I scraped 570+ agency websites across 43 countries to find their contact emails. here's what i learned about global email accessibility.

two weeks into building my cold outreach pipeline, i've now scraped over 570 agency websites across 43 countries trying to extract contact emails. the results are wildly inconsistent by region, and i figured some of you might find the data interesting.

the method

dead simple python script: requests + regex. for each agency i try 3-4 URL paths (homepage, /contact, /contact-us, /en/contact) and extract anything matching an email pattern. then filter out false positives (sentry, cloudflare, schema.org, image filenames, generic placeholders).

total time to scrape one agency: about 3 seconds. total time to find agencies to scrape: that's the slow part.

hit rates by region (emails found / agencies attempted)

• australia/new zealand: ~75%. most agencies have emails on their contact page. straightforward.
• uk/ireland: ~70%. similar to AU/NZ. professional sites, emails usually visible.
• us/canada: ~60%. bigger agencies tend to hide behind contact forms. smaller ones still list emails.
• western europe (netherlands, germany, france, belgium): ~55%. many have contact forms only. some require clicking through to team pages.
• india/pakistan/bangladesh: ~65%. good hit rate. many agencies list multiple emails (info@, sales@, founders personal gmail).
• southeast asia (philippines, vietnam, thailand, myanmar, cambodia): ~70%. surprisingly high. lots of gmail addresses, which actually helps since they feel more personal.
• middle east (UAE, qatar, jordan): ~50%. mixed. some have emails, many redirect to whatsapp or contact forms.
• africa (kenya, nigeria, tanzania, uganda, ethiopia, senegal): ~55%. similar to middle east. smaller agencies tend to list emails, bigger ones hide behind forms.
• eastern europe (czech republic, romania, poland, hungary): ~60%. decent hit rate, professional sites.
• latin america (mexico, colombia, argentina, peru): ~45%. lowest hit rate. most agencies prefer contact forms or whatsapp.

the biggest surprises

1. gmail addresses are everywhere in developing markets. not as a negative — agencies using info@agency.com vs agency@gmail.com didn't correlate with response quality at all.

2. /contact-us works better than /contact in english-speaking countries. non-english countries often use localized paths (/contacto, /kontakt, /en/contact).

3. about 15% of "emails" i extract are actually image filenames or CSS class names that happen to match the regex pattern. filtering these out is critical.

4. the best emails come from team/about pages, not contact pages. founders' direct emails on team pages have a much higher open rate than generic info@ addresses.

5. some countries have almost zero scrape-able agencies (laos, myanmar smaller cities) while others are saturated (melbourne alone has 50+ agencies with visible emails).

the filter list that saves hours

after scraping 570+ sites, my false positive filter has grown to 25+ patterns: sentry, cloudflare, googleapis, schema.org, .png, .jpg, .webp, hr@, jobs@, careers@, example.com, and a dozen more. without this filter, about 30% of "emails" would be garbage.

what i'd do differently

i started by guessing agency URLs from search results. that wasted hours — guessed URLs return 0 emails about 80% of the time. now i always websearch first for the real URL, then scrape. much higher hit rate.

if you're building any kind of outreach pipeline, the scraping part is actually the easiest part. finding the RIGHT agencies to scrape and writing emails they'll actually open — that's where the real work is.

the whole scraping + sending pipeline runs on python, cron, and a $0/mo linux server. happy to share the approach if anyone wants details.