The raw material was a wall of numbers
The Weekly Business Review deck was comprehensive by design — agenda items covering seller performance, SBU-level GMV bridges, inbound and outbound operations, forecasting, refunds, reverse logistics, and open action items, each with its own dense table. That's exactly what an operating team needs. It is not what a Senior Leadership Team distribution list needs to read cover-to-cover before Monday's call.
| SBU | GMV TY | GMV Plan | vs Plan ($) | vs Plan (%) |
|---|---|---|---|---|
| Apparel | $6.8M | $6.6M | +$0.2M | +3.0% |
| Consumables | $15.1M | $16.2M | −$1.1M | −6.8% |
| Hardlines | $17.9M | $20.4M | −$2.5M | −12.3% |
| Home | $19.0M | $21.1M | −$2.1M | −10.0% |
| Food & Other | $3.6M | $3.1M | +$0.5M | +16.1% |
| Grand Total | $76.8M | $81.0M | −$4.2M | −5.2% |
- Miss driven by Hardlines and Home — largest sub-categories: Furniture, Patio & Garden, Home Management
- Partly offset by beats in Impulse Merchandise and Food
Multiply that one table by roughly thirty, across seller segments, fulfillment centers, and shipping lanes, and the actual weekly artifact was somewhere between an Excel workbook and a small book. Somebody had to read all of it, every week, and decide what actually mattered.
Every metric, shown next to its goal and its trend — always
This is the actual raw material the narrative was written from: an Executive Summary scorecard where no metric was ever allowed to stand alone. Every single tile carried the same three things — the actual value, the goal it was measured against with a color-coded gap, and the year-over-year comparison. That consistency is what made it possible to read forty-plus metrics in one sitting and know, at a glance, exactly which ones needed a sentence in the narrative and which ones didn't.
Every one of those tiles, week after week, is where the narrative's "WoW / YoY / vs. Plan" framing actually came from — the scorecard already carried two of the three lenses natively; the manual craft was picking which dozen tiles out of forty deserved a sentence, and writing the third lens (the trend story) that the tile grid alone couldn't tell.
Every number, framed three ways
The manual work wasn't just summarizing — it was consistently answering the same three questions for every metric that mattered, so leadership never had to ask "compared to what?":
Is this getting better or worse right now, independent of the calendar?
How does this compare to the same week last year — is the business fundamentally healthier?
Regardless of trend, are we going to hit the number the org committed to?
Same headline metrics, all three lenses at once (illustrative values)
A metric can look great WoW, terrible vs. plan, and fantastic YoY, all at the same time — and each of those is a different conversation with leadership. Collapsing all three into one clean bar per metric was the difference between a leader glancing at a screen for 10 seconds and actually walking away informed.
What actually landed in the leader's inbox
Every week, this became one email to the full Senior Leadership Team distribution — short enough to read on a phone before a meeting started, with a runway of pointed questions at the bottom instead of just a data dump.
Week 22 Summary
Week 22 is in the books, and GMV strengthened roughly $1.8M versus the prior week, ending at $76.8M (−5.2% to plan, +24.3% YoY). Penetration missed plan but beat the year-ago mark, landing at 41.4% (−22bps vs. plan, +118bps YoY). Year-to-date, WFS is running roughly −$91M behind the annual goal.
Sellers & SKUs Sellers grew at a healthy clip, adding 402 last week for a total of 43,850 live sellers, 6,120 of them new this fiscal year. Seller retention (adjusted) landed almost exactly on plan at 89.6%, while overall retention missed by −7.4pp at 81.7%. In-Demand SKU coverage is the flag this week at 641,900 (−24.8% to plan).
Outbound On-time delivery softened slightly WoW to 92.95% (−1.4pp to plan, −1.9pp YoY), driven mainly by fulfillment-center processing delays rather than carrier performance, which actually improved WoW.
Inventory In-stock rate recovered WoW but still trails plan at 73.1% (−0.6pp to plan, −3.9pp YoY). Refund rate came in at 8.2%, missing plan by 0.9pp, while restock rate continues to outperform at 15.9%.
Key Callouts This Week
- GMV and Penetration continue to lag plan despite a comfortable YoY beat — what closes the gap to plan?
- In-Demand SKU coverage is well behind goal — what's the remediation plan and timeline?
- Overall seller retention is meaningfully behind goal — do we understand the root cause?
- OTD degraded again WoW — what stops it from falling further next week?
- Refund rate is trending the wrong way — what corrective actions are already in flight?
Bringing in the outside world
The part that turned this from a status update into an actual narrative was a closing "Macro Economy" section — a handful of curated, e-commerce-relevant industry items scanned and added every week, so leadership could tell whether a number moved because of something WFS controlled, or because of something happening across the entire industry. This is where "craft a story from data" stopped being about Walmart's own numbers and started being about competitiveness.
Consumer-protection regulators in a major market issued formal notices to a cluster of e-commerce platforms over deceptive "dark pattern" checkout flows, alongside new consumer-education tools.
Shifting cross-border tariff exposure was pushing several marketplace sellers to diversify manufacturing out of a single low-cost country and toward alternate regions.
A well-funded robotics startup began piloting sidewalk delivery robots capable of navigating stairs and uneven terrain — a long-standing "last 50 feet" problem for every fulfillment network.
A major apparel brand that had sat out a rival marketplace for several years announced it was returning to sell there directly — a signal worth watching on where big brands see growth.
None of these items came from an internal dashboard. They came from actively reading trade press every week and asking, deliberately, "does this change how we should interpret our own numbers?" That's the piece that's easy to skip under deadline pressure, and the piece leadership actually remembered.
Same data, fundamentally different outcome
| Raw scorecard export (~30 pages) | Weekly narrative email | |
|---|---|---|
| Time for a leader to get the point | 10–20 minutes of scrolling, if they open it at all | Under 2 minutes, readable on a phone |
| Context per metric | Numbers in isolation, no framing | WoW, YoY, and vs. plan, every time |
| External/competitive context | None — the deck only knows about WFS | A weekly industry scan, tied back to the numbers |
| What drives the meeting agenda | Whoever speaks up first | A short list of pointed callouts, set in advance |
| Who actually reads it | The analysts who built it | The entire Senior Leadership distribution |
What this actually delivered, every week
This was, in hindsight, exactly the kind of work that's now a natural fit for AI assistance — structured synthesis, consistent framing, and an external research scan, done on a strict weekly cadence. At the time, it was a standing habit: read everything, decide what mattered, and always come back with the "so what," not just the "what."