Creating Custom Characteristics · Nielsen

Overview

RoleUser Experience Architect

UsersNielsen data analysts

ToolsUser flows · Wireframes · InVision

StatusPrototyped at mid-fidelity · Not officially released

ProblemThe categories in the database didn't match the segments analysts needed to report on.

OutcomeA guided workflow that turns filtering logic into reusable, reportable business segments.

One step in Nielsen's end-to-end reporting system. See Reporting Clarity →

01Problem

When the data doesn't match the question.

An analyst might need to report on "Alcohol Free," even though no such value existed in the dataset. They needed a way to say:

Find the records that meet these conditions, group them, name the group, and let me report on it.

02Insight

Build on a word analysts already used.

At Nielsen, "characteristic" was established vocabulary. Instead of inventing a new model, I built the experience around it, so the capability felt like part of the platform, not a new tool to learn.

03Workflow

A progressive sequence, not a wall of configuration.

Select data → Create characteristic → Name it → Create values → Define rules → Handle remaining data → Report

Defining a value inside a Custom Beer Segment: a list of characteristic values on the left, and a rule builder on the right composing World Beer from category, type of beer, and country conditions. — Defining "World Beer" from category, type, and country rules: filtering logic in business terms.

04Key decisions

Designed for experts doing it many times.

Built for many values.After each rule, the next action was ready, so repetitive setup kept its momentum.
Give "everything else" a home.A catch-all value made unmatched records predictable in reporting.
Prove it at mid-fidelity.The InVision prototype tested the interaction model before visual design.

A single characteristic holding many user-built values: Alcohol Free, World Beer, Craft Beer, Standard Lager, Premium Lager, Super Strength Lager, and Price Fighter Lager, with an All Other row for unmatched records. — One characteristic, many reusable values. "All Other" catches everything outside the rules.

The larger data-selection experience: a New Extract screen with a prompt rail (dataset, product, market, period, fact), a product dimension list, a characteristics column, and a live summary panel. — The finished characteristic returns to the larger data-selection experience for reporting.

05Outcome

A technically complex operation became a guided workflow.

Define → Filter → Group → Name → Report

Analysts could organize data around their questions instead of predefined categories, without modifying the underlying database. The workflow was evaluated at mid-fidelity; the feature was not officially released.