Automating Keyword Research with Python & Google Keyword Planner API

← Back to Blog

Keyword research is core to SEO. In this guide we’ll use Python with the Google Keyword Planner API to pull keyword ideas at scale, then export them to CSV for analysis and planning.

Why automate keyword research?

• Save hours of manual lookups in Google Ads.
• Quickly surface high-volume, lower-competition opportunities.
• Standardise research across markets and teams.

Setup

1. Create/activate a Google Ads account.
2. Generate a google-ads.yaml credentials file (Developer token, OAuth client, refresh token, login-customer-id if needed).
3. Install the client library:

   pip install google-ads

Python example: fetch keyword ideas & export CSV

Updates change over time; adjust versions/fields if Google updates the SDK.

from google.ads.googleads.client import GoogleAdsClient
from google.ads.googleads.v16.services.types import GenerateKeywordIdeasRequest
import csv

# Load credentials from google-ads.yaml in the same directory or HOME path
client = GoogleAdsClient.load_from_storage(path="google-ads.yaml")

# Replace with your Google Ads customer ID (no hyphens), language & location constants
CUSTOMER_ID = "0000000000"
LANGUAGE_ID = "1000"  # English
LOCATION_IDS = ["2826"]  # United Kingdom (example geo target)

def keyword_constant(resource_type, value):
    svc = client.get_service(resource_type)
    return getattr(svc, f"{resource_type[0].lower()+resource_type[1:-7]}_path")(value)

def generate_ideas(seed_keywords):
    kp_service = client.get_service("KeywordPlanIdeaService")
    request = GenerateKeywordIdeasRequest(
        customer_id=CUSTOMER_ID,
        language=client.get_service("LanguageConstantService").language_constant_path(LANGUAGE_ID),
        geo_target_constants=[
            client.get_service("GeoTargetConstantService").geo_target_constant_path(loc)
            for loc in LOCATION_IDS
        ]
    )
    request.keyword_seed.keywords.extend(seed_keywords)
    resp = kp_service.generate_keyword_ideas(request=request)
    rows = []
    for r in resp.results:
        metrics = r.keyword_idea_metrics
        rows.append([
            r.text,
            metrics.avg_monthly_searches,
            metrics.competition.name if metrics.competition is not None else "UNSPECIFIED"
        ])
    return rows

if __name__ == "__main__":
    ideas = generate_ideas(["seo consulting", "technical seo audit", "local seo services"])
    with open("keyword_ideas.csv", "w", newline="", encoding="utf-8") as f:
        w = csv.writer(f)
        w.writerow(["Keyword", "Avg Monthly Searches", "Competition"])
        w.writerows(ideas)
    print("Saved keyword_ideas.csv")

Next steps

• Cluster by intent: use TF-IDF + KMeans to group ideas into topics.
• Filter by difficulty: enrich with SERP difficulty/competition to shortlist targets.
• Map to content: assign clusters to blog/product pages and add internal links.

SEO benefits

• Faster iteration → more up-to-date topical coverage.
• Data-driven focus → less time on low-impact queries.
• Clear input to briefs → better on-page optimisation and internal linking.

← Back to Homepage