Mary Varnelvi · Chief Data Officer · India
I could help directors examine the claim a data product asks the enterprise or its customers to rely on. The relevant question may concern who was counted, how a new result can be compared with an earlier one or which population the study actually observed. I would begin by understanding the receiving company’s product and the decisions it supports, rather than applying a preferred survey method to every problem. My experience includes an incorrect pooled-person claim that I approved, so I know how a familiar record can become an unsupported conclusion through management judgement. A board contribution should make that movement easier to question before it affects a material commitment. Directors need to understand the remaining uncertainty and the assurance available, while research specialists retain responsibility for design, analysis and professional interpretation.
Presented to the Board or a committee
Board research-product and commercial-evidence reviews
April 2022–present
Presents data-definition, comparability and coverage issues through the chief executive; holds no statutory seat.
Reported to a committee
Research-product correction review through the CEO
April 2017–March 2022
Reported the pooled-interview correction and the limits of the distinct-respondent statement.
| Capacity | Committees |
|---|---|
| Ready to join | Data-dependent strategy and customer-evidence discussions · Research-product risk oversight where constituted |
I would explore an initial non-executive contribution with a chair who wants research assumptions examined without making the board a parallel analytics team. My present experience is executive presentation and reporting, not statutory directorship. Further governance learning and appointment-specific assessment would come before acceptance. I could participate in data-product, customer-evidence or strategy-risk discussions alongside established financial, legal and governance skills, without presenting myself as ready to chair a statutory committee. The chair should identify which recurring data claim directors find hard to challenge and what assurance management currently supplies. I would be useful where a narrower but supportable conclusion can still inform action, and where colleagues can distinguish the need for better evidence from a desire for a more persuasive chart. The responsibility for professional methods and product delivery would remain with accountable management.
The selected file contained 2,400 interviews from 1,650 distinct respondents. I would ask directors how the product preserves that distinction when the larger count moves into a commercial account. Identifying distinct respondents did not establish population reach, representativeness or independence of observations. Oversight should make the intended inference explicit and ask whether management has evidence for it. The question is not whether every director can repeat the analysis, but whether the board understands what the number actually supports. My approval error shows why a familiar data source is not sufficient assurance for every conclusion drawn from it. The correction needs to reach the place where a client or executive makes a decision, not remain only in a working-file note that the persuasive headline contradicts.
After eight telephone waves, the web wave could have appeared as a directly comparable ninth point. The 120-person paired comparison showed response differences, but did not establish a universal cause or adjustment. I would ask the board how the product communicates the break in method and what further work would support a particular comparison. Directors should not decide the statistical design themselves. They should understand whether the commercial proposition is relying on continuity the research has not established. A separately labelled series can remain useful while uncertainty is examined. The contribution I could make is to distinguish a decision that needs comparability from one that can use the new wave on its own terms, with methodological colleagues responsible for the evidence supporting either interpretation.
I would trace one significant data-dependent decision from the research question through recruitment, collection and the product used by management or a customer. Through the chair’s agreed arrangements, I would meet methodology, data, commercial and risk colleagues to understand how a disputed claim is resolved. I would examine a correction alongside a successful product launch, because the organisation’s response to an attractive but unsupported inference matters. Twenty-two proposed annual days must cover learning, preparation, meetings and exceptional work as well as executive responsibilities. I would ask the chair whether my questions improve the board’s understanding or draw it into methodological decisions it should receive assurance on instead. At the first-year review, I would want a material discussion to state more clearly what the evidence counts, which comparison it supports and what uncertainty remains. I would also ask whether that clarity changed the actual decision or simply added a note no one used.
My role exposes private client questions, respondent arrangements, panel terms, research methods and unpublished results. I would disclose employer, trust, investor, client, panel-provider and adviser relationships before receiving another enterprise’s confidential evidence. The company considering an appointment must assess eligibility and independence in its own circumstances, and my employer must consent to an outside commitment that can be met. I would not use participant records or a client’s unpublished research to help another board gain a commercial advantage. Recusal might address an isolated overlap, but recurring exclusion from the main evidence agenda could make the role unsuitable. Capacity must include the possibility of a material correction in my executive work. I would decline a seat that expected access to private respondent files or treated my appointment as a substitute for commissioning research capable of answering its actual question.
Employer, researcher-trust, investor, client, panel-provider and adviser relationships require specific review. I hold no personal equity or trust interest, but do not infer independence from ownership status alone.
IICA Independent Directors’ Databank: Not yet registered
The governance preparation and any registration required for an intended appointment remain ahead of me. No assessment pass or exemption is represented in this fictional account.