Chief Product Officer / Japan
Three decisions that cost us something measurable
- Household electricity contracts
- 94,000
- March 2026; LP gas serves a further 176,000 homes
- Households on the two-band tariff
- 33,400
- One cheaper window, 10:00–15:00 daily
- Solar-and-battery subscriptions
- 5,200
- Homes whose feed-in tariff contracts ended
Decisions & consequences
01Choosing a two-band tariff over a better-priced four-band design
In the autumn of 2022 our supply and trading team proposed a four-band time-of-use tariff for smart-meter households. It had a cheap overnight band, a cheaper midday band when solar output depresses wholesale prices, a standard band and an expensive evening peak, with hours varying by season. Ryohei Sakuramine, who leads supply and trading, modelled an annual procurement benefit of about ¥150 million if 30,000 households joined and followed it. The model was sound; I doubted customers would follow it.
A trial through the winter of 2022–23 put 800 households on each design, the alternative having one cheaper window from 10:00 to 15:00 daily. After six weeks, 38 per cent of the four-band group could name their cheapest hours without checking, against 81 per cent of the two-band group. The four-band group moved more consumption overall, 7.8 per cent against 6.9 per cent, but the gain came mostly from younger households with programmable appliances. Customers over seventy on four bands moved 2.1 per cent and generated most of the trial's billing calls.
I recommended the two-band tariff for general sale. Ryohei argued that we were pricing away a real cost signal and that customers would learn. The president accepted my recommendation on condition that the four-band structure remain available to households whose battery or heat-pump water heater could follow it automatically. Adjusted for the trial results, the choice gave up about ¥25 million a year of modelled procurement benefit.
By March 2026, 33,400 households were on the two-band tariff and 2,100 on the automated four-band version. Billing complaints on the two-band tariff ran at 0.4 per 1,000 bills. Supply and trading now publishes an annual estimate of what the simpler design costs.
What I would do differently: when we moved the window by one hour in spring 2025, 600 customers called. I should have published in advance the conditions under which the window could change.
Paying power-saving rewards that did not penalise frugal households
Our first power-saving events, in the winter of 2022–23, paid only for measured reduction, calculated against each household's average use on recent comparable days. Within a month the customer centre was hearing from older customers who had taken part and received nothing. Their evening use was already low, and the baseline left them no room to reduce. Several said they had switched off their heaters to try. Evenings in Matsumoto regularly fall below minus five degrees in January.
For the winter of 2023–24 I proposed two changes. No event message would suggest reducing heating; we would name the washing machine, the dryer and water heating, and say that warmth comes first. Each participating household would also receive ¥50 in points for acknowledging and taking part in an event, in addition to ¥15 for each kilowatt-hour reduced. Kazumi Oterada, our chief financial officer, objected that the participation payment rewarded households for nothing measurable. She was partly right, and we agreed a limit of ten participation payments per winter.
The scheme ran nineteen events with 27,600 households enrolled by app, telephone or postcard. Rewards cost ¥41 million, against an estimated ¥26 million under a reduction-only design. Average measured reduction was 0.28 kilowatt-hours per participating household per event. Peak procurement and imbalance costs avoided were estimated at ¥17 million, so the events did not pay for themselves that winter, and I have never presented them as though they did.
What changed was participation among older customers, which rose from 9 per cent of those enrolled to 31 per cent, and the end of complaints about unrewarded effort. Enrolled households also left us at half the rate of others, although I cannot separate that from the kind of customer who chooses to enrol.
What I would do differently: I would have published the baseline method in plain language from the first event, and asked the customer centre to review event messages before they were sent rather than after the first complaints arrived.
Showing solar owners a battery comparison in which our product often lost
Between 2019 and 2023, about 15,000 households among our gas and electricity customers reached the end of their ten-year feed-in tariff contracts. Their export price would fall from the forty-yen range to single digits per kilowatt-hour. Our home-services sales team, led by Kosuke Tamanoi, planned a battery subscription at ¥6,800 a month for ten years, including installation and maintenance. Their draft app comparison showed a battery's electricity savings without deducting the fee.
I insisted that the comparison use each household's own half-hourly data, deduct every fee and sit beside two alternatives: selling surplus power to us at ¥9 per kilowatt-hour, or remaining with the existing purchaser. Kosuke argued, fairly, that gross savings were how other sellers presented batteries and that customers valued backup during outages, which no calculation captured. We agreed to describe backup in words, including the hours a typical battery could run a refrigerator, lighting and a heater fan, without converting it into yen.
For 72 per cent of eligible households, the battery produced a net cost of between ¥1,500 and ¥4,500 a month. We showed that. The sales plan assumed 8,000 subscriptions in the two years after the October 2020 launch. We reached 3,900. Over the same period 11,200 households signed surplus purchase contracts, which earn us far less per home.
First-year cancellations on the subscription were 1.2 per cent, against 5.8 per cent in a 2019 pilot that had used the savings-only comparison. Complaints about our battery sales to local consumer affairs centres were zero across those two years; the smaller pilot had produced four. The subscription base has since grown to 5,200 homes, mostly through customers who valued backup and said so.
What I would do differently: I would have explained much earlier what happens at the end of the ten years, including battery replacement and ownership. That became the main question in our 2025 renewal survey, and we answered it late.