The regional arithmetic was sound; my combined account was not

I signed off the seventy per cent figure because the two regional percentages looked comparable. One region had ninety complete order-review records out of one hundred; the other had four hundred and fifty out of nine hundred. Averaging ninety and fifty gave seventy. Combining the underlying records gave five hundred and forty out of one thousand, or fifty-four per cent. The arithmetic in each region was sound. The way I allowed the regions to stand for the whole was not.

That 2019 report cost ₹19 lakh to repair. The correction was sixteen percentage points, not evidence that performance had suddenly fallen. I had approved the summary, and I wanted the explanation to retain that fact. A data leader cannot ask colleagues to examine their assumptions while treating her own aggregation as an anonymous dashboard problem. The incident changed the questions I asked about the population represented by a convenient executive number.

I grew up in Madurai and completed statistics at the fictional Tezqurmi Statistical College in May 1994. In July I joined Senvqari Information Services as a statistical analyst. The firm employed eighty people and recorded ₹8.4 crore revenue for FY ending March 1995. I worked in a six-person team without direct reports or spending authority. My early work taught me to describe what had actually been observed before trying to explain what it meant for a larger population.

In June 2000 I became analytics manager at the same firm. Eleven colleagues worked within ₹95 lakh annual expense, reporting through the analytics director. Senvqari had 210 employees and ₹36 crore revenue for FY ending March 2001. The role broadened my responsibility from producing an analysis to helping people use it. I learned that a technically correct calculation could still be attached to the wrong business question, and that explaining the limitation was only useful if we could identify a workable next step.

I joined Renvqathi in September 2007 as head of business analytics. The manufacturer employed 260 people and recorded ₹84 crore revenue for FY ending March 2008. I led ten colleagues with ₹1.7 crore annual expense under the CFO. Working inside one industrial enterprise made the consequences of a summary more tangible. A forecast informed a mix decision; a returns analysis could influence a product discussion; a timing measure could shape expectations about the work. I needed operating and technical colleagues to explain the context that the records alone did not contain.

Enterprise-data responsibility followed in April 2015. I led eighteen colleagues through three leads within ₹4.1 crore annual expense, reporting to the CEO. The company then employed 410 people and recorded ₹166 crore revenue for FY ending March 2016. The regional-average error occurred during this appointment. It showed that building a more coordinated data function did not itself prevent a poor inference. The meaning of the combined figure still required careful judgement from the people responsible for presenting it.

I became CDO in January 2021. Twenty-six colleagues now work through four leads within ₹6.9 crore annual expense. CEO Ishaan Mezqavi owns the executive direction; COO Devika Nevqurmi retains operations, and finance retains financial reporting. Our present ownership is 48% founders, 42% an industrial holding company and 10% employees. My 2% holding is included in that employee portion, and I have no statutory seat. I lead the data contribution without claiming authority over product acceptance or the decisions an automated recommendation might affect.

Recent cases have made the distinction between a number and its use more concrete. Hinge wear in sixty-three returned totes could not establish the failure rate of every shipment. Two opposing family forecast errors disappeared in a correct total. A predeployment model labelled all orders routine and achieved 99.2% accuracy while missing forty special-documentation cases. A timing review found that adding six-minute and eight-minute stage medians did not produce the nineteen-minute median of the matched complete jobs. Each example required a different correction to the question being answered.

I do not present these interventions as measured improvements in safety, service or profit. The reviewed accounts contain costs, held work and better-defined evidence. Their value is to show how I help colleagues decide what can be concluded now and what needs further investigation. A criticism that leaves the business unable to move is incomplete; an attractive number that invites an unsupported decision is worse.

I have no statutory board appointment. In this fictional career I registered with the IICA independent directors databank in June 2026, with assessment still to complete and no exemption claimed. Twenty-four days is a provisional allowance for a first outside role. I would bring thirty-two years of analytical and industrial experience, together with a clear memory of a summary I approved too easily. I would also need to learn the receiving institution beyond data and contribute to the board's collective judgement rather than become its informal analytics department.