Natural language processing & automated SEO for Kasaz

Automated SEO – NLP – Data Improvement

Client: Kasaz

Challenge: Improve the real estate listing data quality and increase their organic reach in Google.

8910

SEO pages created

35

extra real estate keywords (NLP)

80M

possible SEO page combinations

Who is Kasaz?

Kasaz is the real-estate portal of the future, the first truly buyer-focused property portal in mainland Europe. Where the status quo is duplicates, fake listings, and inaccurate information, Kasaz offers unparalleled transparency and listing quality combined with superior user experience.

Goals

  • Real estate data enrichment: one of the goals was to enrich the real estate listing data. Out of the descriptions we get extra real estate related keywords, for examples ‘pool’, ‘garage’, ‘airco’… This enables us to extract more features and to build more relevant long-tail SEO landing pages.
  • More organic reach and SEO improvement: Improve the organic reach on Google and get more visitors to the Kasaz website.

Solutions

How Natural language Processing enriches the listing data and improves user experience

A real estate listing description usually contains a lot of information, such as real estate related keywords. Words that aren’t available in the searchable features (metadata) can still be found in listing descriptions. Some examples of keywords are: renovated/not renovated, amount of bedrooms, pool, balcony, and many more.

Natural language processing real estate

We enrich the listing data quality of Kasaz by scanning their listing descriptions with Natural Language Processing algorithms. With those NLP algorithms, we find real estate related keywords such as terrace, open kitchen, bath, garage, and more. This information can be used to complete searchable features (metadata).

After a thorough analysis of Kasaz’s data, we concluded that structured data such as pricing and address were okay in general. But Co-libry noticed that 80% of real estate related keywords (such as ‘garage’, ‘balcony’) were not available in the searchable features yet. We were about to change that.

Here’s an example of how Natural language processing (NLP) finds keywords such as ‘garaje’, ‘ascensor’, ‘trastero’, and ‘aire_acondicionado’ (In English: garage, elevator, storage room, and air conditioning).

Some search engines can also find exact keyword matches. The power of Natural Language Processing goes beyond that. It detects the exact keywords and misspelled words, synonyms and negatives (for example: ‘no garage’).

In this example, our NLP engine looks for the word ‘ascensor’ (elevator) in the descriptions. And comes with much more results than a one-on-one match.

More organic reach and SEO improvements

The classic SEO approach with marketing agencies or in-house SEO has its limitations. Typically, the agency or team will perform a manual SEO analysis looking for high-value keywords that match user intent. But for an online marketplace with thousands of listings, the challenge lies in finding a scalable way to create qualitative pages for all the keywords while avoiding duplicate content.

Kasaz saw the potential to increase new user visits, and ultimately, revenue through dedicated landing pages focused on thousands of high-value long-tail keywords. This could only be done through a solution that automates the process of creating these landing pages.

They noticed that tons of Google searchers are looking for precise, long-tail keywords. Some examples are ‘4 bedroom house in Madrid with balcony’ or ‘2 bedroom apartment with swimming pool’.

Those keywords have lower competition and here’s why:

The best way to be the number 1 in Google is to create a specific landing page for certain keywords. For example: create a landing page containing only listings with four bedrooms + house + in Madrid + with balcony.

Almost none of the real estate marketplaces we analyzed in Spain are creating pages for these keywords, which is a big opportunity for Kasaz.

The power of Natural Language Processing and improving SEO

The previous chapter about Natural language processing was a crucial step in improving Kazas’ SEO.

We combined keywords like ‘city’ + ‘amount of bedrooms’ + ‘house/apartment’. Besides that, we also combined those with keywords found in the description of the listings.”

Step-by-step process: 

In the first phase, Co-libry built base combinations consisting of

  • a listing type (eg. ‘apartment’)
  • an offering type (eg. ‘for sale’)
  • a city (eg. ‘Madrid’)

These base combinations are then extended with (single or both)

  • a possible number of bedrooms (eg. ‘3’)
  • NLP characteristics (eg. ‘open kitchen’)

Combining all these, had a total of 80.196.375 combinations. Every combination is checked whether it contained at least 10 hits in the database (10 existing listings). These combinations were eventually selected to generate the sitemap.

In the end, 8910 URLs were generated and placed in the sitemap.

What does Co-libry precisely automate?

  • We automatically combine real estate keywords relevant to SEO pages. A combination usually looks like ‘house/apartment’ + ‘sale/rent’ + ‘city’ + ‘amount of bedrooms’ + ‘real estate keyword’.
  • We create certain parameters so our client can create an H1 header, meta-title and meta-description.We show the relevant listings, based on the url, in the pre-build website templates.”

REQUEST YOUR DEMO

AI and personalization made easy

Start today!

REQUEST YOUR DEMO

AI and personalization made easy

Start today!