Why Companies Need a Digital Marketer

# Analysing a Google Ads campaign
# Every performance marketer does this weekly.

import pandas as pd

# Sample export from Google Ads (usually CSV)
data = {
    "campaign": ["Brand", "Brand", "Generic", "Generic", "Retargeting"],
    "keyword": ["nike shoes", "adidas shoes", "running shoes", "sports shoes", "cart-abandon"],
    "impressions": [12000, 8000, 45000, 32000, 6000],
    "clicks": [480, 320, 900, 640, 240],
    "cost": [2400, 1600, 7200, 5120, 1800],
    "conversions": [24, 16, 18, 12, 36],
    "revenue": [18000, 12000, 9000, 6000, 27000],
}

df = pd.DataFrame(data)

# Calculate the core marketing metrics
df["ctr"] = (df["clicks"] / df["impressions"] * 100).round(2)          # Click-through rate
df["cpc"] = (df["cost"] / df["clicks"]).round(2)                        # Cost per click
df["cvr"] = (df["conversions"] / df["clicks"] * 100).round(2)           # Conversion rate
df["cpa"] = (df["cost"] / df["conversions"]).round(2)                   # Cost per acquisition
df["roas"] = (df["revenue"] / df["cost"]).round(2)                      # Return on ad spend

# Sort by ROAS to see what's working
print(df.sort_values("roas", ascending=False).to_string(index=False))

# Output:
#    campaign         keyword  impressions  clicks  cost  conversions  revenue  ctr   cpc   cvr    cpa    roas
#   Retargeting   cart-abandon         6000     240  1800           36      27000  4.00  7.50 15.00  50.00   15.00
#         Brand      nike shoes        12000     480  2400           24      18000  4.00  5.00  5.00 100.00    7.50
#         Brand    adidas shoes         8000     320  1600           16      12000  4.00  5.00  5.00 100.00    7.50
#       Generic   running shoes        45000     900  7200           18       9000  2.00  8.00  2.00 400.00    1.25
#       Generic    sports shoes        32000     640  5120           12       6000  2.00  8.00  1.88 426.67    1.17

# What this shows a marketer:
# - Retargeting has the best ROAS (15x) — scale it
# - Brand keywords are reliable (7.5x)
# - Generic keywords are burning cash (1.2x) — pause or optimise
#
# A digital marketer looks at this and makes decisions.
# That is the job.

// GA4 event tracking — what a marketing analyst sets up
// Without events, you cannot measure what users actually do.

// Send an event when a user signs up
gtag("event", "sign_up", {
  method: "email",                // how they signed up
  value: 0,                       // optional numeric value
  currency: "INR",
});

// Send an event when a user adds a product to cart
gtag("event", "add_to_cart", {
  currency: "INR",
  value: 1499,
  items: [
    {
      item_id: "SKU_12345",
      item_name: "Wireless Mouse",
      price: 1499,
      quantity: 1,
    },
  ],
});

// Send an event when a purchase completes
gtag("event", "purchase", {
  transaction_id: "T_12345",
  value: 1499,
  tax: 149,
  shipping: 0,
  currency: "INR",
  items: [
    {
      item_id: "SKU_12345",
      item_name: "Wireless Mouse",
      price: 1499,
      quantity: 1,
    },
  ],
});

// Why this matters:
// - Without events, GA4 can only see pageviews
// - With events, you can see the full funnel:
//     pageview → add_to_cart → purchase
// - You can then attribute revenue to specific ads
// - You can report "₹X in ad spend → ₹Y in tracked revenue"
//
// Setting up events correctly is one of the most valuable
// digital marketing skills — and one of the most overlooked.

-- BigQuery / GA4 export — a real marketing query
-- Used by growth and analytics teams to make decisions.

-- Weekly revenue by channel
SELECT
  DATE_TRUNC(event_date, WEEK(MONDAY)) AS week,
  traffic_source.source AS channel,
  COUNT(DISTINCT user_pseudo_id) AS users,
  COUNT(DISTINCT IF(event_name = 'purchase', user_pseudo_id, NULL)) AS buyers,
  SUM(IF(event_name = 'purchase', ecommerce.purchase_revenue, 0)) AS revenue
FROM `analytics_123456.events_*`
WHERE event_date >= DATE_SUB(CURRENT_DATE(), INTERVAL 90 DAY)
GROUP BY 1, 2
ORDER BY 1 DESC, 5 DESC;

-- Funnel by channel (last 30 days)
WITH base AS (
  SELECT
    traffic_source.source AS channel,
    user_pseudo_id,
    COUNTIF(event_name = 'view_item') AS viewed,
    COUNTIF(event_name = 'add_to_cart') AS added,
    COUNTIF(event_name = 'purchase') AS purchased
  FROM `analytics_123456.events_*`
  WHERE event_date >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
  GROUP BY 1, 2
)
SELECT
  channel,
  COUNT(DISTINCT user_pseudo_id) AS users,
  SUM(viewed) AS view_events,
  SUM(added) AS add_to_cart_events,
  SUM(purchased) AS purchases,
  ROUND(SUM(purchased) * 100.0 / COUNT(DISTINCT user_pseudo_id), 2) AS cvr_pct
FROM base
GROUP BY 1
ORDER BY 6 DESC;

-- What a marketer does with this:
-- - Identify which channels are driving revenue
-- - See where users drop off in the funnel
-- - Decide where to invest more budget
--
-- This is the difference between "marketing" and "performance marketing".
-- A digital marketer can write or at least read this kind of query.