Build a Multi-Channel Retail Advertising ROI Engine with Google ADK and Gemini
Flipkart's advertising vertical drove significant revenue growth for Walmart's e-commerce ecosystem. In this challenge, you will construct a multi-channel ad attribution and spend optimization engine using Google ADK (Agent Development Kit) integrated with Vertex AI and Gemini multimodal models. The system will ingest ad campaign performance metrics across marketplace channels (Flipkart Ads, Amazon Sponsored Products, Google Shopping), perform dynamic Return on Ad Spend (ROAS) optimization, and execute cross-channel budget reallocation recommendations.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an advertising spend allocation agent using Google ADK and Vertex AI to optimize marketplace ROAS.
How work is evaluated
Evaluates accuracy of campaign attribution calculation and performance output.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
budget_conservation_test
Ensures total recommended spend equals original budget.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
roas_maximization
Efficiency of budget allocation optimization. • target: 0.92 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build multi-modal agent workflows using Google Agent Development Kit (ADK).
Process visual banner assets alongside tabular ad performance metrics.
Compute incrementality and multi-touch attribution models.
Generate automated real-time ad spend rebalancing suggestions.
Reference links and supporting material
Cross-channel campaign telemetry data for consumer electronic brands advertising on Indian e-commerce portals.
Challenge setup
Evaluates accuracy of campaign attribution calculation and performance output.
Challenge input
JSON containing list of channel metrics with spend, revenue, and impressions.
Google ADK
Required framework for building Gemini-powered agents.
Evaluated output
JSON with reallocation_plan and expected_roas_uplift.
- Ensures total recommended spend equals original budget.
- Efficiency of budget allocation optimization. • target: 0.92 • range: 0-1
- Benchmark: AdTech Optimization Standard
- Roas Maximization target: 0.92
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Google ADK · Required
- Google Gemini · Optional
- Google · Optional
Evaluation contract
- budget_conservation_test · Weight 1
- roas_maximization · Weight 1
Metrics
- roas_maximization · Target 0.92
Benchmarks
- AdTech Optimization Standard
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing list of channel metrics with spend, revenue, and impressions.
Output: JSON with reallocation_plan and expected_roas_uplift.
Execution harness: Python
Evaluation tools
- Google ADK · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.