Quota Manager

n-Infinite deliberately generates more synthetic respondents than you'll ultimately need — for example, generating 2,000 to deliver a final sample of 200. Quota Manager selects exactly the right subset of that larger pool to match your required sample structure — by age, gender, region, or any combination of fields.

Open it from a finished project: Download Merged → Quota Manager (or the DATASETS menu in the project detail). The flow has three steps: Source → Define Quotas → Review & Download.

Step 1 — Source: three ways to define targets

1. Manual — pick fields from the project and set quotas by number or percentage.
2. Population Pyramid — use a country's age/gender population pyramid as the quota target.
3. Upload Quota File — upload an Excel/CSV quota sheet and let AI propose a mapping to the project's fields. The proposed mapping is always shown for review and adjustment before anything is applied.

Step 2 — Define quotas

Quota Settings:

Setting	Meaning
Mode	Targets as Percentage or counts
Answers	The size of the final sample to deliver
Interlocked	Cross-tabulated targets (e.g. a specific age-by-gender cell) instead of independent per-field targets
Priority	Which rows fill the quotas first — e.g. Real rows first, then Generated
Guarantee the full count	When quota cells can't be filled exactly, top up with the best-fitting leftover rows so the output still reaches the requested number of answers
Include all real cases	Keeps every real case in the output even if that overshoots the requested answers or skews the quota distribution

Dimensions and buckets. Add a dimension per quota field; each dimension holds buckets with a label, a type (Values or ranges), the categories they cover and their target. The sum indicator turns green at 100%:

Balance options (soft targets — the quotas always win): Keep profile fields balanced, and SmartQuotas to keep the selected output's answers distributed like the real panel (train), subject to the quotas.

Step 3 — Review & download

The fulfillment report, on screen, before you download anything:

• Headline cards: requested Answers, Achieved, rows from real data, and total Shortfall.
• Per-cell table: Target · From Real · From Generated · Achieved · Available in Pool · Shortfall · Relaxed — if a target can't be fully met from the available pool, the shortfall is reported explicitly rather than silently under-filled, and the Relaxed column shows where relaxation priorities were applied.

Selection is deterministic and reproducible. Download produces the filtered, ready-to-use dataset, as Excel (.xlsx) or SPSS (.sav).

A collapsible Post-processing (optional) card sits above the download button — "Repair, refine or deduplicate the selected rows before download." — with the same three stages described in Post-processing. All three start off here, and turning Remove duplicates on or off re-runs the quota preview, because removing duplicates changes the pool of rows competing for each cell.