NRS Insights Blog · Topic

Data Quality & Governance

Definitions, item masters, suppression, restatement and the practices that keep a number trustworthy.

All topics

Data Quality & Governance

Newest first

Data Quality & Governance
Sell-in vs. sell-through: reconciling shipments and scans

Sell in vs sell through: why shipments and scan data never match exactly, how to reconcile them, and what the gap says about inventory.

Aug 26, 20264 min read
Data Quality & Governance
Data dictionaries: agreeing on what a number means

A data dictionary defines every field and metric so teams mean the same thing by the same number. What to include and how to keep it current.

Aug 14, 20264 min read
Data Quality & Governance
Data latency: how fresh does retail data need to be?

Data latency is the gap between a sale and when you can analyze it. How fresh retail data needs to be depends on the decision it supports.

Aug 12, 20264 min read
Data Quality & Governance
Data quality checks every retail analyst should run

Data quality checks to run before any retail analysis: completeness, duplicates, negative units, price sanity and store counts, with fixes.

Aug 10, 20264 min read
Data Quality & Governance
Outliers in sales data: keep, cap, or flag?

Outliers in sales data can be errors or real events. A practical method for deciding when to keep, cap or flag an unusual store-day figure.

Jun 19, 20264 min read
Data Quality & Governance
Knowing when a number is good enough

Precision has a cost and a point of diminishing return. How to decide when an estimate is sufficient for the decision.

Jun 6, 20264 min read
Data Quality & Governance
Working with a research agency

The agency relationship determines the quality of the output more than the brief does. What good looks like on both sides.

Jun 1, 20265 min read
Data Quality & Governance
Confidentiality and competitive sensitivity in shared analysis

Collaboration between a brand and a retailer involves information that matters to third parties. Where the lines sit and why.

May 30, 20265 min read
Data Quality & Governance
Sharing analysis with retailers and wholesalers

What you send to a trading partner becomes their view of you. What is appropriate to share, and what should stay inside.

May 29, 20264 min read
Data Quality & Governance
Turning a finding into a recommendation

Analysis stops at what is true; a recommendation says what to do. The step between them has its own discipline.

May 22, 20265 min read
Data Quality & Governance
Reproducibility in commercial analytics

An analysis that cannot be re-run is a claim nobody can check. What reproducibility requires in a commercial setting.

May 9, 20265 min read
Data Quality & Governance
Writing an analysis brief that gets answered

Most disappointing analyses were badly specified rather than badly executed. What a brief needs to contain before work starts.

May 6, 20264 min read
Data Quality & Governance
Attribute tagging and why taxonomies drift

Tags let a brand analyse the market its own way. They also decay, and a decayed taxonomy answers questions incorrectly.

Apr 1, 20264 min read
Data Quality & Governance
Item master hygiene: the unglamorous foundation

Attributes decide what every report can say. Why item masters drift, what breaks, and how to keep one usable.

Mar 30, 20265 min read
Data Quality & Governance
Reconciling two sources that disagree

When shipment data and scan data tell different stories, the gap is usually structural. A method for finding where it comes from.

Mar 27, 20265 min read
Data Quality & Governance
Data latency and what a fresher number costs

Faster data costs more and is worth less than people assume. How to decide what latency a decision actually needs.

Mar 25, 20265 min read
Data Quality & Governance
Restatements: when history changes

Published figures get revised. Why history moves, what a restatement policy should say, and how to keep an analysis reproducible.

Mar 23, 20264 min read
Data Quality & Governance
Suppression, minimum cell sizes, and small numbers

Small cells are unreliable and sometimes identifying. Why providers suppress them and what a buyer should do with a suppressed figure.

Mar 21, 20265 min read
Data Quality & Governance
Outlier handling: rules before results

Every dataset contains values that do not belong. Deciding what to do with them before seeing the answer is what keeps the answer honest.

Mar 20, 20265 min read
Data Quality & Governance
Weighting and projection: turning a sample into a market

A panel measures its own stores. Projecting to a market is a modelling step with assumptions that decide the answer.

Mar 16, 20264 min read
Data Quality & Governance
Store churn, mix effects, and the number that is not growth

A panel that gains stores reports growth that is not market growth. How churn distorts totals and what a like-for-like basis fixes.

Mar 14, 20265 min read
Data Quality & Governance
Data dictionaries: why definitions decide the answer

Two providers can report different numbers for the same market and both be right. The difference is usually in the definitions, not the data.

Feb 23, 20265 min read

Data Quality & Governance

All topics

Data Quality & Governance

Sell-in vs. sell-through: reconciling shipments and scans

Data dictionaries: agreeing on what a number means

Data latency: how fresh does retail data need to be?

Data quality checks every retail analyst should run

Outliers in sales data: keep, cap, or flag?

Knowing when a number is good enough

Working with a research agency

Confidentiality and competitive sensitivity in shared analysis

Sharing analysis with retailers and wholesalers

Turning a finding into a recommendation

Reproducibility in commercial analytics

Writing an analysis brief that gets answered

Attribute tagging and why taxonomies drift

Item master hygiene: the unglamorous foundation

Reconciling two sources that disagree

Data latency and what a fresher number costs

Restatements: when history changes

Suppression, minimum cell sizes, and small numbers

Outlier handling: rules before results

Weighting and projection: turning a sample into a market

Store churn, mix effects, and the number that is not growth

Data dictionaries: why definitions decide the answer