Blog | Registrant analysis of the .ie domain

.ie domains

.IE Tech

Data and analytics

by Priyadarshini Tiwari

15 Jan 2025

Introduction

A domain registrant is the individual, organization, or entity that registers a domain name, effectively reserving a unique identifier for their use.

This registration designates them as the responsible party for the domain’s purpose and content. Registrants include individuals, businesses, non-profits, educational institutions, government bodies, and community or sports organisations.

For instance, individuals may register domains for personal blogs, businesses for branding and e-commerce, nonprofits for public causes, and educational or government entities for academic resources and public services.

In .ie , registrants currently classify themselves at the time of registration, meaning the data is self-declared rather than system-assigned. While this approach offers flexibility, it can potentially lead to inconsistencies that complicate data validation and analysis.

Such inconsistencies are particularly significant in light of new regulations like the European Union’s NIS2 directive, emphasising the importance of accurate and standardised registrant data.

To address this, we’re leveraging refined categories such as Statutory/Governing Bodies, Educational Institutions, Community/Union/Charity, and Sports Clubs. These categories aim to provide sharper insights into registration trends, support compliance efforts, and enhance data reliability. For those interested in the details of implementing this approach, check out our blog on predictive modelling for registrant classification.

Our Registrant Classification model has been crunching numbers and predicting registrant types based on the available data. However, there are noticeable differences between what the model predicts and what is self-declared by registrants.

These discrepancies highlight the challenges in refining the system to achieve more accurate classifications, but they also present an opportunity to enhance the data’s reliability and compliance with emerging standards.

Data Overview

The dataset used for this analysis contains information about the registrant type (such as CHA, COM, or OTH), the label predicted by the machine learning model (rant_label), and the total number of domains (domain_count) associated with each combination.

Breakdown of Registrant Types and Model Predictions

Let’s understand the various self-declared registrant types in the dataset:

CHA: Represents entities registered as charities, typically non-profit organisations focused on public benefit activities.
COM: Refers to companies or other commercial entities, primarily business-oriented organisations engaged in trade, services, or commerce.
OTH: This category typically includes entities that don’t neatly fit into the other classifications. It might encompass a natural person whose classification is ambiguous or unclear.

On the other hand, the model classifies entities into three categories:

Natural Person – Domains registered by a person.
Company – Domains registered by a company or organisation.
Others – Registrant that cannot be attributed to a specific person or company, also includes Charity domains.

Here’s a breakdown of Registrant Type and Registrant Label from your provided data, summarised in percentages:

The registrant types are categorised as COM (Companies), OTH (Others), and CHA (Charities). Below is the percentage distribution based on domain count, with data collected as of 2024-08-01:

COM (Companies):
Total Domain Count: 157,312
Percentage: 48.1% of total domains
OTH (Others):
Total Domain Count: 162,129
Percentage: 49.6% of total domains
CHA (Charities):
Total Domain Count: 7,534
Percentage: 2.3% of total domains

Note: The total domain count adds up to 326,975.

The labels used here are Company, Natural_Person, and Others. Below is the percentage distribution for each label:

Company:
Total Domain Count: 151,956
Percentage: 46.5% of total domains
Natural_Person:
Total Domain Count: 147,471
Percentage: 45.1% of total domains
Others:
Total Domain Count: 30,547
Percentage: 8.4% of total domains

Note: The total domain count also sums up to 326,975.

Non-Matching Registrant Labels and Registrant Types

The category with the highest domain count is Company under the COM (Commercial) rant type, with 147,438 domains registered, accounting for 96.99% of the total. A smaller portion, 4,579 domains (3.01%) are categorized as not matched.
The OTH (Other) registrant type dominates the Natural Person category, with 145,495 domains matched, representing a significant 98.6% of the total. In contrast, only 2,065 domains (1.4%) fall into the not matched category.
The Others registrant label reflects a diverse distribution across registrant types. While 21.29% of the domains (5,865) are classified as matched, the vast majority, 78.71% or 21,681 domains, are categorized as not matched.

The predictive model versus self-declared registrant

We identified several instances where the Registrant Labels did not align with their corresponding Registrant Types. By quantifying these mismatches in percentage terms, we can gain insights into potential inconsistencies and compare them against the expected error rate of our predictive model.

Company Mismatches:

CHA Registrant Type: 1,396 registrants are categorized as Companies under the Charity (CHA) category. This accounts for 1.16%, which is significantly below the model’s expected error rate of 5%. This suggests the model performs well, with the potential to investigate specific mismatches.
OTH Registrant Type: 3,122 registrants are classified as Companies under the Other (OTH) category. This represents 2.58% and remains below the 5% threshold. The presence of Companies in the “Others” category suggests potential inconsistencies in registrant self-declaration that may require further review.

Other Mismatches:

CHA Registrant Type: 5,866 registrants are categorized as Others under the Charity (CHA) category, representing 4.81%. This percentage is close to the 5% benchmark, suggesting the model performs reasonably well. However, the significant number of self-classifications under “Others” raises questions about registration clarity and criteria.
COM Registrant Type: There are 8,067 registrants listed as Others under the Commercial (COM) category, accounting for 6.54%, which exceeds the 5% threshold and warrants further investigation. Many of these registrations date back to before 2019, when the new registry system was introduced. During data migration, essential information like valid CRO (Company Registration Office) or RBN (Registered Business Name) numbers may not have been carried over, potentially leading to errors in categorisation.

Does the registrant influence domain usage and security stance?

Back in 2023, our team introduced the .ie Domain Scorecard, a measure reflecting the adoption of modern security standards for domain names. According to the scorecard, a domain can receive a grade from A (best) to F (worst). Let’s take a look at the different Registrant Labels that have different security scores.

The dataset represents domains categorised by their Registrant Labels (Company, Natural_Person, Others) and their corresponding domain score. With new and reasonably accurate knowledge at hand, we aim to uncover patterns that can help improve the overall health and trustworthiness of the domain ecosystem.

Score Distribution Across Registrant Labels

The analysis reveals that the majority of scores for all RANT labels—Company, Natural_Person, and Others—fall within the lowest score group (F), following a consistent pattern across the database. Specifically:

Company domains have their majority scores at 58.95% as F, aligning closely with the database-wide trend.
Natural_Person domains show a similar majority score concentration at 54.45% as F.
Others domains primarily hold scores around 53.64% as F, mirroring the overall distribution.

There are no significant differences between registrant labels, but domains registered by Others and Natural Persons show slightly better security scores than domains registered by companies. This is a surprising finding, as it’s expected that companies would use their domains for business purposes, requiring them to have higher security standards to deal with the risks posed by being online.

Domain Usage

Since 2021, we have crawled and categorised every single .ie domain into using six categories, as explained in this blog post. The categories used to indicate the level of web usage are:

Unknown: No web page was fetched, due to a variety of errors including DNS failures.
Not used: The page obtained represents a default web page for a specific type of software
Blocked: The page indicates this is a suspended account
Parked: A holding page was returned
Upcoming: Under construction or CMS default page
High Content: The page has some content, but it doesn’t match any of the low content categories above

The distribution of domains using the Registrant Label and the Web Usage is in the figure below

General Observations:

High Content dominates across all Registrant Labels.
Parked Domains are moderately represented for Natural_Person and Company.
Blocked Domains and Upcoming Categories are rare for all labels.
Unknown Usage is particularly notable for Company (45.62%) and Natural_Person (38.82%), accounting for a significant portion of domains where the usage cannot be estimated. This is typically due to the web page being non-functional. The 7-point difference in Unknown Usage between these two categories highlights a meaningful distinction.

Conclusion

The analysis of domain registrants and their classifications has uncovered insights, particularly regarding the mis-classification of RANT and the need for refined categorisation in the RANT Classification model.

Key Findings

Mis-classification: One of the most pressing issues identified is the mis-classification under the Others (OTH) category. The model has shown a tendency to incorrectly tag these registrants as Companies (COM), leading to a mismatch percentage of 6.54%. This discrepancy exceeds the expected error rate of 5% for the predictive model, suggesting challenges in distinguishing between registrants with diverse characteristics that fall under the broad “Others” category and those fitting the profile of companies.
Diverse Registrant Types: The analysis categorized registrants into three primary buckets: Companies, Charities, and Others. Within the Others category, sole traders are often overlooked, highlighting a gap in the model’s ability to accurately classify these individuals. This not only affects the reliability of the data but also obscures the actual landscape of domain ownership.
Need for New Categories: The introduction of new registrant categories, such as Statutory/Governing Bodies, Educational Institutions, Community/Union/Charity, and Sports Clubs, was initially considered to provide more granular classifications. However, as detailed in Vindhya’s blog, these categories were not fully implemented due to limitations in training data and practical challenges. While the idea of enhanced classification remains valuable for gaining deeper insights and ensuring regulatory compliance (e.g., with EU NIS2), the current focus is on refining the existing classification framework for accuracy and reliability.

In summary, while the current classification system provides a foundational understanding of domain registrants, it falls short in accurately representing sole traders. By addressing these gaps through improved categorisation and data collection practices, we can enhance the reliability of the RANT Classification model and ensure that it meets the evolving needs of the domain registration landscape. By prioritizing the accurate representation of sole traders, we not only improve data integrity but also contribute to more effective compliance with regulatory standards, ultimately fostering a clearer and more trustworthy domain registration environment.

What is .ie

Get your domain

Manage your domain

Domain tools