Case study · Retail analytics

From 500 Walmart store doors and a spreadsheet to a live intelligence portal.

A CPG brand newly live at Walmart across 6 states had no store-level visibility, manual weekly reporting, and no way to know when a store went dark. We onboarded them onto the Lucerna portal. The spreadsheet is gone.

Client
CPG brand, anonymous
Retailer
Walmart in-store
Service
Full platform deployment
Scale
514 store locations
The problem

Seven questions they couldn't answer in real time.

The brand had Retail Link access and weekly POS data from ABI Retail Hub. What they didn't have was any structure around it. The raw export — rows of store-week-SKU combinations with no visualization and no way to ask follow-up questions — wasn't an insight. Insights happened quarterly, in spreadsheets, after the window to act had already closed.

01

Which stores had never sold a single unit?

They were counting ghost doors in their ACV number — stores in the distribution estate with zero lifetime sales — without knowing they existed.

02

Which stores used to sell and had gone silent?

Lost doors were invisible until a quarterly broker review revealed them — often months after the store had gone dark and recovery was no longer possible.

03

Was a velocity dip a demand problem or an OOS problem?

Without inventory data layered against POS, they couldn't tell the difference — and the wrong response to each is completely different. Call the 3PL or call the field rep?

04

Which SKU-store combinations had quietly dropped off?

A SKU could go dark in a store for 10 weeks before anyone noticed. The store kept selling other items — so nothing looked wrong at the store level.

05

Where should the broker team focus this week?

Field rep time was allocated based on gut feel and relationships, not data. There was no prioritized action list — just raw signals buried in the weekly export.

06

Were new doors ramping as expected?

When new stores activated, there was no way to track opening velocity or flag early underperformers before they became chronic problems.

The platform

Twelve dashboard panels. One view of the business.

We deployed a purpose-built retail intelligence portal processing Walmart POS and inventory snapshot data into a layered analytics environment. Every panel was designed around the specific questions their team needed to answer week over week. The period selector — L4 / L12 / L24 / L52 / YTD / individual weeks — controls every panel simultaneously.

01

KPI Grid

Walmart POS sales, week-over-week delta, and active selling store count — surfaced in the first 10 seconds on the page, with directional indicators so the team answers the most important question before anything else loads.

BeforeThe team didn't know current-week numbers until they built a pivot table, often days after the upload.
02

Weekly Sales Trend

A dual-axis chart pairing POS revenue against units sold across every week in the period. Clicking any bar opens a drill-down showing top performers alongside every no-sale store, root-cause-tagged, with a consecutive-silence counter.

BeforeTrend data was a static table. No drill-down, no week-by-week story with store context.
03

Inventory Health

When inventory snapshot files are present, a stacked bar chart shows OOS, thin-stock, and healthy stores per week. A KPI strip surfaces OOS rate, pipeline coverage %, and — most urgently — OOS combos with no recovery pipeline in flight.

BeforeInventory and POS data lived in separate exports. Connecting them was a multi-step manual process that almost never happened before quarterly review.
04

Store Tiers

The entire store universe segmented into Tier A (top 20%), B (middle 60%), and C (bottom 20%) by average weekly revenue. An insight line quantifies the concentration ratio: how much more an A store contributes versus a C store, per door per week.

BeforeStore tiers were never formally calculated. "Good" and "bad" stores were known anecdotally by the broker — not systematically, and not with numbers.
05

SKU Weekly Trend

Each SKU tracked individually, normalized by its own distribution footprint — not blended into a total that masks per-SKU dynamics. Color-coded week-by-week UxSxW table with a period-average column.

BeforeSKU-level velocity was blended into a single number. A strong SKU could mask a weak one for months.
06

SKU Pareto Analysis

Which stores drive 80% of each SKU's cumulative volume over the trailing four weeks. An "All-SKU" tab surfaces stores in the Pareto-80 tier for every SKU simultaneously — the brand's most critical doors for broker prioritization.

BeforePareto segmentation was never applied to the store base. Resource allocation for field visits had no data foundation.
The reports suite

Twelve specialist reports. One for every category of risk.

Beyond the main dashboard, twelve purpose-built reports give the team and their broker the tools to diagnose and act on specific signals — each week, not each quarter.

Action Items

Aggregates signals from every other report into a single prioritized list — Urgent, High, and Medium — tagged by signal type and store. The hub the broker opens before every call.

Lost Doors

Stores that were active sellers but have gone completely silent, ranked by weeks dark. Shows last sale date and peak weekly velocity before the store went cold.

Never-Sold Stores

Every store in the distribution estate with zero lifetime sales — doors counted in ACV that have never actually moved a unit.

OOS → No Recovery

Contiguous out-of-stock periods with no pipeline coverage in flight — tracked by streak length, resolved vs. unresolved, and average recovery time.

Anomalous WoW Drops

Store-weeks where velocity collapsed 40–70% versus the prior week, each root-caused: OOS-Driven, Demand-Driven, or Unknown. Filters by cause type and minimum baseline to suppress noise.

SKU Dropout

Stores that used to carry a specific SKU and have gone silent on that item — while potentially still selling others. The SKU-level version of Lost Doors.

New Door Gains

Stores that recorded their first-ever sale within a configurable lookback window. Shows opening velocity and weeks since activation — flags slow starters early.

Streak Leaders

Every store ranked by longest consecutive no-sale streak across all history. Separates a one-time blip from a chronic pattern of going dark.

Chronic Underperformers

Stores that have fallen in the bottom velocity percentile for a configurable number of consecutive weeks — not just stores that had a bad stretch.

SKU Distribution

Full store-by-store distribution picture for any SKU: active carriers, stores that have never carried it, and lapsed stores. Geographic density map included.

Store Lookup

Enter any store number and see its complete sales and inventory history — every week it has ever appeared in an upload, with no date filter. Deep-linked from every panel.

Upload Gap Audit

Identifies stores with missing weeks between first and last upload date — so analyses are never run on incomplete histories without knowing it.

The vocabulary

A shared language for the whole team.

Beyond individual panels, the platform introduced a standard set of metrics that the brand, their broker, and their leadership now share. When everyone uses the same definitions, the right conversations happen faster.

Metric
What it means
Why it matters
UxSxW
Units per Store per Week, normalized by each SKU's own distribution count
Apples-to-apples velocity comparison across SKUs and time periods
No-Sale Rate
% of distribution universe selling zero units in a given week
Distinguishes selling velocity from distribution presence
OOS-Driven vs. Demand-Driven
Root cause classification of every no-sale event
Tells you whether to call your DC or run a promotion
Pipeline Coverage
% of OOS stores with a reorder currently in transit
Distinguishes recoverable OOS from unresolved stockouts
Consistent Sellers %
Stores that sold every single week of the selected period
The healthiest distribution signal — frequency, not just average
Streak
Consecutive weeks of no sales in a single store
Makes chronic silence visible at a glance
Pareto-80 Store
A store in the top-80% volume tier for a given SKU
Prioritization framework for field rep and broker attention
Tier A / B / C
Store segments by revenue percentile — top 20% / mid 60% / bottom 20%
Quick-sort for resource allocation across hundreds of doors

The question was never whether the data existed. It was whether anyone could see it clearly enough to act on it.

The outcome

Week-level decisions. Not quarter-level surprises.

The portal didn't just save time — it changed how the team and their broker operate. Store-level intelligence that used to take hours to assemble is now available the moment anyone opens a browser.

Distribution leakage surfaced on the first upload. The Never-Sold report identified a meaningful number of doors the brand had been counting in their ACV — stores in the estate with zero lifetime sales they didn't know existed.
OOS streaks quantified for the first time. The OOS → No Recovery report revealed multi-week stockout events that had resolved without anyone knowing — and others still unresolved — distinguishing supply-chain urgency from historical context.
Broker prioritization transformed. The Action Items report replaced the pre-meeting prep spreadsheet. The broker team now opens the portal before every touch, pulling a prioritized list organized by signal type and urgency.
Root-cause clarity changed who gets the call. The team can now answer the question a VP always asks at quarterly review — "Is this a demand problem or a supply problem?" — with a chart rather than a guess.
SKU dropout caught weeks before quarterly review. A SKU that had quietly dropped in a cohort of stores was surfaced by the SKU Dropout report with enough lead time to investigate and recover before it compounded.
Distribution Build map used in investor meetings. The animated week-by-week expansion across 500+ store dots became a visual asset the brand uses to tell their growth story — in board decks, buyer meetings, and pitch calls.
Ready to see it?

Your Retail Link data. Your portal. This week.

Request a demo and we'll walk you through the platform using your own Walmart footprint — store count, SKUs, current distribution. No slides, no sandbox data. Just your actual numbers in Lucerna.

Get started

Selling at Walmart? You're already halfway there.

If you have a Retail Link login, you have everything you need to go live in Lucerna this week. Request a demo and we'll show you exactly what your stores look like inside the platform.

Request a demo →