From: gregisenberg
Frey, an “online directory King”, teaches how to build an online directory that can generate between 10,000 a month in passive income, requiring only about 15 minutes of work per week [00:00:08]. His first directory, built in October 2022, has been monetized for 18 months, is Evergreen and location-based, and generated around $2,300 in revenue in a recent month [00:01:27]. While AI tools can be recommended for certain parts, they are not strictly necessary to get started [00:02:25].
The process of starting an online directory includes finding and validating a niche, optimizing it for SEO, gathering and cleaning data, enriching it, and then implementing it on a content management system (CMS) like WordPress [00:02:45].
Getting Data for Directories
The primary method for acquiring data for directories is web scraping [00:27:41].
Web Scraping Tools
Frey uses OutsScraper, specifically its Google Map scraper [00:27:47]. Before scraping, it’s recommended to go to Google Maps and identify actual Google categories for the niche, as this makes scraping much easier [00:28:02]. For example, for “dog parks near me,” a dedicated “dog park” category exists [00:28:19].
If a specific Google category doesn’t exist, a general query can be used, but this will result in significantly more data, much of which will be “junk” [00:29:04]. Frey prefers creating Nationwide directories to take advantage of all search volume [00:32:24].
Cleaning Junk Data
When a plain query results in a large amount of irrelevant data (e.g., businesses with “dog” mentioned in reviews but not actual dog parks), several strategies can be employed for cleaning:
- Delete listings without addresses [00:30:05].
- Remove entries with no reviews or only one review [00:30:18].
- While AI tools like ChatGPT or Claude can help parse data, a solid prompt is essential [00:30:50]. Frey suggests manually removing large irrelevant entries first to prevent AI from discarding valuable listings [00:31:14].
Key Parameters to Scrape
When configuring the scraper, essential parameters to include are:
- Name [00:32:45]
- Phone number (especially for indoor parks) [00:32:47]
- Address (Street, Postal Code, State) [00:32:50]
- Reviews [00:32:53]
- Street view [00:33:01]
- Working hours [00:33:02]
- Location link (crucial for later data enrichment) [00:33:06]
Data Enrichment
Data enrichment is a critical step to make a directory more useful and comprehensive than existing competitors [00:35:22].
Identifying Desired Features
To enrich data, Frey analyzes Google Map reviews for recurring tags and features that users mention. For dog parks, common desired features include:
- Water fountains for dogs [00:36:51]
- Shade [00:36:55]
- Benches [00:36:58]
- Dog bags [00:37:02]
- Trash cans [00:37:05]
- Specific types like “indoor dog park” or “off-leash dog park” [00:37:13]
These features align with the “fragmented search intent” found during keyword research, where users are looking for specific types of amenities [00:15:47].
Manual vs. Automated Enrichment
Manually enriching data for thousands of listings is “God awful” and “extremely tedious” [00:38:20]. Frey’s first directory only had about 130 listings, making manual enrichment feasible [00:39:41].
For larger datasets, AI automation is key [00:39:49]. Frey is developing a tool that uses the location ID/Google URL to automate data enrichment by reading reviews and extracting information on specific features (e.g., “does this dog park have shade?“) [00:38:32]. This saves immense time compared to manual entry [00:39:35].
Benefits of Detailed Enrichment
Detailed data enrichment is what differentiates a new directory from dominant competitors, making it more useful and encouraging repeat visits [00:41:21]. For example, providing information on whether a dog park has dog bags or water fountains solves a real problem for users who otherwise would need to search reviews or call the park [00:41:33].
Implementation on a CMS
After data is cleaned and enriched, it can be formatted onto a CMS like WordPress [00:42:11]. While WordPress might seem “dumb” or basic for modern builds, it’s effective [00:42:38]. Other CMS options like Framer or Bolt.new are also viable [00:42:24].
This detailed data, especially with comprehensive features, forms the backbone of a static pillar page directory, which is highly effective for SEO by targeting high-volume location-based keywords [00:43:41].
The initial steps of finding an idea, validating it, and gathering/parsing data are consistent regardless of whether one aims to build a large or small directory [00:53:37].