us hide the withdrawal of service behind a new name. That position cost revenue in the short term. I would accept a separate brand where the service promise is genuinely different, such as a prepaid product for visitors to Japan, but not as a device for postponing an explanation.

Most of our older customers meet Hinoura in a shop run by an agency company, not by us. The commission we pay those companies decides what their employees emphasise. I have worked on both sides of that arrangement, at an agency and at the operator, and I read the commission schedule as the most persuasive copy we write. My practice is to pay for verified explanation and for customers who remain with us, and content, after six months, rather than for options attached at the counter. I also treat partners as businesses with payrolls. When we change a scheme, I fund a transition and publish the timetable early. Our 2022 redesign included ¥2.2 billion of transition support over eighteen months. It still hurt two smaller agencies more than we had modelled, and I now involve the smallest partners in designing a change rather than consulting them on the result.

I distrust any measure that cannot be wrong. Our attribution model once credited paid search with 44% of customers who switched to us from another operator, mostly because search was the last thing they did before arriving. I asked our marketing science team to test television by withholding it from three prefectures for eight weeks, compared with three matched prefectures. We now run a withholding test every year on at least one large budget line, outside the February-to-April switching season. Before each test, finance and marketing agree what result would move money and by how much. I also accept the cost: the 2023 test lost an estimated 6,800 customers who would have joined from other operators. I rely on specialists for the statistics. My contribution is to make sure the question matters commercially, that the answer is capable of embarrassing the marketing department, and that the board sees the confidence interval as well as the central estimate.

An operator knows where its customers sleep, work and travel. That knowledge creates commercial temptations, and Japan's Act on the Protection of Personal Information sets a legal floor rather than a standard of trust. I sometimes take positions on data use that are stricter than our legal colleagues require, and I say so openly in internal papers so that the cost is visible. For our advertising audience product we required explicit opt-in, offered a monthly data allowance in return and showed customers a plain description of what advertisers would and would not receive. I work closely with our privacy office and legal team, and I do not give legal advice. My role is to ask whether a customer who read the complete description would feel used, and to test that question with real customers, including customers over seventy, before anything is launched.

Japan's mobile customer base is ageing faster than its tariff tables. Many of our customers over seventy-five still visit a shop every few months, often for help that has nothing to do with a sale. When we closed the 3G service in March 2026, about 412,000 lines remained on it. I treated the closure as a marketing responsibility, not only an engineering milestone. We contacted 97% of those lines by letter, telephone or shop visit, offered a simple handset at a reduced price within the handset-discount rules, and kept shop appointments free for affected customers. Some customers moved to other operators, and some lines had been kept open for relatives who had died. I want those outcomes reported plainly. The work confirmed that a network transition is experienced as a broken promise unless someone explains it patiently, in person, more than once.

Generated images and voices are now inexpensive enough that every agency proposes them. I set our rules on three principles. We do not show a generated person as though they were a customer or an employee. We label generated audio and imagery wherever a reasonable viewer might assume otherwise. And no generated voice explains a contract term, a fee or a cancellation right, because a customer who later disputes that explanation deserves to know that a person stood behind it. These rules have cost us production savings that our agency estimated at about ¥400 million a year. I accept that generated material is useful for storyboards, internal testing and first drafts of translation. The question I ask is whether the customer would feel misled if they learnt how the material was made. When the answer is yes, we do not use it, however good it looks.