DEV Community

Md. Saquib Jawed
Md. Saquib Jawed

Posted on AI-assisted

I tested my website contact scraper against the most-used one on Apify (same 63 small-business sites)

Earlier I wrote about getting emails for Google Maps leads. Since then I improved my tool and wanted an honest answer to one question: is it actually better than what people already use?

So I ran both on the same list, on the same day, on Apify's cloud.

The setup

  • The list: 63 real small-business websites, the kind people pull from Google Maps: dentists in London, plumbers in Austin, marketing agencies in Pune, tax advisers in Munich, bakeries in Toronto.
  • The rival: the most-used contact scraper in the Apify Store, on its default settings (up to 20 pages per site, 2 links deep, contacts merged per site).
  • Mine: Company Contacts & Tech Stack Lookup, on its defaults (up to 5 pages per site: home, contact, about, imprint; plus the privacy page if no email turned up).
  • How I counted: a site "has an email" if at least one address came back. An email counts as "another organisation's" when it isn't on the company's domain and isn't a Gmail/Outlook-type inbox, for example a trade body, a template placeholder or the hosting company.

The results

Same 63 websites Mine Most-used contact scraper
Websites with an email 57 (90%) 51 (81%)
Websites with a phone number 54 (86%) 40 (63%)
Websites with a LinkedIn page 22 19
Emails belonging to other organisations 5 22
Run time 81 seconds 416 seconds
Cost of the run about $0.90 $1.92

Where the difference comes from

Fewer pages, chosen better. The rival read 958 pages for 63 sites, about 15 each, and charges per page. Mine reads the 3–5 pages where contacts actually live. It's faster, and the cost doesn't depend on how big a site is.

Finding the email where it hides. On 8 sites only my tool found an email. One was a London dental practice whose only address sits on its complaints and cookie pages; mine checks those pages when the main ones show no email.

Other organisations' emails. Most of the rival's 22 extra addresses came from legally required pages. German Impressum pages list the tax advisers' chamber (info@stbk-muc.de showed up on 5 different firms), the bar association and the arbitration board. Elsewhere there were template leftovers like info@mysite.com and filler@godaddy.com. Import those into a cold email tool and you're emailing the chamber of commerce. Mine keeps only addresses on the company's own domain, plus Gmail-type inboxes when the business uses one.

Phones. Mine checks every number against the country's numbering rules and reads labelled numbers ("Tel:", "Telefon:") as well as click-to-call links. That's where most of the 14-site gap comes from.

Where the rival is better

  • It found an email on 2 sites where mine didn't: one failed to load during my run, and one keeps its emails on a sister domain.
  • It can crawl deeper and use a full browser for sites that need JavaScript. Mine doesn't.
  • It has optional paid add-ons for finding named people at a company. Mine only returns what the company publishes.

Caveats

63 sites is a small sample, and each tool ran once. Results move a few points between runs because some sites are slow or block bots on some days. The site list and both result files are available if you want to check them; ask in the comments.

Disclosure: I built one of the two tools, and AI helped me write this post.

Top comments (0)