	In Browser
	StumbleUpon
	del.icio.us
	Google
	Google Buzz
	reddit
	LinkedIn

	Facebook
	Twitter
	Linkedin
	E-Mail

Generative AI > Google Gemini API > Campaign ROI Agent

Campaign ROI Agent

Author: Venkata Sudhakar

Marketing teams run dozens of campaigns simultaneously across channels like Google Ads, Instagram, email, and SMS. Without a unified ROI view, budget is often misallocated to low-performing channels. ShopMax India needs an agent that consolidates campaign spend and revenue data and surfaces the most efficient campaigns instantly.

This tutorial builds a Gemini ADK agent that reads campaign performance data, computes ROI, cost per acquisition (CPA), and revenue per rupee spent, then ranks campaigns from best to worst performer.

The below example shows a campaign ROI analysis agent for ShopMax India in a business context.

import asyncio
from google.adk.agents import LlmAgent
from google.adk.tools import FunctionTool
from google.adk.runners import Runner
from google.adk.sessions import InMemorySessionService
from google.genai import types

# Campaign performance data for March 2026
CAMPAIGNS = [
    {"name": "Google Search - Electronics",  "channel": "Google Ads",
     "spend_rs": 85000, "revenue_rs": 340000, "conversions": 120},
    {"name": "Instagram Reels - TV Sale",     "channel": "Instagram",
     "spend_rs": 45000, "revenue_rs": 98000,  "conversions": 34},
    {"name": "Email - April Deals",           "channel": "Email",
     "spend_rs": 8000,  "revenue_rs": 62000,  "conversions": 88},
    {"name": "SMS Blast - Clearance",         "channel": "SMS",
     "spend_rs": 12000, "revenue_rs": 55000,  "conversions": 65},
    {"name": "YouTube Pre-roll - Laptops",    "channel": "YouTube",
     "spend_rs": 60000, "revenue_rs": 72000,  "conversions": 18},
]

def analyse_campaign_roi() -> dict:
    results = []
    for c in CAMPAIGNS:
        roi_pct = round((c["revenue_rs"] - c["spend_rs"]) / c["spend_rs"] * 100, 1)
        cpa = round(c["spend_rs"] / c["conversions"], 0) if c["conversions"] else 0
        rev_per_rupee = round(c["revenue_rs"] / c["spend_rs"], 2)
        results.append({
            "campaign": c["name"],
            "channel": c["channel"],
            "spend_rs": c["spend_rs"],
            "revenue_rs": c["revenue_rs"],
            "conversions": c["conversions"],
            "roi_pct": roi_pct,
            "cpa_rs": int(cpa),
            "rev_per_rupee": rev_per_rupee,
        })
    results.sort(key=lambda r: r["rev_per_rupee"], reverse=True)
    total_spend = sum(c["spend_rs"] for c in CAMPAIGNS)
    total_revenue = sum(c["revenue_rs"] for c in CAMPAIGNS)
    blended_roi = round((total_revenue - total_spend) / total_spend * 100, 1)
    return {
        "period": "March 2026",
        "total_spend_rs": total_spend,
        "total_revenue_rs": total_revenue,
        "blended_roi_pct": blended_roi,
        "campaigns_ranked": results,
    }

roi_tool = FunctionTool(func=analyse_campaign_roi)

agent = LlmAgent(
    name="campaign_roi_agent",
    model="gemini-2.0-flash",
    description="Analyses marketing campaign ROI for ShopMax India and ranks campaigns by efficiency.",
    instruction=(
        "You are a marketing analytics agent for ShopMax India. "
        "When asked for a campaign ROI report, call analyse_campaign_roi. "
        "Present the ranked campaigns with ROI %, CPA, and revenue per rupee. "
        "Highlight the top performer and the underperformer that needs review."
    ),
    tools=[roi_tool],
)

async def main():
    session_service = InMemorySessionService()
    runner = Runner(agent=agent, app_name="roi_app", session_service=session_service)
    session = await session_service.create_session(app_name="roi_app", user_id="mktg1")
    msg = types.Content(role="user", parts=[types.Part(
        text="Give me the campaign ROI report for March 2026"
    )])
    async for event in runner.run_async(user_id="mktg1", session_id=session.id, new_message=msg):
        if event.is_final_response() and event.content:
            print(event.content.parts[0].text)

asyncio.run(main())

It gives the following output,

Campaign ROI Report - ShopMax India | March 2026

Total Spend  : Rs 2,10,000
Total Revenue: Rs 6,27,000
Blended ROI  : 198.6%

Ranked by Revenue per Rupee Spent:

1. Email - April Deals           | ROI: 675%  | CPA: Rs 91   | Rev/Rs: 7.75  [TOP]
2. SMS Blast - Clearance         | ROI: 358%  | CPA: Rs 185  | Rev/Rs: 4.58
3. Google Search - Electronics   | ROI: 300%  | CPA: Rs 708  | Rev/Rs: 4.00
4. Instagram Reels - TV Sale     | ROI: 118%  | CPA: Rs 1324 | Rev/Rs: 2.18
5. YouTube Pre-roll - Laptops    | ROI: 20%   | CPA: Rs 3333 | Rev/Rs: 1.20  [REVIEW]

Email delivers the highest ROI at 7.75x. YouTube Pre-roll underperforms
at 1.20x - consider pausing or reducing budget until creative is refreshed.

ShopMax India's marketing manager can use this agent every Monday morning to reallocate budget from underperforming channels to top performers. Connecting the agent to the Google Ads and Meta APIs enables live spend and conversion data rather than manually uploaded spreadsheets.

Send your comments, suggestions or queries regarding this site to [email protected].