A registration is not a kitchen

Ask a city gas company how fast it is growing and the answer will usually come in connections. The number is large, it rises every quarter and it fits neatly on a slide. In my experience it is also the least reliable number a utility marketer handles, because one word is used for four quite different things. Unless a board knows which of them it is being shown, it cannot tell whether a city is adopting piped gas or merely signing up for it.

I learnt this in someone else's network before I saw it in ours. At a listed distributor in Gujarat, I once asked for a list of homes where a connection had been recorded but no gas had been billed for a year. It ran to 52,000 addresses. Nobody had hidden them. Each had been counted honestly at one stage of the process, and nobody had been asked to count the stage that mattered.

Four things called a connection

The first is a registration: a household has filled in a form and usually paid something. The second is an installation: the service line, riser and internal piping are in place and a meter is fixed. The third is gas-in: the line has been tested, purged and commissioned, and a flame has been lit in front of the customer. The fourth is consumption: the household cooks on piped gas as its main fuel, which in our cities means at least five standard cubic metres a month once the novelty has passed.

Each stage has a cost. A registration that never becomes an installation costs us the agency fee and the survey. An installation that never reaches gas-in can leave ₹17,500 or more of pipe, fittings, meter and labour idle on a kitchen wall. A gassed-in home that keeps its LPG cylinder as the main fuel costs us the whole installation and pays a small, irregular bill. Only the fourth stage earns back the network.

The gaps between stages are not small. When we opened Amravati in 2019, about 35 per cent of registrations had reached gas-in within six months, and three in ten gassed-in homes were barely using the gas. A slide showing 62,000 connections would not have been false. It would have been a description of intentions.

Paying for the flame

The earlier stages dominate for a simple reason: they are what most acquisition contracts pay for. An agency paid per form will produce forms, and it will produce them fastest where signing is easy and conversion is hard: crowded markets, festival stalls and housing societies whose landlords have not been asked. A contractor paid per installation will install where he can, not where gas can soon flow. None of this is misconduct. People optimise the number on their invoice.

In 2020 we changed the invoice. Agencies were paid only for households still consuming sixty days after gas-in. Installation contractors were paid on gas-in within forty-five days of registration. We cut the upfront payment to ₹1,000 and spread the remaining connection charge over twelve bi-monthly bills, because the deposit was the main reason families who wanted gas delayed the final step. Registrations fell by about 40 per cent for two quarters, and three of our seven agencies left. Those that stayed changed their methods within weeks. They began working through society committees, arranged gas-in evenings where neighbours could watch a meter turn, and declined registrations from rented homes until the landlord had agreed.

By 2022–23, 78 per cent of Amravati registrations reached gas-in within ninety days, and 91 per cent of gassed-in homes were consuming after sixty days. Our all-in acquisition cost per consuming household fell by almost a third. The headline count grew more slowly. The business grew faster.

There is an obvious objection. City gas authorisations carry commitments, and the regulator's reporting has its own definitions, which a company must meet whatever its marketing department prefers. I accept that completely. Our projects and regulatory teams report licence progress on the regulator's terms, and nobody in my team touches those reports. My argument concerns management information, not licence compliance. A company can report what its authorisation requires and still decide that its board, its incentives and its marketing budget will be run on consumption.

What the board pack should count

Our board pack now carries four lines for each area: registrations, installations, gas-in and consuming households, with the conversion rate between each stage and how long each backlog has been waiting. Consuming households lead. When a director sees registrations rising while gas-in conversion falls, the question writes itself, and the answer usually points to a contractor shortage, a meter delay or an agency working the wrong streets.

I would add three disciplines that apply well beyond city gas. First, define consumption in a way the customer would recognise; five cubic metres a month is a household cooking, while one is a household experimenting. Second, report abandoned installations as a cost rather than as work in progress; a meter that has hung on a wall for a year without gas is not an asset in any useful sense. Third, show the conversion rates to the people paid on them. Our agencies see the same funnel the board sees, area by area, every month.

The approach has costs I should name. It slows the visible numbers in the years when a young company most needs to impress its lenders and its investors. It moves risk onto agencies, which then charge more per household and become fewer. And it demands measurement we did not have at the start: reliable meter readings within sixty days, which in our first year meant sending people to read meters by hand.

I still think the trade is right. A city adopts piped gas kitchen by kitchen, through families who find it safe, convenient and fairly billed, and who tell their neighbours so. A marketing function that measures forms will be tempted to manufacture them. One that measures cooking has to earn it.