⚡ QuickScan: Know exactly which workflows to automate in 2 weeks. Start your QuickScan →
Sapnity logo
RETAIL · OMNICHANNEL · DYNAMIC PRICING

Dynamic Pricing Automation for Omnichannel Retail

Sapnity implemented a Dynamic Pricing Automation Hub for a global fashion & lifestyle retailer, using Power Platform and Azure ML to optimise SKU–region–channel prices across stores, distributors and e-commerce — while protecting brand and margin guardrails.

Core Platforms: Power Apps, Power Automate, Dataverse, Azure Machine Learning, Azure Data Lake, POS / ERP, E‑commerce & Distributor Portals · Region: Europe, Middle East & Asia · Complexity: High – 40k+ SKUs across 7 channels.
Dynamic Pricing Engine SKU–Region Optimisation Omnichannel Retail Azure ML & Power Platform
10-Day Sprint
One-SKU Cluster Pricing Sprint
Take one high-impact SKU cluster (e.g., sneakers in tier‑1 cities) from Excel pricing to a governed dynamic pricing flow in ~10 days.
View 10-day sprint →
2-Week Scan
Price Elasticity & Margin QuickScan
Identify pricing pockets with the highest elasticity, margin leakage and competitive risk across SKUs, regions and channels.
Start QuickScan →
3-Week Pack
Dynamic Pricing Starter Pack
Stand up a Dynamic Pricing Hub pattern with elasticity models, pricing guardrails and approval workflows for a pilot category.
Explore Starter Pack →

1. Business Problem — Spreadsheet-Driven Prices

The retailer sold across flagship stores, franchise outlets, distributors, marketplaces and its own e‑commerce site. Pricing decisions were driven by channel-specific spreadsheets and intuition rather than a unified view of demand, competition and inventory.

  • Merchandisers built price ladders in Excel and shared them via email and chat.
  • Promo calendars varied by country and channel, often overlapping in unplanned ways.
  • Competitor price checks happened manually via store visits and website sampling.
  • Inventory build-up in certain regions was not systematically reflected in prices.
  • Finance saw margin leakage but could not pinpoint which price decisions caused it.

Everyone agreed pricing was a lever, but no one could see a single, trusted picture of price, demand and margin by SKU × region × channel.

2. Sapnity’s Mandate

Group leadership engaged Sapnity to:

  • Design a Dynamic Pricing Automation Hub that could propose price changes by SKU, region and channel based on demand, elasticity and competitor signals.
  • Preserve strategic guardrails around brand positioning, price corridors and discount caps.
  • Provide a single cockpit for pricing managers to simulate, approve and deploy price changes.
  • Integrate seamlessly with POS, ERP, distributor and e‑commerce systems.
  • Establish a reusable pricing pattern across categories and seasons.

3. Before — Weekly Price Wars in Spreadsheets

Every week, merchandisers and pricing analysts ran “pricing marathons” to decide which SKUs to discount, where and by how much. By the time price files reached channels, conditions on the ground had already shifted.

  • Price files were shared as CSVs and PDFs to different channel teams.
  • Mid-week competitor moves were handled with one‑off price overrides.
  • Inventory and sell‑through data arrived late and in inconsistent formats.
  • Test‑and‑learn experiments were limited to a few stores due to manual effort.

Operational Heatmap — Where Pricing Got Stuck

StepPrimary OwnerCommon Issue
Demand review Merchandising Lagging data; no granular elasticity insight.
Price proposal Pricing team Multiple Excel versions; limited scenario analysis.
Approval Country heads Email-based approvals; unclear rationale.
Deployment Channel ops Manual uploads into POS and e‑commerce; delays and errors.
Post‑event review Finance / Analytics No clear view of which decision drove which outcome.

4. After — Sapnity Dynamic Pricing Automation Hub

Sapnity introduced a Dynamic Pricing Automation Hub that ingests demand, inventory and competitor data, runs elasticity‑aware models, and generates guardrail‑compliant price recommendations that flow to channels via governed workflows.

RETAIL DYNAMIC PRICING PATTERN

Demand, Inventory & Competition Feeds

POS, e‑commerce, distributor, web‑scraped competitor prices, promo calendars.

Data Prep & Feature Store

Cleaned, aligned SKU–region–channel time‑series with price, volume and margin features.

Elasticity & Uplift Models

Azure ML models estimating demand response to price and promotion by cluster.

Pricing Guardrail Engine

Rules enforcing brand corridors, minimum margins, competitor bands and discount caps.

Pricing Workbench (Power Apps)

Pricing managers simulate, compare and approve price scenarios by SKU and region.

Channel Price Deployment

Approved prices flow automatically into POS, ERP, distributor and e‑commerce systems, with effectivity dates and audit trail.

Pricing managers now operate a single pricing cockpit instead of wrestling with disconnected spreadsheets and manual uploads.

5. Implementation Story

Phase 1 — Discovery & Data Reality Check

  • Ran workshops with merchandising, pricing, finance and channel ops teams.
  • Mapped how prices actually changed week‑to‑week across channels and regions.
  • Audited data sources: POS, e‑commerce, distributor reports, competitor scrapes.

Phase 2 — Pricing Blueprint

  • Defined pricing strategies by category: everyday value, fashion‑forward, clearance, seasonal, outlet.
  • Designed a pricing guardrail framework with finance and brand teams.
  • Agreed a minimum viable feature set for elasticity models (price, discount depth, promo type, seasonality, inventory).

Phase 3 — Data & Modelling

  • Built Azure data pipelines to unify sales, inventory and competitor feeds.
  • Clustered SKUs into pricing families (e.g., “white sneakers <₹5k” in tier‑1 cities).
  • Trained and validated elasticity models; surfaced model quality indicators in the UI.

Phase 4 — Dynamic Pricing Hub Build

  • Developed the Dynamic Pricing Workbench in Power Apps for pricing teams.
  • Implemented approval flows and digital sign‑offs for country heads.
  • Integrated with POS, ERP and e‑commerce to deploy prices with proper effectivity dates and rollback capability.

Phase 5 — Rollout, Test & Optimisation

  • Piloted with two categories across three countries; compared uplift vs. control stores.
  • Refined guardrails and alerts based on real outcomes and exceptions.
  • Rolled out to the broader portfolio with training for merchandisers and finance analysts.

6. Technical Architecture — Layered View

Experience Layer Power Apps Dynamic Pricing Workbench for pricing teams; dashboards in Power BI for executives and finance; Teams notifications for key approval events.
Data & Feature Layer Azure Data Lake and Dataverse tables holding unified sales, inventory, promotions, competitor prices and derived features.
ML & Optimisation Layer Azure Machine Learning for elasticity and uplift models; pricing optimisation algorithms embedding guardrails and constraints.
Workflow & Governance Layer Power Automate flows to orchestrate pricing cycles, approvals, exceptions and compliance checks with full audit trail.
Channel Integration Layer APIs and connectors pushing approved prices to POS, ERP, distributor portals and e‑commerce engines; scheduled price sync jobs.
Analytics & Experimentation Layer Power BI models for margin, volume and uplift tracking; A/B test monitoring and experiment archives for future learning.
DYNAMIC PRICING PATTERN

7. Reusable Dynamic Pricing Pattern for New Categories

Instead of building separate pricing solutions for each category or region, the retailer now uses a single Dynamic Pricing pattern. New SKUs and geographies are onboarded by configuring models, guardrails and workflows — not by writing new code.

Unified demand & price data model
Elasticity models by SKU cluster
Configurable brand & margin guardrails
Scenario simulation & approvals
Automatic channel deployment
A/B testing & uplift measurement

8. Outcomes & KPIs

KPIBeforeAfter Sapnity Dynamic Pricing Hub
Price update cadence for key SKUs 7–10 days 2–3 days, with daily simulations possible
Manual price overrides per week 150–220 < 50, mainly for strategic exceptions
Promo margin leakage (flagship categories) 3–5% < 1.5% after guardrails and optimisation
Time to evaluate pricing scenarios 2–3 days of Excel work < 1 hour for multiple what‑if scenarios
Incremental gross margin (pilot categories) Baseline +1.5–2.3 pts vs. comparable control period

The Dynamic Pricing Hub is now a strategic engine for planning promotions, reacting to competition and managing end‑of‑season sell‑through.

9. Sapnity Differentiators

  • Retail‑first thinking: We started from merchandiser and channel realities (seasonality, hero SKUs, fashion cycles), not just from generic ML patterns.
  • Guardrails by design: Brand and finance owned the guardrails; ML was built inside those boundaries, not the other way around.
  • Explainable recommendations: Pricing workbench shows “why this price” — drivers, constraints and sensitivity, not just a number.
  • Built on the Microsoft stack: Power Platform + Azure ML + Power BI gave a secure, enterprise‑ready foundation with low IT friction.
  • Pattern, not project: The same architecture now supports dynamic pricing pilots in new regions and categories without re‑platforming.

For this retailer, Sapnity transformed pricing from a spreadsheet fire‑fight into a data‑driven, test‑and‑learn capability that the business can scale.