| When | April 2021 |
| Customer | A large Indian brand, recently onboarded at the time |
| The defect | A bought-in batch of wall fan oscillation cords corroding, impeding oscillation |
| How we found it | During lot inspection for a different customer |
| Units already shipped | ~1,000, across three warehouses in three parts of the country |
| What we did | Disclosed to the brand the same day; sent three teams of five to rework on site |
| Time to complete | Under 60 hours, working nights |
| Permanent fix | Spring cords replaced with nylon across the range — failure mode eliminated, not monitored |
| Lasting outcome | The brand’s QC head moved us from in-person audits to self-inspection |
April 2021. We did not find this on their line. We found it while inspecting lots for a different customer, and the fans in question had left our factory the week before.
The defect
A batch of the cords used to drive oscillation on wall fans was corroding. Corroded cord, impeded oscillation. Bought-in component, single batch, and the failure would not always show on the line — it would show in a customer’s home a few weeks later.
We fixed it immediately for the customer whose lot we were inspecting. That was straightforward: the stock was in front of us.
Then came the question that mattered. What about the roughly 1,000 fans already shipped to a large Indian brand the previous week — sitting in three warehouses, in three different parts of the country?
The arithmetic of saying nothing
Worth setting out honestly, because the incentives all ran one way.
A recall would delay supplies and cost the brand season sales, which is real damage to them and to us. On-site rework meant deploying teams across three states at our cost, up front, and asking a customer we had only recently started supplying to stop selling for three days in the middle of their season.
The alternative was available and cheap: do nothing, and absorb the rejections as they came. The failures would be intermittent and spread over months. They would arrive as ordinary warranty claims, handled mostly by the brand’s own service team, attributed to any number of causes. Nobody was going to trace an oscillation complaint in one city back to a cord batch at our plant. We would almost certainly never have been caught.
So this was not a case of having no choice. We had a cheaper choice, a quieter choice, and a choice that would have worked. Something about it didn’t sit right.
What we did
We called the brand’s procurement head the same day and told them what we had found, what was likely already in their warehouses, and what we proposed to do about it.
Then we sent three teams of five people from our own factory to the three warehouse locations, and asked the brand to hold sales for three days while we reworked the stock in place.
Rework on site rather than recall was the right call for the brand, not for us. A recall is administratively simpler for a supplier — you take the stock back, you fix it at home, you redeliver. It also costs the brand weeks of a season they cannot get back. Going to the warehouses was harder for us and far cheaper for them.
Working nights
A warehouse cannot stop moving goods during the day, so our teams worked at night. Fifteen people, three cities, night shifts, reworking units one at a time and re-checking oscillation on each.
The rework was completed in under 60 hours — inside the three-day window we had asked for — and the stock was certified fit for sale.
We took the flak
This is the part usually left out of a supplier’s case study, so it is worth stating plainly: the brand was not delighted with us. We took criticism from several departments, and we deserved a fair amount of it. A defect had reached their warehouses, we had asked them to stop selling in peak season, and we had created work for people who had not caused the problem.
We would make the same call again. The alternative was letting their customers find our defect for them.
The permanent fix
Reworking 1,000 fans deals with the batch. It does nothing about the next batch.
The oscillation cords were steel spring cords, and steel in a humid supply chain will eventually find a way to rust. Adding an incoming inspection step for corrosion would have been the obvious response, and the weaker one — an inspection catches a failure that has already happened, and it depends on somebody catching it every single time.
We changed the component instead. The range moved from spring cords to nylon, which does not corrode. The failure mode is not controlled or monitored. It is gone.
That distinction matters more than it sounds. An added check is a control, and controls decay — people move on, volumes spike, a sample gets waved through in March. A design change that removes the cause does not decay. It is the difference between managing a risk and no longer having one.
What it bought
Some months later the brand’s QC head moved us from in-person audits to self-inspection. We certify our own lots; they accept the certification.
That is close to the largest thing a brand can hand a supplier, and it was not something we asked for. In-person auditing exists for one reason: the brand does not yet know whether a supplier will report honestly on their own output. Moving to self-inspection means that question has been answered and their own verification step has been removed. It saves them travel, manpower and time at our gate. It also puts the entire burden of being right on us, permanently.
We did not earn that by passing inspections. We earned it by reporting a failure the brand had no way of detecting.
What a brand is actually outsourcing
Worth being precise about what changes hands when a brand appoints an OEM, because it is considerably more than a production schedule.
Our name is not on the fan. Theirs is. If a unit fails in someone’s home, that customer does not know we exist — they form a view about the brand printed on the carton. The brand absorbs damage for a defect they did not cause and had no way to detect.
That asymmetry is the central problem with outsourced manufacturing and it is better said out loud than papered over. The party doing the building carries none of the brand risk. A manufacturer’s exposure ends at the purchase order; a brand’s runs for as long as their name does. It is precisely why the quiet option existed in the case above — the cost of saying nothing would have landed on somebody else’s reputation, not ours.
So a brand outsourcing to us is not handing over capacity. They are handing over the most valuable thing they own — what people think of when they see their name — and trusting us to treat it the way we treat ours. We take that literally. The name on the carton is ours to protect for as long as the goods are in our hands, and our own name is worth precisely as much as the last brand we did not let down.
Since it would be dishonest to leave this out: profitability matters to us, and a supplier who is not profitable is a supplier who will not be here in ten years, which is its own risk to a brand. But profit is what we optimise inside the constraint, never by loosening it. Trust is the constraint. That one is not negotiable, and those 1,000 fans are what it looks like when holding to it costs something.
Why this is on our website
Every manufacturer has batch defects. Anyone who tells you otherwise is either new or not being straight with you. Components come in from outside, batches vary, and a proportion of problems will not reveal themselves until after the goods have shipped.
So “we have no defects” is not a claim worth making, and not one a serious buyer believes. The question that actually matters when you choose an OEM is narrower and much harder to get an honest answer to: what does this supplier do in the first hour after they find something, when nobody else knows yet and saying nothing is cheaper?
This is our answer, with the numbers attached. We would rather you judged us on it than on an assurance.
If you want to talk about how we would handle your programme — including what happens when something goes wrong — start here, or call +91 9177 068 845. More on how the plant is set up is on the Hyderabad plant page.