Chief Technology Officer / Japan
Three decisions that each gave something up
- Company revenue
- ¥6.4 bn
- Year ended March 2026
- Transmitter sites
- 40
- Main transmitter and 39 relay stations across Akita Prefecture
- Technology staff
- 38
- Plus about twenty contractor technicians
Decisions & consequences
01Renewing eight relay stations, moving two valleys to fibre and keeping a third on air
In 2023 eleven of our relay stations had digital transmitters installed between 2007 and 2009 and approaching the end of the maker's support. Renewing all eleven would cost ¥236 million over three years. Three of them served very small valleys of 160, 240 and 410 households, and each needed a four-hour journey by snow vehicle in winter. Our relay team had nine engineers, with three due to retire by 2027.
I proposed renewing eight stations and replacing the three smallest with delivery over a municipal fibre network to the community reception facilities those valleys already used. The news director objected that we had a duty to reach every household by our own transmitters, and our head of sales feared that advertisers would read any withdrawal as decline. One mayor opposed the change at first. I accepted that the plan gave up a principle our station had held since 1969, and that residents would see it as a reduction in service.
We worked with the municipalities, the other broadcasters at those sites and the regulator's procedures, which took eleven months. We paid our share of receiving equipment and fibre connection, ¥26 million in total, and held resident meetings in each valley.
Two valleys, with 400 households, moved to fibre delivery in 2025, and we saved about ¥9.4 million a year in maintenance, power and site rent. The third did not move. A survey showed that its fibre route crossed a slope prone to landslides, and I would not replace a working transmitter with a single cable in that place. We renewed that station for ¥14 million. Including the eight renewals at ¥171 million, we spent ¥211 million against the ¥236 million first estimated.
The relay team now has two fewer winter journeys, and the network has a tested method for future decisions. I would hold the first resident meetings before agreeing terms with the municipalities. We did it the other way round, and residents reasonably felt the decision had been made without them.
Keeping microwave on the backbone and moving fourteen spur feeds to leased circuits
By 2023 the twenty-six microwave hops carrying our signal from the studio to the transmitters were reaching the end of support. Renewing all of them would cost ¥290 million. Our finance director proposed moving every feed to circuits leased from carriers, which would avoid most of that capital spending. Tomoe Shibaike, who manages relay engineering, argued for renewing every hop. In March 2011 our microwave paths had stayed up on battery power while some carrier routes failed, and she did not want our signal to depend on another company's restoration order.
I proposed a division by consequence. The nine backbone hops feeding the main transmitter and the six largest relay stations, serving 86% of households, would stay on microwave. Three spur hops with no second fibre route would also stay. Fourteen spur feeds would move to leased circuits from two carriers over separate routes. The finance director lost most of the capital saving he wanted, and Tomoe lost her preferred design. I wrote into each contract the restoration priority, notification times and route information we needed, because the carriers' own obligations did not cover our broadcast.
The twelve microwave hops were renewed for ¥128 million, and the leased-circuit equipment cost ¥34 million. Annual circuit fees of ¥21 million replaced about ¥8 million of microwave maintenance, so over ten years the two approaches cost almost the same. I told the board that the reason for the change was the smaller team that would maintain the links, not the money.
In January 2026 heavy snow brought a tree down on poles near one spur station. Both carriers' cables ran on those poles for 3.2 kilometres, a fact neither route map had shown clearly. The relay station was off air for 47 minutes, affecting about 2,100 households.
We now require an engineer to walk any section where routes appear to converge. I would have done that before signing the contracts, not after the outage.
Streaming disaster news in one lower-quality rendition so that it stayed available
We began streaming our local news in April 2023, with an annual delivery budget of ¥18 million and a normal evening peak of about 3,800 concurrent viewers. Network programmes stayed off the stream because their rights were not ours to stream. During heavy rain or an earthquake, demand could rise many times over, and a delivery service priced by the gigabyte could exhaust a year's budget in a few days. I also worried that a stream which collapsed during an emergency would fail exactly the viewers we were trying to reach.
In 2024 I proposed an emergency mode: a single 1.2-megabit rendition, no advertising insertion and surge capacity agreed in advance with our delivery provider at a fixed rate. Our head of sales wanted advertising to continue, since emergency streams would draw our largest audiences. The news director wanted high definition so that maps and evacuation information stayed legible. I agreed with her concern but not her remedy. We redesigned the emergency graphics with larger type and fewer details, and the sales team accepted the loss of advertising during emergency periods only after the president decided the matter.
In August 2025 a heavy-rain emergency lasted thirty-one hours. The stream peaked at 58,000 concurrent viewers, more than fifteen times the normal peak, and stayed available throughout. Delivery cost ¥2.1 million, compared with an estimated ¥7.6 million at our normal quality ladder. We gave up advertising worth about ¥3.2 million and received about 140 complaints about picture quality.
One part failed. Switching to emergency mode was a manual decision, and it took twenty-two minutes after the first official warning. During that period the stream buffered for many viewers.
The station now treats the stream as part of its disaster broadcasting rather than a digital extra, and the newsroom rehearses it each winter. I would link the switch to official warnings from the start, with a person able to reverse it, rather than wait for a meeting to decide.