BIGGIES Winner: How the Homeowners Bot convert online readers to subscribers, using automatic real estate texts

Thursday, August 9th, 2018 Case Study Comments

Converting visitors on news sites to paying customers is today the focus of media houses around the world. By opening up a new digital revenue stream, alongside advertising, total digital revenue can increase. But that requires content that makes readers log in to the site and – preferably – become paying, recurring customers.

News about real estate business have a proven high reading among local newspaper readers and visitors to local media sites. Analyzes of how the content is used now also shows that real estate business news appeals to loyal, often paying visitors. (In addition, they are pure click magnets!) At the same time, it appears that almost no such texts are published on the local media site. (Case: MittMedia, Sweden)

The housing market has long been an important advertising category for the daily newspapers. Online, the media houses have had a hard time creating a relevant environment for offers from real estate agents. With a flow of automated real estate texts, such an environment can be created.

Bring real estate listings into the Internet age!

By automatically produce a unique text for every house sold, with images, maps and geographical data, we bring the old listings into the internet age. (United Robots operates in the same way withs sports results, stock market listings and new companies and bankruptcies.) Not only does the number of texts about real estate deals increase dramatically, in addition, the following happens:

  • We automatically finds and attaches to the text a high resolution Google Street View image, a satellite image and a map with coordinates.
  • Now also smaller news sites can launch their own property sites, as the automated real estate texts allow a constant flow of new content. (For a normal Swedish news site that means about 20 articles per weekday.)
  • Each property sale gets its own URL, which dramatically increases the number of pageviews.
  • Real estate texts have proven ability to convert readers to paying subscribers.
  • Interested in testing personalization or superlocal apps? This requires super-local materials, such as real estate sales. And it is a highly relevant content for readers in local communities.
  • When our smart algorithms find information that might be worth paying extra attention to, an Alert will be sent to the newsroom. For example when the most expensive property so far this year was sold, or the largest, or when the highest square meter price in an area was recorded, etc.

We analyze and compile – completely automatically

But there is more to gain from using United Robot’s analysis and natural language platform.

All property sales in a particular area (such as a county or municipality) are stored in our database and analyzed over time. Our platform keeps track of average prices, price trends, the most expensive hoses sold in the area and much more.

Fully automatically, our algorithms write an unlimited amount of summary texts and list the most expensive deals – a dull and time-consuming job, currently performed manually in most newsrooms.

List: The most expensive villas in Övikstrakten right now

The house prices in Övikstrakten continue to decline. Last month, prices fell by seven percent.

Four municipalities have downward curves. For example, prices fell by 14 percent in Härnösand and 12 percent in Kramfors. The average price for a villa in the region is SEK 830,000.

Here are the most expensive houses sold in Övikstrakten in the last month:

• Sjöbodvägen 141, SEK 4 000 000:-

• Mesanvägen 23, SEK 3 530 000:-

• Brunnshusgatan 34, SEK 3 000 000:-

• Frejgatan 53, SEK 2 950 000:-

• Blåhakevägen 21, SEK 2 950 000:-

• Hummelviksstrand 138, SEK 2 800 000:-

• Andra Långgatan 11, SEK 2 600 000:-

• Ulvö Hamngata 212, SEK 2 355 000:-

• Lövvik 202, SEK 2 350 000:-

• Klippvägen 14, SEK 2 325 000:-

Automatisk text från United Robots

Data source and method

The Swedish Mapping, Cadastral and Land Registration Authority registers changes of real estate ownership in Sweden. Similar authorities exist in most countries.

United Robots receives from the Authority about 500 data posts about changes in real estate ownership daily. The data is stored and analysed, to find the most relevant angle for a story about every property deal that happened in the country last day. After that United Robots Natural Language Generation plattform turns those findings into a finished text, an publishes it immediately to our customers news sites.

In addition to the text, we also use Machine Learning software to find the most relevant high resolution Street View pictures from Google, as well as a map and a satellite picture. Geo coordinates are also added.

Example of a story written by The Home Owners Bot, from MittMedia in Sweden (translated)

Results

Early measures and analysis shows that the real estate texts generates a massive amount of page views. The big Swedish local news publisher MittMedia has put the texts behind its paywall, and seen them generate hundreds of new paying customers. Another local media house has created a stand alone destination site about home ownership.

For United Robots the Homeowners Bot has been a success, as the four larges local media companies in Sweden has signed up for the service during its first six months.

Enter the 2018 BIGGIES EMEA competition, with 12 categories focused on Big Data- and AI-driven products and processes at media companies. Deadline is Sept. 7. Enter now!