Cost updates worked. The experience around them didn't.
- Role
- Product Manager
- Focus
- B2B supplier tooling, cost management, competitiveness
- Users
- Supplier partners, Merchandising, Operations
- Methods
- Discovery interviews, journey mapping, analytics, roadmap prioritization, cross-functional delivery
In 30 seconds
- The problem
- A company-wide competitiveness strategy was being enforced through a supplier workflow that gave suppliers almost no information or feedback — a 2,000-SKU CSV export, weak upload errors, and no warning before a product went offline.
- My vantage point
- I used and supported these tools in Operations, then worked with them in Merchandising, before inheriting Manage Cost as its Product Manager.
- What I owned
- Two-week discovery, the five-year roadmap and prioritization with Analytics, the supplier workflow and competitive context, CSV and error handling, cost-review logic, and delivery with Engineering, Merchandising, and Operations.
- Outcome
- 30% reduction in SKUs offline due to competitiveness, partners using Manage Cost daily, and Operations shifting from upload escalations to coaching.
30%
Reduction in SKUs offline due to competitiveness after the changes shipped.
- 2,000 → 10,000
- CSV export ceiling, with filtering before export
- Daily
- Partner use of Manage Cost after the changes shipped
- 15% / 30 days
- Increase review threshold, with a 30-day cumulative look-back
Background
01
I used these tools before I owned them.
I came into this from Operations and Merchandising, where I had personally used and supported the supplier tools. When I became the PM responsible for Manage Cost, I already knew which parts of it people quietly worked around.
I was given a two-week discovery phase to identify business, partner, and customer-experience gaps, and to come out the other side with a five-year roadmap.
What was broken
02
The workflow I inherited
Inherited cost update flow
- 01Supplier logs into Supplier Oasis and opens Manage Cost
- 02Scrolls a huge SKU list with poor filtering
- 03Exports a CSV — essentially SKU plus current cost/price
- 04Edits costs by hand, where a formatting change can break the file
- 05Uploads and gets little or no useful failure feedback
- 06Manually verifies whether anything actually changed
The export ceiling
Exports capped at 2,000 SKUs. Large partners could exceed 50,000. That gap turned a routine task into an escalation chain: Operations, then the PM, then Engineering, to produce multiple files.
The feedback gap
Uploads gave little useful failure information. A formatting change could break the file, and the only way to find out was to check the catalog afterward and guess.
Why it mattered
03
Suppliers found out their products were gone after they were gone.
Overstock increasingly removed uncompetitive SKUs from the site. Suppliers often had no visibility that a product was at risk or already offline, so Merchandising became the middleman. Weekly partner meetings sometimes only got through the top 10 SKUs.
The pain loop
- 01SKU becomes uncompetitive
- 02Supplier has no visibility that it is at risk
- 03SKU goes offline
- 04Selling opportunity is lost
- 05Supplier asks Merchandising what happened
- 06Investigation and explanation, then back to Manage Cost with an uncertain outcome
…and the loop starts again
Discovery vision
04
From a transaction form to a workspace.
- More filters and category or subcategory exports
- Bulk changes instead of one-at-a-time edits
- Sorting by top sellers, Competitive, At Risk, and Uncompetitive
- Competitive insight from outside the four walls of the catalog
- Reasons a product is offline, not just the fact that it is
- Direct cost actions with less manual work around them
Prioritization
05
Impact and reach, argued with numbers I didn't own alone.
Analysts supported the evaluation so sequencing came from evidence rather than the loudest request. The roadmap was reviewed with top partners and with senior Product, Merchandising, and Operations leadership, including executive leadership.
Inputs to the ranking
- —Partners and SKUs affected
- —SKUs offline due to competitiveness
- —Internal time spent
- —GMS opportunity and impact
- —Top-10 SKUs versus the broader catalog opportunity
What we built
06
Competitive status, in the same place as the cost field.
UPCitemDB expanded competitive analysis beyond manual checks of a handful of top SKUs on sites like Wayfair and Amazon. That let us surface status directly in the tool instead of explaining it in a meeting.
Conceptual view — competitive status
Competitor and channel prices
Links to the competing listing
Percent uncompetitive
At Risk lasts 5 days before a SKU becomes Uncompetitive and can be removed under business logic
Enter a new cost or price directly, or toggle to a percentage reduction
Offline status and reason — content, stock or sellability, internal action, and others
Offline reasoning is included in the CSV as well. Conceptual diagram — not a product screenshot.
Data quality
07
Competitive comparisons are only as good as the UPC.
Inaccurate and fake UPCs undermined competitive pricing and created SEO and data-quality issues. UPCitemDB mismatches were what exposed the scale of it.
We added GS1 validation during SKU creation to check UPC structure and ownership validity, so bad or recycled UPCs were caught before we trusted any comparison built on top of them.
The CSV, rebuilt
08
Before
- Export ceiling of 2,000 SKUs
- Minimal context — SKU and current cost/price
- Weak upload feedback
- Troubleshooting the whole file when anything failed
After
- Filter before exporting, including category, subcategory, and status
- Export ceiling raised from 2,000 to 10,000 SKUs
- Fields for site status, competitive status, competitor links and pricing, and offline reason
- Valid rows process; failed rows come back in a separate error file
- An error-message column explains why each row failed
- Correct the failed rows and re-upload just those
Known errors surfaced back to the supplier included formatting and readability problems, and cost increases above 15% that require review.
Cost workflow
09
Make decreases easy. Make increases reviewable.
Cost change request — routing logic
Decrease
Auto-approved
Submitted as a dollar amount or a percentage reduction, and applied in bulk or by category, so eligible SKUs can return onsite faster.
Increase above 15%
Merchandising review
Internal users filter by category, approve or deny, and leave a reason. Suppliers can see their own pending requests. The threshold was set with Merchandising leadership.
Cumulative-increase logic looks back 30 days, so repeated small increases can't quietly bypass the threshold.
Suppliers can schedule a temporary cost decrease and have the original cost restored automatically afterward, without the restore incorrectly tripping the normal increase-window logic.
Tradeoffs
10
What didn't get finished, and why.
Automatic sale badge
A temporary reduction should have shown an onsite sale badge automatically. The PDP, Cart, and Checkout teams had other priorities, so the workaround was Merchandising manually flagging SKUs in separate Product Maintenance tooling.
Counteroffer workflow
A denied increase above 15% should have opened a continued negotiation rather than a dead end. That workflow wasn't completed during my ownership, so partners had to submit a new request.
Impact
11
30%
Fewer SKUs offline due to competitiveness.
- Partner engagement increased, and partners were using Manage Cost daily.
- Operations work shifted from repeated upload escalations toward coaching partners.
- More uncompetitive SKUs could be handled in the product instead of only the handful covered in a meeting.
- Partners used the competitive data to check their pricing on other channels, and sometimes to find other sellers carrying their products.
Reflection
12
The lesson wasn't “improve the CSV.”
Manage Cost went from a transaction form to a workspace — one that gave suppliers enough context to understand what was happening to their catalog and act on it, without needing an internal team to explain every issue first.
Credits
13
KEG / PINT
The greatest engineering team on the planet. I'll fight you on this.
Genuine thanks to Engineering, Merchandising, Operations, Analytics, Product leadership, and the supplier partners who told us the truth about what wasn't working.