Case Study: Seasonal E-commerce Achieves 350% ROAS with Dynamic Budget Management

1. Unpredictable Seasonal Demand

The business had three major seasonal peaks that drove 60% of annual revenue:

• Valentine’s Day (Feb 10-14): 5-day window, roses and romantic gifts
• Mother’s Day (March): 10-day window, plant arrangements and gift sets
• Christmas (Nov-Dec): 6-week window, festive plants and gifts

Problem: Manual budget adjustments were too slow. By the time they reacted to increased demand, peak selling days had passed.

Impact: Lost 20-30% of potential revenue during seasonal peaks due to budget constraints.

2. Stock Management Nightmare

Living plants have finite inventory—you can’t suddenly produce more when demand spikes. The client had two recurring problems:

• Stockouts from over-advertising: Advertising sold out products, damaging customer experience
• Unsold inventory from under-advertising: Conservative budgets left stock unsold (perishable products = waste)

Impact: Either disappointed customers or wasted inventory—no middle ground.

3. Product Margin Variation

Not all products were equally profitable:

• High-margin: Rose arrangements (65% margin), premium plants
• Medium-margin: Standard plants (45% margin), gift sets
• Low-margin: Entry-level plants (25% margin), promotional bundles

Problem: Previous Performance Max setup treated all products equally, wasting budget on low-margin items while underfunding high-margin winners.

4. Geographic Delivery Constraints

Living plants can’t be shipped everywhere. The business delivered to UK mainland only, but ads were showing to customers in Northern Ireland, Scottish islands, and Channel Islands—wasting budget on un-deliverable locations.

Impact: 8-12% of ad spend wasted on geographic areas they couldn’t serve.

🤖 Dynamic Budget Allocation

How it works: Daily budget adjustments based on:

• Current ROAS performance (last 7 days)
• Days until peak event (Valentine’s Day countdown)
• Available stock levels (via inventory feed integration)
• Weather forecasts (warm weather = higher plant demand)

Example: 10 days before Valentine’s Day, budgets automatically increased 20% per day if ROAS remained above 300%.

📦 Inventory-Aware Bidding

How it works: Bid adjustments based on stock levels:

• High stock (100+ units): Normal bidding
• Medium stock (20-100 units): 20% bid reduction (preserve for organic traffic)
• Low stock (<20 units): 50% bid reduction
• Out of stock: Product automatically excluded from feed

Result: Never over-advertised limited stock, never stockouts from advertising.

💰 Overall Seasonal Performance (Full Year)

• Valentine’s Day: 350% ROAS, £85k revenue
• Mother’s Day: 345% ROAS, £92k revenue
• Christmas Period: 330% ROAS, £168k revenue
• Off-Season Average: 380% ROAS (better than previous peaks!)

Annual Impact: £127k additional revenue vs previous year (+38%)

📊 Efficiency Improvements

• 40% higher conversion rate during peak periods (vs Standard Shopping)
• Zero stockouts from over-advertising (vs 8 incidents previous year)
• 15% reduction in wasted inventory (better demand forecasting)
• 10% budget savings from geographic targeting improvements

⚙️ Operational Benefits

• 95% reduction in manual budget adjustments (automation handled it)
• Real-time inventory integration prevented customer disappointment
• Weather-responsive bidding captured unexpected warm-weather demand spikes
• Predictable peak season performance (no more guessing budgets)

📈 Valentine’s Day Learnings → Mother’s Day Improvements

• Finding: Customers searched “same day delivery” heavily on Mother’s Day itself
• Action: Added “same day delivery” headlines, increased bids 50% on Mother’s Day morning
• Result: Captured 180 same-day orders at 320% ROAS

🎄 Christmas Strategy Evolution

• Finding: Corporate gift orders (bulk purchases) had 2x average order value
• Action: Created separate campaign targeting businesses, “corporate gift” keywords
• Result: £42k B2B revenue at 380% ROAS (entirely incremental)

✅ What Worked

• Inventory integration: Zero stockouts = happy customers + more sales
• Seasonal asset rotation: Relevant creative = 40% conversion rate lift
• Weather-responsive bidding: Captured impulse purchases during sunny weekends
• Geographic precision: 10% budget savings from excluding undeliverable areas
• Campaign separation: Seasonal vs year-round strategies optimised independently

⚠️ What Didn’t Work Initially

• Too conservative scaling: Initial Valentine’s budget was 2x (should have been 3x) – left money on table
• Generic peak messaging: “Order now” wasn’t urgent enough; “Delivery by Feb 14” performed 60% better
• Delayed asset activation: Year 1, Valentine’s assets went live Feb 5; Year 2, Jan 25 start captured early planners