Cybex Retail AI
BI Analytics · Building Block 3

BI Analytics Series

The daily routine over the data platform. Each series is a governed pack of matrix, list, query and dashboard templates for one retail function, with drill-through to the product and a user guide behind the Note button.

258 templates in productionDistribution Centre series: 22Guide behind every Note buttonA new report is a row, not a release
Retail BI analytics series: governed matrix, list, query and dashboard templates over the data platform
Matrix · List · Query · Dashboard

What a template is

A template is a saved, governed report definition: the dataset it reads, the rows and columns, the measures, the filters a user may change, the drill path, and the guide that explains it. Templates are data, not code, so a new report is a row, not a release.

Matrix

Cross-tab views such as store by class, size by colour, week by department, with totals, variances and heat shading.

  • Size matrix with drill to the product
  • Plan versus actual by month
  • DC flow by day and lane

List

Ranked and filtered lists: exceptions, reorder candidates, slow movers, cashier watch lists.

  • Sortable, exportable
  • Selection feeds the applications
  • Saved filters per role

Query

Parameterised extracts for analysts and for the AI insights engine.

  • One query per governed dataset
  • Same definitions as the matrix
  • Feeds CSV, Excel and the insight briefs

Dashboard

KPI tiles and rollups for a store manager or an executive.

  • Store dashboard with PWP tickets
  • Daily flash versus last year
  • Traffic, conversion and staffing

The series in production

Each series belongs to one retail function and reads the governed dataset for that function. New templates join a series; a new function gets a new series.

SeriesFunctionWhat the pack answersDrill path
Distribution Centre 22 templatesDC productivity, replenishment lanes, stock rebalancingReceipts, shipments and touches by day; labour by shift; units per hour and cost per shipped unit; lanes to each store; rebalancing proposals with five dialog parametersDC day to SKU movement to product
MerchandisingMerchandise plan, OTB, assortment, sizePlan versus actual by class and region; open-to-buy from open orders; size matrix by style and colour; white spaceClass to style to size, with the product card
Store OperationsSales, traffic, conversion, staffingDaily flash versus last year; footfall and conversion by store-week; staff hours against visits and salesChain to region to store to hour
Sales AuditRevenue integrity, loss preventionExceptions by type, store, terminal and cashier; cash-out variances; tender reconciliation; weekly margin bridgeStore day to ticket to line
Replenishment and PurchasingReorder, suggested PO, direct-to-storeWeeks of supply, restock and keep lines, suggested purchase orders, direct-to-store coverClass to style-colour to SKU
Private LabelBlanks, private label production, name dropsBlank requirement for 26 and 52 weeks, private label style performance, production schedule against blanks on handBlank to private label style to name drop

The Note button

Every template carries its own user guide: what the numbers mean, how the filters change them, and what to do with the answer. The guide is stored with the template and shown in the viewer, so the person who opens a report at seven in the morning has the explanation beside the numbers, not in a manual somewhere else.

  • Guides written for power users, not for the vendor
  • No database object names, functional language only
  • Same guide on test and on production
  • Updated by a row change, not a release

How the series fit the other blocks

Datasets

Templates read the governed datasets. If the dataset is right, every template is right.

Solutions

A list template's selection drives the applications: reorder, allocate, produce, ship.

AI Insights

The query templates are the extracts the insight briefs read.

Consulting

A series is scoped, built and handed over as one engagement stage.

Building block 3

See a series on your own data.

A Data Platform Foundation pilot ends with two series in production over your data. Bring the function that hurts most.