Show HN: Script that finds local businesses with bad websites and cold emails
jumbotronic.gumroad.comI built this because I was trying to get freelance web dev clients. The standard approach (manually Googling, clicking each site, finding an email) was taking hours per dozen leads.
What it does: - Takes a city + business type (e.g. "Las Vegas plumbers") - Searches for local businesses via DDG - Visits each site and runs a DOM audit: mobile viewport, SSL, page speed, meta tags, outdated copyright, missing favicons - Scores each site 0-100, filters to only the bad ones (score >= 40 = real fixable problems) - Tries to extract a contact email from the site - Generates a personalized cold email referencing the specific issues found - Outputs a CSV with business name, URL, score, issues, email, and draft pitch
Stack: Python 3.9, requests, BeautifulSoup, no external APIs needed.
I tested it on Las Vegas nail salons and cleaning companies. First run: 22 emails sent, 6 replies. It finds real problems because it actually visits the site rather than using a database.
Would love feedback on the DOM audit heuristics - currently scoring on ~13 signals. Is website quality actually a good proxy for "willing to pay for a redesign"? This is a clever script! Regarding your question on heuristics: technical errors (SSL/Speed) are good, but I've found conversion friction is a much stronger proxy for 'willingness to pay.' Business owners care less about a missing favicon and more about 'why users aren't clicking the Call button.' I actually built something that automates the 'conversion audit' side of this. It analyzes the page and generates a prioritized improvement plan. It might be a great 'high-value' attachment to those cold emails to prove you've done deeper thinking than a standard bot.
Happy to let you run a few of your Las Vegas leads through it to see if the pitch resonates better! If you want you could run it and if possible just give me the feedback after or if you need any help with that. :) https://websiteroasters.com/