AppScout aggregates online business listings.

Several months ago, I picked up the habit of scouting out online businesses listed for sale. While most of the many millions of apps and websites that exist do not make money, some of them do. And some of those, occasionally go up for sale. Unlike browsing Zillow or for real estate, or BizBuySell for brick and mortar businesses, there’s nothing like an MLS for listing online businesses.

Instead, the market is quite fractured. There are many dozens of very small boutique brokerages, perhaps half a dozen well respected, higher-volume brokerages, several For-Sale-By-Owner type classified sites, and then there’s Flippa, which has the largest volume, but is oft-avoided by serious owners and investors due to the frequency of scams.

An investor in search of a deal either spends a significant amount of time scanning these sites for new listings, or has an assistant/employee who does that for them. Due to the size of these businesses for sale (many small, what you’d call “micro” startups), I think it’d be more often the former. Several sites offer newsletters that send out when new listings are posted, but these often lag behind.

I created AppScout to save time and give myself a competitive advantage should I ever seriously pursue purchasing an online business. While this is a tool for personal use, the backend can fully support multiple users, and I designed a landing page with custom illustrations:

The heart of the service is a feed of the latest listings. You can bookmark ones that interest you, add listings manually (and/or privately, if you’d prefer to keep an a deal secret), and as the admin I can edit listings to fix typos, populate metadata etc. Search and filtering tools on the right are not yet functional but were carefully thought out to include what an investor would require without welcoming feature creep or an overwhelming interface.

How does it run?

AppScout is a Python-Flask app with a Postgres database and currently runs locally. On demand, I ping its API to fetch new listings, though this could and would be automated with a cron job if I were to make the app more widely available.

When a fetch is requested, AppScout uses a combination of custom built BeautifulSoup and Headless Selenium scrapers to fetch listings from each of several dozen brokers. Old listings are ignored, and only unseen listings are collected. In turn, each of these listings is visited at their individual pages and more specific data is collected, such as a full descriptions, and revenue details.