DEV Community

Lindenbyte
Lindenbyte

Posted on Edited on

Spy on your competitors' Google ads (Search, YouTube, Display) and get alerted when they launch new ones

Disclosure: I'm Lindenbyte, the small studio that built the Apify Actor used below, and this post was
drafted with AI assistance. Google's own Ads Transparency Center is free and works fine for a quick look;
the Actor adds bulk export, filters you can script, and daily runs.

Every advertiser on Google has a public page in the Ads Transparency Center:
every Search, YouTube and Display ad they run, with first-shown and last-shown dates. That is free
competitive intelligence, but the UI has no export, no "what's new since yesterday" and no alerts.
Here is how to get it as data, find a competitor's likely winners, and get pinged when they launch something new.

1. Pick the competitor: use the domain, not the name

Open the Google Ads Transparency Center Scraper
and give it domains:

{
  "queries": ["notion.com", "hubspot.com", "shopify.com"],
  "region": "US",
  "format": "all",
  "maxAdsPerAdvertiser": 300
}
Enter fullscreen mode Exit fullscreen mode

A domain query returns ads that point to that domain, from all advertiser entities. A name query has to
guess which legal entity you mean, and it can guess wrong. When I ran "Notion", the log said:

"Notion" -> Notion Limited (AR01024198173649797121, ~80 ads). Other matches (pass the ID or a domain
to target them): Notion OU AR03733763719230914561; Notionmind AR08497814192110174209; ...
Enter fullscreen mode Exit fullscreen mode

That picked a different company: 17 image ads, none shown since July. The notion.com query returned
125 ads, all from Notion Labs, Inc. So use the domain; if you do use a name, read that log line and pass
the right AR... ID (or raise maxAdvertisersPerQuery).

The domain query has a bonus: it also catches partners and resellers advertising your competitor.
For hubspot.com, 78 of the 300 ads came from 49 other advertisers (agencies, app vendors) pointing to hubspot.com.

Each row looks like this (trimmed):

{
  "advertiserName": "Hubspot, Inc.",
  "creativeId": "CR01111966079451660289",
  "format": "video",
  "firstShown": "2026-02-17T21:36:11.000Z",
  "lastShown": "2026-11-13T06:27:26.000Z",
  "daysActive": 269,
  "landingDomain": "hubspot.com",
  "transparencyUrl": "https://adstransparency.google.com/advertiser/AR10072600183532683265/creative/CR01111966079451660289"
}
Enter fullscreen mode Exit fullscreen mode

Image ads carry an imageUrl, video and some text ads a previewUrl, and every row links to the ad on Google's own page.

2. Narrow it: region and format

region takes a country code (US, GB, DE, BG ...) or anywhere; format is all, text, image or video,
the same filters as Google's UI. Results come newest-first (by last shown), so maxAdsPerAdvertiser gives you
the most recently active ads and caps your cost.

Watch the cap with big advertisers. With "maxAdsPerAdvertiser": 1000 and "onlyNewSince": "2026-11-12",
shopify.com in the US hit the cap: 1,000 ads, all last shown in the previous ~10 hours. If a competitor is
that large, split the job by format or region so you are not looking at an arbitrary slice.

3. Analyze: what's new, what's working, what mix

Export as JSON and a few lines of pandas give you the picture:

import pandas as pd
df = pd.read_json("ads.json", convert_dates=["firstShown", "lastShown"])
now = pd.Timestamp.now(tz="UTC")
for q, g in df.groupby("query"):
    new30 = (g.firstShown >= now - pd.Timedelta(days=30)).sum()
    print(q, len(g), "ads |", new30, "first shown in last 30 days |", g.format.value_counts().to_dict())
    print(g.sort_values("daysActive", ascending=False)
           [["advertiserName", "format", "daysActive", "transparencyUrl"]].head(5).to_string(index=False))
Enter fullscreen mode Exit fullscreen mode

What my run on 13 Nov 2026 (US) showed:

Domain Ads returned First shown in last 30 days Format mix (text / video / image) Running 180+ days
notion.com 125 (all) 7 79 / 32 / 14 22
hubspot.com 300 (cap) 22 216 / 74 / 10 196
shopify.com 300 (cap) 62 185 / 76 / 39 178

How to read it:

  • Long-running ads are probably winners. Nobody pays to keep a losing ad live for months. HubSpot's oldest still-active creatives were first shown in October 2021 (daysActive 1,846) and were still showing this week. Open those transparencyUrls first: that is the messaging they keep betting on.
  • Fresh ads show what they're testing now. Notion launched 3 new ads on 9 Nov; HubSpot's 22 new ads this month split 12 text, 9 video, 1 image, so they are pushing video harder than their overall mix.
  • Format mix shows where the budget goes. Roughly a quarter of HubSpot's and Shopify's recent ads are YouTube/video.

4. Get alerted when they launch something new

Set newAdsLastDays and the Actor returns only ads first shown in that window. Ads that are older but
still running are skipped, and you aren't charged for them. With HubSpot in the US and
"newAdsLastDays": 30, it returned 32 ads (all first shown on or after 15 Oct). That's more complete than
the 22 from step 3, which only looked at the 300 most recently shown ads.

Because the window is relative ("last N days"), the input never changes, so a plain Apify Schedule works:

{
  "queries": ["notion.com", "hubspot.com", "shopify.com"],
  "region": "US",
  "newAdsLastDays": 1,
  "maxAdsPerAdvertiser": 500
}
Enter fullscreen mode Exit fullscreen mode

Save that as a task, add a daily Schedule, and attach a Slack or email integration (or a webhook) to
"run succeeded". If you'd rather handle the results in your own code:

# pip install apify-client
from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("lindenbyte/google-ads-transparency-scraper").call(run_input={
    "queries": ["notion.com", "hubspot.com"],
    "region": "US",
    "newAdsLastDays": 1,
})
for ad in client.dataset(run.default_dataset_id).iterate_items():
    print(f'{ad["advertiserName"]} launched a {ad["format"]} ad: {ad["transparencyUrl"]}')
Enter fullscreen mode Exit fullscreen mode

(onlyNewSince is different: it keeps ads last shown after a date, i.e. everything still running. Use
it for "what is live right now", not for alerts.)

Limits (honest ones)

  • Google's text format bucket also contains image-rendered ads: in my HubSpot sample, 210 of 216 "text" ads had an imageUrl. Treat format as Google's filter label, not a strict ad type.
  • No ad copy extraction yet: you get the image/preview link and Google's page, not the headline text as a field.
  • No per-region impression stats or audience targeting details; firstShown/lastShown are the dates Google publishes.
  • Very large advertisers (Shopify) can have more live ads than a sensible cap; filter by format/region.

Cost: $0.001 per ad plus $0.003 per run, so the 125-ad Notion pull is about $0.13. A daily
newAdsLastDays: 1 check only pays for the few ads launched that day: HubSpot's 32 new US ads in 30 days
come to about $0.12 a month, plus $0.09 for the 30 run fees.


Not affiliated with Google. Data comes from Google's public Ads Transparency Center; it is advertiser
business data, not personal data.

Top comments (0)