A Stage Gate Process For New Products At A D2C Brand

Most small D2C brands in India do not run a new product process. They run a conviction. A founder likes an idea, a vendor sends a sample on WhatsApp, a launch date gets picked off the festive calendar, and the first honest cost sheet appears only after the stock has landed. By then the decision is already made. The money is spent, the SKU is live, and the only question left is how long the brand will quietly subsidise it.

A stage gate fixes the sequence. It does not add process for its own sake. It moves the expensive decision to the point where it is still cheap to make.

Why conviction launches fail late

A conviction launch is not wrong because the idea is bad. Plenty of them are good ideas. It fails because every piece of evidence arrives in the wrong order. Demand evidence comes after the tooling deposit. Real per unit cost comes after the minimum order quantity is committed. Return rate and breakage come after the first thousand orders ship. Each of those is a fact that could have killed or reshaped the product, and each one lands after the point of no return.

In the numbers it shows up as dead stock, as a catalogue with forty listings and four that matter, as working capital sitting in a warehouse in Bhiwandi. Founders describe it as a demand problem. It is usually a sequencing problem.

The five gates worth having

Five is enough for a brand under fifty crore. Any more and people start managing the process instead of the product.

Gate one: concept

Evidence needed: a written statement of who buys this, what they use instead today, and why they would switch. A rough target shelf price taken from what the category already charges, not from what you hope to charge. One page. If the switching reason cannot be written in a single sentence, the concept is not ready. Cost of getting this gate wrong: a few hours.

Gate two: feasibility

Evidence needed: at least two suppliers who say they can make it, an indicative cost range from each, a view on whether the format runs on an existing line or needs custom tooling, and a regulatory read. In India that regulatory read is not optional. Food, cosmetics, ayurvedic claims, electronics and toys each carry licensing, labelling and testing requirements that can add months to a calendar. Find that out here. Cost of getting it wrong: sample charges and a few weeks.

Gate three: costed sample

This is the gate that earns the whole process. Evidence needed: a physical sample you have used yourself, and a full cost sheet built up from the bill of materials, conversion, packaging, inbound freight and applicable duties, with the channel stack loaded on top. Marketplace or quick commerce commission, fulfilment, a returns provision and a realistic marketing cost per order all sit above the factory number.

Most products that should die, die here. The sample is fine. The economics are not. A brand that kills at gate three has spent sample money and six weeks. A brand that kills after production has spent the whole first buy and now owns the stock.

Gate four: pilot run

Evidence needed: a small quantity made on the real line with the real materials, sold to real customers through one channel. You are not testing appeal any more. You are testing consistency unit to unit, damage in transit, return reasons, review sentiment, and whether the cost sheet survived contact with reality.

Gate five: scale

Evidence needed: pilot results measured against thresholds you wrote down before the pilot ran, a supply plan that can hold the lead time through a festive peak, and a decision on what this product replaces in the range.

Who decides at each gate

One named person per gate. Not a committee. In a small brand it is usually the founder at gates one and five, and it should usually not be the founder at gate three. Give gate three to whoever owns the cost sheet, because the founder is the one carrying the conviction and conviction is precisely what gate three exists to test.

Write the pass criteria before the evidence is gathered. Thresholds set after the fact are not thresholds. They are justifications.

The discipline of killing at gate three

Killing a product feels like failure in a small team. It is the process working. A brand that never kills anything at gate three is not being rigorous, it is being polite.

Three habits make it easier. Name the kill rate you expect up front, so that a dead concept is an outcome rather than an embarrassment. Keep a one line record of what killed each concept, because the same objection resurfaces every year and you should not have to relitigate it from memory. And separate the person from the product in the language you use. The concept did not clear the gate. Nobody failed.

Running it without bureaucracy

The whole thing lives in one document per product. Five sections, one per gate, filled in as you go. A fortnightly review where every live concept gets thirty seconds and exactly one of three outcomes: pass, hold with a named missing piece, or kill. No fourth option. Concepts parked without a decision are the way this degrades back into conviction.

What you get is not a slower brand. It is a brand where the expensive decisions happen early, on paper, against samples that cost thousands rather than production runs that cost lakhs. The launches that do happen carry evidence behind them. And the catalogue stays a range, rather than an accumulation of things somebody once believed in.