Decisions that slowed growth so the product would hold

Retiring 66 inverter models to make the hybrid the default

When I joined Dhruvanta in April 2021 we sold 112 inverter and UPS models. The range had grown through decades of distributor requests: a cheaper square-wave model for one state, a different cabinet for a large retailer. Our sales data showed something the range did not reflect. In nine of our fourteen main states, replacement demand for backup inverters had fallen for three years as supply improved, while in Bihar, Jharkhand and eastern Uttar Pradesh it was still growing. Backup was not disappearing, but in most of our markets it had stopped being a reason to buy something new.

I proposed cutting the range to 46 models and making a solar-ready hybrid inverter the default at every capacity from 1 kVA upwards. The hybrid carried a built-in solar charge controller for up to three kilowatts of panels and cost the customer about 18 per cent more than the equivalent backup model. Our national sales head disagreed. He argued that entry-level square-wave models were how we competed with unbranded assemblers, that eastern distributors earned their margins on them and that a dearer range would lose shelf space. He was right on each point. We compromised in the east, keeping nine entry models for Bihar, Jharkhand and eastern Uttar Pradesh until March 2024.

The cost was clear. We gave up about ₹140 crore of annual revenue from retired models, and backup revenue fell 6 per cent in 2022–23. Two large distributors in Bihar moved part of their business to a competitor. In return, finished-goods inventory fell from 52 days to 34, because the Haridwar plant was building far fewer variants.

By 2025–26 hybrids were 41 per cent of the inverter units we sold. More important, the hybrid became the start of a rooftop conversation. Thirty-four per cent of our rooftop leads now come from registered owners of Dhruvanta inverters, and those leads convert at 2.3 times the rate of others.

I would do two things differently. For the first fourteen months we sold the hybrid without a priced path to add panels, and many buyers never learnt that their inverter could take them; I would launch the upgrade kit with the inverter. And I would have met the Bihar distributors myself before the announcement, not after their complaints reached the managing director.

Quoting the full rooftop price before the subsidy

When the central government launched its residential rooftop programme in February 2024, demand changed within weeks. A household could receive a fixed central subsidy per kilowatt, tapering above two kilowatts and capped at three, paid into its own bank account after the distribution company had inspected the system and fitted a net meter. Many vendors began advertising prices after subsidy, and some offered to deduct the subsidy at the counter and wait for it themselves.

Our sales leadership and two state heads wanted us to do the same. Their argument was strong: a quote tens of thousands of rupees dearer than a rival's lost the sale at the counter. I disagreed for two reasons. The subsidy was the household's money, released only after steps neither we nor the household controlled, and the wait for inspection and net meters varied from weeks to several months across the areas we served. Advancing it would have tied up ₹50 crore to ₹70 crore of working capital. Worse, households told a net price would blame us for every delay.

We quoted the full price, with the expected subsidy and its conditions printed beneath it, and offered a bridge loan through our lending partners to households that needed one. We built a file tracker into the app covering nine steps from portal registration to subsidy credit, and opened liaison desks staffed by technicians who had worked on meter testing.

The cost came first. From April to June 2024 our lead-to-order conversion was 14 per cent against a plan of 19, and I estimate we lost about ₹60 crore of orders in the first two quarters. Our Uttar Pradesh state head left for a competitor that advertised net prices.

By March 2026, 51,800 households had received the subsidy for Dhruvanta systems. The median time from installation to subsidy credit fell from 97 days in mid-2024 to 38 days in the January to March 2026 quarter, and order cancellations fell from 12 per cent to 4 per cent. Days to net meter and days to subsidy now appear in every quarterly board pack.

I would have opened the liaison desks before launch rather than after. One discom took more than 110 days to fit net meters during a meter shortage in 2024, and our customers there waited longest because we learnt about the shortage from them.

Cutting installer partners from 410 to 160 to fix the roof

Our rooftop business grew faster than our ability to install it well. Between 2021 and mid-2023 we signed 410 installer partners across eleven states, most of them small electrical contractors recommended by our retailers, and paid them on commissioning. By June 2023 the evidence was uncomfortable. Systems installed in 2022–23 were generating 38 service tickets for every 100 in their first year. The largest causes were water leaking through roof anchors on concrete roofs, poor earthing, undersized DC cable and inverters tripping on high evening voltage at the ends of rural feeders. The monsoon of 2023 brought 2,900 leakage complaints.

I proposed three changes: keep only partners who could pass a new certification, move installation standards and partner management from sales into the product organisation, and pay partners a share of each fee only after ninety days without a quality fault. Our national sales head and our chief financial officer both disagreed with the timing. The plan for 2023–24 assumed 26,000 installations, and the market was about to grow. They proposed certifying partners over eighteen months while continuing to sell. I did not believe we could retrain 410 contractors while they worked at full pace, and the leaks were damaging our retailers' standing in their own towns.

The managing director backed the change, on condition that I report progress to the board every month. We certified 160 partners, opened training academies in Lucknow and Coimbatore, standardised a kit with pre-crimped cables and ballasted mounts for suitable flat roofs, and paused new orders in three districts for a quarter. Our order backlog stretched from three weeks to seven, and about ₹95 crore of orders slipped into the following year or were cancelled. We installed 19,500 systems in 2023–24, not 26,000.

First-year service tickets fell to 11 per 100 systems for installations made in 2025–26. Leakage complaints in the 2025 monsoon numbered 310, on an installed base three times larger. Certified partners now earn more per installation than before, because rework has almost disappeared.

I would have certified before we scaled rather than after, and tied partner pay to ninety-day quality from the first contract. Both lessons were plain from my years managing technicians at Taravani, and I let the pace of growth persuade me otherwise in 2021.

Declining a pay-per-unit model for small factories

In early 2024 our strategy team, backed by a renewable-energy fund that had approached the company, proposed a pay-per-unit model for small manufacturers. Dhruvanta would own rooftop systems on their sheds and sell them the electricity at a fixed rate for fifteen years. The model removed the need for a loan, which many small factory owners avoid, and the fund offered to finance the assets. Our managing director found it attractive, and at first so did I.

Before recommending it, I ran a pilot with 64 factories in Tiruppur, Ludhiana and Rajkot: knitwear units, cycle-component makers and small engineering workshops. The pilot showed three problems. Forty-one per cent of the units operated from rented sheds on leases shorter than five years, so the system's owner would depend on a landlord neither party controlled. Payments from units with seasonal order books arrived late in nine months of twelve. And roof strength varied so widely that a third of the sheds needed structural work before installation. A fifteen-year contract would have made Dhruvanta a creditor to hundreds of small businesses whose risks we were not equipped to judge.

I recommended a different product: a system the factory owns, financed by our lending partners, with the instalment held within 80 per cent of the expected monthly saving on the factory's current tariff. For tenants we added a relocation option, under which the system can be moved once to a new shed within five years for a fixed fee. The managing director disagreed and wanted to test the pay-per-unit model at scale; the board approved my product and kept the other option open for review. The trade-off was explicit. Tenants without a landlord's consent, about 35 per cent of interested factories, could use neither product.

We launched in October 2024. In 2025–26 we installed 1,400 factory systems with an average size of 22 kilowatts, worth ₹154 crore. Our lending partner reports 2.1 per cent of those loans more than thirty days overdue, and 46 owners have used the relocation option.

I would do one thing differently. The pilot was too small to rule out the pay-per-unit model for owner-occupied factories, and I presented it to the board as though it did. I would now test both products, with landlord consent agreements, before calling the question closed.