No more demand surprises. Ward sees them first.
Insight cards surface demand patterns your dashboards miss.
Why demand matters
in fashion retail.
Most fashion SKUs have zero sales history, they're new every season, so time-series models fail. Ward takes an attribute-based approach, clustering new styles against historical analogues by silhouette, colorway, price point, and fabric weight, then calibrating in real time as early sell-through data arrives.
Industry benchmarks
Fashion forecast accuracy: 30-45% MAPE pre-season, dropping to 18-28% by week 4 of selling. Operators using attribute-based modeling typically reduce week-1 first-allocation error by 25-40% and recover 1-3 points of full-price sell-through.
Pre-season buy planning, fall collection
The buying team is finalizing quantities for hundreds of new fall styles with no sell-through history. Ward maps each to attribute clusters from prior seasons and adjusts for current trend velocity. The result is store-cluster-level buy recommendations that materially reduce first-allocation error, meaning fewer stockouts on winners and less dead inventory on misses.
What Ward actually tracks
Ward uses attribute-based similarity models, trend velocity indicators, store cluster demand profiles, and early-signal calibration from the first weeks of sell-through. It also tracks fashion cycle timing to anticipate when trends peak and decay.
Data signals
Style attributes from PIM, prior-season sell-through analogs, store cluster demographics, trend velocity from internal and external signals, and weekly sell-through during the selling window.
Three pitfalls Ward catches
in fashion demand.
- 01 Pre-season buys are sized off prior-year category totals, ignoring that the trend mix has shifted (more elevated denim, less basic tee) within the category.
- 02 First-allocation curves use chain-average size profiles when each store cluster has a meaningfully different size mix.
- 03 Early sell-through (weeks 1-2) is dismissed as noise when in reality it's the highest-signal indicator of full-season trajectory.
How Ward runs demand
for fashion retailers.
-
01
Build attribute-based analog mapping
Every new style gets matched to 3-5 historical analogues by silhouette, colorway, fabric, and price point, producing a probabilistic demand curve rather than a single point estimate.
-
02
Calibrate on early sell-through
After weeks 1-2, Ward updates the trajectory using the actual signal and re-allocates remaining inventory across stores.
-
03
Feed misses back into the next pre-season
Categorical biases (overforecasting basics, underforecasting trend) are surfaced as recurring patterns to inform next-season planning.
What a Ward card looks like.
72-hour heat wave predicted for Dhaka region. Historical model suggests +18% on beverages, +12% on ice cream. Pre-position recommended.
Chat
Ask anything. Ward routes to the right agent and returns cited answers.
I pulled Store 37’s last 28 days against the chain baseline. Two root causes, both compounding.
| Signal | Finding |
|---|---|
labor_efficiency | Rev/labor-hour −22% vs. cluster, staffing mismatch at 11a–1p peak |
inventory.fresh | Fresh fill 83%, backroom replenishment lag at 2–4p |
promo.lift | BOGO crackers cannibalized Brand Y by 28%, net category +6% |
Recommend: re-baseline Store 37 schedule against true peak, raise replen window to 1p, and review the BOGO before next cycle.
labor_scheduling…
Dashboards
Pinned views built from saved data-lake queries.
Models
Browse, search, and manage data–lake model definitions for your tenant.
| Name | Namespace | Version |
|---|---|---|
retail_pos_transactions | retail | 1.0 |
retail_inventory_snapshot | retail | 1.2 |
retail_labor_scheduling | retail | 1.0 |
retail_promo_calendar | retail | 1.1 |
retail_supplier_performance | retail | 1.0 |
sap_inventory_shrinkage | sap | 1.0 |
ga4_daily_events | marketing | 1.0 |
meta_ads_ad_level | marketing | 1.0 |
Sources
Connect external systems to the data lake.
| Name | Type | Last sync |
|---|---|---|
sap_pos_transactions | import | 2m ago |
sap_inventory_shrinkage | import | 2m ago |
sap_labor_scheduling | import | 14m ago |
retail_inventory_weekly | import | 1h ago |
retail_google_ads_daily | import | 1h ago |
retail_meta_ads_daily | import | 1h ago |
retail_ga4_website_daily | import | 1h ago |
Architecture
Two ways to connect. Federate against your live systems, or ingest into Ward’s data lake. Toggle below.
sap.possnow.inventoryPipelines
Move data from sources into models on a schedule.
| Name | Source | Model | Status | Schedule |
|---|---|---|---|---|
sync_sap_pos_transactions | sap_pos_transactions | pos_transactions | enabled | hourly |
sync_sap_labor_scheduling | sap_labor_scheduling | labor_scheduling | enabled | daily |
sync_sap_inventory_shrinkage | sap_inventory_shrinkage | inventory_shrinkage | enabled | daily |
sync_retail_inventory_weekly | retail_inventory_weekly | inventory_weekly | enabled | weekly |
sync_retail_google_ads_daily | retail_google_ads_daily | google_ads_daily | enabled | daily |
sync_retail_ga4_website_daily | retail_ga4_website_daily | ga4_website_daily | enabled | daily |
Streams
Real-time ingestion pipelines.
pos.txnstore_037, basket $42.18inv.movedc_west → store_104labor.clockstore_022 shift_startpos.txnstore_211, basket $19.04
Policies
Browse and manage Cedar access policies for your tenant.
| Policy ID | Effect | Resources |
|---|---|---|
merch-read-default | permit | Model::* |
finance-read-shrinkage | permit | Model::"shrinkage" |
vendor-blocked | forbid | Model::"labor_*" |
region-west-only | permit | Tenant::"acme" |
Entities
Principals and resources referenced by Cedar policies.
| Entity UID | Type | Tenant |
|---|---|---|
Tenant::"acme" | Tenant | acme |
Model::"sap.pos_transactions" | Model | acme |
Model::"sap.inventory_shrinkage" | Model | acme |
Model::"sap.labor_scheduling" | Model | acme |
Model::"retail.toast_pos_daily" | Model | acme |
Model::"retail.ga4_website_daily" | Model | acme |
Providers
Manage LLM API keys and the model profiles that use them.
| Name | Provider | Used by | Created |
|---|---|---|---|
anthropic-default | Anthropic | 3 profiles | Apr 22 |
openai-default | OpenAI | 2 profiles | Apr 22 |
gemini-default | Gemini | 1 profile | Apr 22 |
ollama-onprem | Ollama | 2 profiles | Apr 22 |
LLM-agnostic. Bring your own key, route per task. No lock-in.
Settings
Manage your dashboard preferences and account.
Light and dark themes are available. Your choice is remembered per browser.
Fashion demand:
the shift.
- ×Markdown timing
- ×Size curve misallocation
- ×Style velocity prediction
- ✓Store-SKU-day level precision
- ✓Weather-driven adjustment
- ✓Event and holiday modeling
Fashion KPI impact.
Ward requires at least 2 full selling cycles to baseline style velocity and markdown timing. Results vary between basics and trend-driven categories.
Questions about fashion demand.
Most fashion SKUs have zero sales history, they're new every season, so time-series models fail. Ward takes an attribute-based approach, clustering new styles against historical analogues by silhouette, colorway, price point, and fabric weight, then calibrating in real time as early sell-through data arrives.
The buying team is finalizing quantities for hundreds of new fall styles with no sell-through history. Ward maps each to attribute clusters from prior seasons and adjusts for current trend velocity. The result is store-cluster-level buy recommendations that materially reduce first-allocation error, meaning fewer stockouts on winners and less dead inventory on misses.
Ward uses attribute-based similarity models, trend velocity indicators, store cluster demand profiles, and early-signal calibration from the first weeks of sell-through. It also tracks fashion cycle timing to anticipate when trends peak and decay.
First demand insight cards arrive within 48 hours. Robust fashion baselines form within two weeks. Ward requires at least 2 full selling cycles to baseline style velocity and markdown timing. Results vary between basics and trend-driven categories.
More Fashion insight cards.
Fashion retailers: see what demand problems Ward catches.
Root causes, not just alerts. See it on your data.
Find out what your data has been hiding.
Tell us about your operation. We’ll show you the problems Ward catches, and the ones your current tools miss.