How do we select the results to show the User?
DomainsBot Name Suggestion is the most widely adopted name suggestion tool on the market, providing relevant domain suggestions in 11 different languages across all gTLDs, nTLDs, and ccTLDs. Market leaders like Network Solutions, One.com, OpenSRS and Weebly rely on our service to help their clients find great available domains and drive more sales.
Because our tool has been created and updated over the last 15 years to analyze and engage with the three components of the domain name.
Our system is designed to identify accordances between the user’s query and the recommendation keywords found by the system itself: this means that our advanced semantic engine will read the user’s query and analyze it.
It finds which category or industry the search term belongs to, understanding if it’s connected to a geographical location, if it’s a proper name and so on. It will then find nouns or combinations of nouns that would be relevant and good quality in the domain name for the user.
Finally, the system matches the query of the user and the semantic suggestion found by itself to any possible TLD. Possibilities are almost endless, so to show just quality results, the system uses a minimum score that every combination has to reach in order to be proposed to the user.
Here’s an example
What happens when a user searches for “Manhattan Restaurant” and www.manhattanrestaurant.com/net/org are taken?
First of all, our system identifies the different pieces of the query, decomposing it in “Manhattan” and “restaurant” and then analyses them bit by bit. “Manhattan” is a proper name of a geographic location, while “restaurant” is a common name of a business connected to food. Now the system can start looking for relevant keywords connected to every piece of the query. It could be “New York” or “East Village” or “Chelsea” for the first word and “Deli”, “Burger” or “Pizzeria” for the second; they now have to be matched in relevant ways.
Once the semantic suggestions have been composed, it’s time to find relevant TLDs: it could be .nyc, .food, .restaurant, .menu and many more. For each one of these TLDs and Keywords, DomainsBot’s Name Suggestion calculates a “confidence score” based on the likeliness-to-convert and just the combinations which are over the specified minimum value will be displayed to the customer in search results.
Here the results: manhattanrestaurant.nyc , manhattancafe.com , manhattan-restaurant.food
Selecting the TLDs
The TLDs shown in the search results are, as said, chosen among the ones which are strictly connected to the user’s query, but also among some generics which belong to a defined special group.
The Specific TLDs
The system, which has previously analyzed the user’s query and found the category/categories it could be put into, will search the TLDs to consider inside the proper category.
The Generic TLDs
The second group, with generic TLDs, includes legacy TLDs – as for example .com, .net, .org, .info, .biz- a small selection of ccTLDs – as .co, .me, .us – and some nTLDs – such as .tech, .app.
To become part of the generic group, the TLDs must prove their ability to generate a relatively high amount of registrations across several different registrars and across different core verticals. Usually, vertical-specific TLD (such as .plumbing) and city TLD (such as .sydney) are not considered to enter the generic TLDs group.
Starting from this concept, our advanced system calculates a default score for each TLD in this group and updates it every week.
Calculating the Default Score
Sorry, that’s our secret ingredient.
What you need to know, though, is that you can edit the def_score parameter, so override the selection of generic TLDs and their default score, and making the Name Suggestion your own customized tool.
Now that you know how the magic happens