Product Data & Lifecycle

How a product moves from trend capture through vendor specs, PIF, PIM, inventory planning, and multi-channel launch.

What this workflow is

How every idea for a new product — a new silhouette, a new finish, a new SKU — becomes something a customer can buy, everywhere Balsam sells.

Scope	Details
Start	Merchandising identifies a product opportunity, or a vendor provides specs
End	Product is live and selling across all channels
Parallel tracks	D2C (19 steps), GE Licensed Products (17 steps), Indirect / Wholesale
Vendor count	65–75 decor vendors + primary tree vendor
Lead time	9–12 months manufacturing; 18-month commitment cycle
Channels	D2C · International · GE Retail · Indirect · B2B · Studio Stores

The flow today

graph LR
    subgraph "Upstream: GTM"
        TREND[Trend Capture<br/>1–1.5 yrs ahead]
        PDDEV[Product Development<br/>+ AI Visualization]
        SEG[Market Segmentation<br/>Brand / Channel / Geo]
        STORY[Product Stories<br/>Balsam Hill, GE,<br/>Tree Classics, Tree Topia]
    end

    subgraph "PIF to PIM"
        VEND[65–75 Vendors<br/>Email / WhatsApp / Excel]
        PIF[PIF Excel<br/>200–300 cols, 20+ files<br/>No version control]
        PIM[ContentServ PIM<br/>+ PxHub]
        MULE[MuleSoft]
        NS[NetSuite ERP]
        SF[Snowflake → Sigma]
    end

    subgraph "Demand → Purchasing"
        CFO[Revenue Targets<br/>→ Category Comps]
        DP[Demand Planning<br/>SKU → Company<br/>Reference SKUs manual]
        SP[Supply Planning<br/>Inventory Position]
        PO[Purchase Orders<br/>Manila Team]
        MFG[Manufacturing<br/>9–12 month lead time]
    end

    subgraph "Multi-Channel Launch"
        D2C[D2C<br/>balsamhill.com]
        INTL[International<br/>Big 5 Europe]
        GE[GE Retail<br/>2 → 10 retailers<br/>400-pg vendor guides]
        IND[Indirect / Amazon /<br/>Wayfair / Wholesale]
        B2B[B2B]
        STORES[BH Studio Stores]
    end

    TREND --> PDDEV --> SEG --> STORY
    VEND -->|Unstructured comms| PIF
    STORY --> PIF
    PIF -->|Manual entry| PIM
    PIM --> MULE --> NS & SF
    CFO --> DP --> SP --> PO --> MFG
    PIF -->|AI-validated| PIM
    PIF -->|Manual copy| D2C & INTL & IND & B2B & STORES
    PIF -->|Manual translation<br/>per retailer format| GE

    style PIF fill:#fecaca,stroke:#ef4444
    style MFG fill:#fecaca,stroke:#ef4444
    style GE fill:#fef3c7,stroke:#f59e0b
    style VEND fill:#fef3c7,stroke:#f59e0b
    style DP fill:#fef3c7,stroke:#f59e0b

How it works

Phase	What happens	Primary tools
Trend capture	Runway shows, competitor assortments, social signals, market intelligence, and customer reviews, 18 months ahead of market. CFO revenue targets cascade into category growth directives.	—
Trend to line list	Trade-show photo capture → manual mood boards → ideation → line-list Excel. 250–300 SKUs/yr across décor and greenery; ~100 dropped pre-launch based on sales signals. ~60% of PD time on product design; vendor spec development (hand-built PPT decks + sample iteration) is the biggest single time sink.	iPhone, Teams, PowerPoint, Excel, ChatGPT (image-gen, translation)
Vendor data & PIF	PD creates the line lists; Merch finalizes the PIF. 200–300 columns per product across 20+ PIF files. 65–75 vendors communicate via email, WhatsApp, Excel, and verbal calls.	Excel
PIF to PIM	Minimum data enters PIM early. Manila data team validates with AI. MuleSoft exports to ERP and Snowflake.	ContentServ PIM, PxHub, MuleSoft
Demand & supply	SKU-level demand plans rolled up to total company. Three-pass commitment cycle: 18 mo → 12 mo → Dec–Jan true-up.	Excel, Sigma, third-party planning tool
Purchasing	Manila purchasing team creates POs, manages vendor production, tracks inbound. 9–12 month manufacturing.	NetSuite (in progress)
Multi-channel launch	Same data must reach D2C, international, GE retailers (manual per-retailer translation), indirect, B2B, Studio.	Hybris, retailer portals

Three parallel tracks

Track	Volume / scope	Why it’s distinct
D2C	Primary, most mature — 19 steps, 18 documented bottlenecks	Highest volume; closest to industry-standard lifecycle
GE Licensed	17-step lifecycle; ~150 new SKUs this season; 2 → 10 retailers by 2027	Retailer approval gates; UL/ETL/CSA electrical certification; 400-page vendor guides with error penalties
Indirect / Wholesale	Amazon, Michaels, Lowe’s, Wayfair	Each retailer has its own portal and template; manual data re-keying

Where it gets stuck

Friction	What it looks like
Trend synthesis is manual	Thousands of phone photos from trade shows get triaged by eye. Themes emerge from memory, not from the data.
Vendor spec iteration runs in PowerPoint	Each SKU gets a hand-built deck, then back-and-forth email + physical sample shipments. First-pass AI image-gen isn’t consistent enough to shortcut this yet.
PIF has no version control	Changes flow by email and Slack. No change log. Errors surface when vendors reject POs.
Inventory planning is the de facto quality gate	~40 hours a month spent on data QA that should have been caught upstream.
Retailer format translation scales linearly	2 → 10 retailers = 5× the manual work.
Reference SKU selection is intuition-based	New products mapped to history by person, not by attribute model. Forecast quality depends on who maps.
9–12 month manufacturing lead times	Blocks mid-year opportunities, custom B2B, and store-specific SKUs.
No unified GTM calendar	Every function runs its own. Upstream date shifts (Chinese New Year, licensor approvals) don’t cascade.
No PLM upstream	Product development runs in Excel, email, Slack, PowerPoint. No concept of “done.”
Per-platform content strategy lives in Feedonomics config, not PIM	Amazon, Target, Meta, Google, and LLM feeds each want distinct descriptions and attributes. Today that’s handled downstream in feed-rule configuration rather than upstream as structured PIM fields.

What a cohort here works on

A cohort on Product Data partners with merchandising, product development, the PIM team, inventory planning, indirect sales, GE licensing, and the international teams. Two groups feel the difference immediately:

Audience	What better looks like
Internal customers (merch, PIM, inventory planning, purchasing, retail ops)	Senior leaders stop doing data entry. Errors get caught at source. 40 hours/month of QA time becomes strategic capacity.
External customers (shoppers everywhere; retail partners; new-market launches)	Complete, accurate product pages on launch day. Retailers get clean data in their template on the first try. New markets open at something closer to a flat onboarding cost.

Places a cohort finds traction:

Governed item master with change notification to every downstream consumer
Validation gates between PIF and PIM so errors get caught at source
Automated translation into the retailer templates that currently absorb the most manual work
Attribute-based reference-SKU model that takes intuition out of demand planning
Structured vendor intake that meets vendors where they already communicate
Single costing source so year-over-year history is visible per SKU
Earlier retail-ready costing — vendor-validated cost inputs, tariffs, duty rates, and loaded costs factored in upstream so margin decisions don’t wait for the full PIF
Per-platform description fields in PIM so channel-specific content (Amazon, Target, Meta, LLM feeds) syndicates automatically via Feedonomics instead of being munged from a single source

Get involved

Cohort 1 is active on this workflow. See Cohort 1 — Product Data for the specific scope and people.

See something wrong on this page? Edit it →
Have an idea that touches this workflow? Submit it →
Want to be part of a future cohort working on this workflow? Say so in your submission — cohorts form from the people who raise their hands.