Category data for market research
How many websites sell pet supplies in Spain? Is the number of AI tool sites still growing? Website category data turns questions like these into counts you can compare across countries and over time. This guide covers the method and the caveats that keep the numbers honest.
Questions website counts can answer
Almost every business, publisher and organization has a website. Counting categorized sites gives a fast, repeatable view of a market's web footprint, which is a useful proxy for how many players are active in a field.
How big is a niche?
Count active sites in a category and country to estimate how many businesses or publishers serve a niche.
Which country is most crowded?
Compare the share of each country's sites that fall in a category to find saturated and underserved markets.
Is a field growing?
Track new domains entering a category each month to see whether a field is attracting new players.
Who are the leaders?
Rank sites in a category by popularity band to see which few names dominate attention.
How fragmented is it?
Compare the number of popular sites with the long tail to judge whether a market is concentrated or spread out.
Where do topics overlap?
Sites with several categories show where fields meet, such as health and fitness or finance and technology.
The shape of the categorized web
Before sizing a niche, it helps to know the overall shape. These shares come from the primary web filtering category of each domain in our database, which covers over 120 million domains.
| Primary category | Share of domains | What it tells a researcher |
|---|---|---|
| Computers & Technology | 17.7% | The largest group, from software firms to IT services and tech blogs. |
| Business | 14.1% | Company sites for a wide range of services, often needing IAB detail to split. |
| Arts & Entertainment | 12.8% | Creative work, music, film, events and hobby sites. |
| Health & Medicine | 6.9% | Clinics, practitioners, wellness and health information. |
| Shopping | 6.7% | Online shops and retail sites across every product type. |
| Home & Garden | 5.5% | Trades, home services, interiors and gardening. |
| Financial Services | 3.0% | Advisers, lenders, insurers and finance information. |
| Automotive & Transportation | 2.9% | Dealers, repair, logistics and transport services. |
| Real Estate | 2.5% | Agents, property listings and developers. |
| Travel | 2.4% | Hotels, tours, travel agents and destination guides. |
Multi category sites: about 9% of domains carry more than one category, ranked by how much of the site each describes. Decide early whether your counts use the primary category only or every category a site carries.
A method for sizing a market
A repeatable sequence that research teams, consultants and investors can follow. Write each choice down, because later comparisons depend on applying it the same way.
Define the market
Write the market in plain words, then map it to IAB categories and web filtering categories from the taxonomy.
Choose the geography
Filter by country, and decide how to treat sites with no country attributed.
Clean the base
Remove parked, for-sale and malicious domains, which inflate counts in commercial niches.
Pick the counting rule
Primary category only, or any category. Primary is stricter, any is broader.
Split by popularity
Separate sites in popularity bands from the long tail, so leaders and small players are visible.
Sample and check
Open a random sample of sites to check that the count reflects the market you meant.
Combine with other data
Pair counts with company registers, surveys or financial data before drawing revenue conclusions.
Repeat on a schedule
Rerun monthly or quarterly with the same rules to follow change over time.
Worked example: pet supplies in two markets
An invented study shows how the steps fit together. A retailer wants to know whether online pet supply competition is heavier in one market than another.
How the analyst approaches it
- Map the market to Pets & Animals plus Shopping, and to the pet supply categories in the IAB taxonomy.
- Filter both countries, keeping sites whose country is attributed.
- Remove parked and for-sale names, which are common in keyword pet niches.
- Count sites that carry both a pet and a shopping category, to focus on shops rather than pet blogs.
- Express the count as a share of each country's shopping sites, so the two markets compare fairly.
- Split each count into sites with a country popularity band and long tail sites.
- Sample fifty sites in each market to confirm they really sell pet supplies.
Result in plain words: the analyst can say which market has more pet shops relative to its online retail base, and whether a few large names or many small shops dominate. That is a sound input to a market entry decision, not the whole answer.
Caveats that change the numbers
Website counts are honest only when their limits are stated. Each caveat below comes with a practical fix.
One company, many domains
Brands run country sites, campaign sites and old names that redirect.
Fix: count registered domains with content, and group known brand families where it matters.
Not every firm has a site
Small trades and local services may rely on marketplaces or social profiles.
Fix: treat counts as web footprint, and check offline heavy sectors against registers.
Parked names inflate niches
Keyword rich names in hot fields are often parked or for sale.
Fix: remove the parked category first, as covered in parked domain detection.
Country attribution varies
Markets with strong local domain endings are easier to attribute than those on generic endings.
Fix: compare shares within countries, not raw counts between them.
Broad categories hide niches
A top level category like Business contains very different kinds of firms.
Fix: use IAB tier 2 to tier 4 for niches, and sample to confirm.
Counts are not revenue
Ten thousand small sites may earn less than one market leader.
Fix: weight by popularity band, and combine with financial sources.
Which fields to use for which question
Each field in a domain row answers a different research question. Matching them up front avoids building a study on the wrong measure, and keeps the final report focused on what the data can actually show.
| Research question | Main field | Supporting fields |
|---|---|---|
| How many players in a niche? | IAB tier 2 to tier 4 category | Country, parked flag |
| How crowded is a market? | Category share within country | Popularity bands |
| Who leads attention? | Country popularity band | Global popularity band |
| Where do fields overlap? | Multiple ranked categories | IAB detail |
| Who is the audience? | Personas | Category, country |
| Is the field growing? | Newly registered domains by category | Monthly snapshots |
Personas describe the likely audience of a site in plain roles. Our guide to website personas explains how researchers use them, and website language and country data covers the geographic fields.
Tracking change over time
A single count is a snapshot. The more interesting story is usually the trend: how many sites enter a field, leave it, or change topic.
New domains by category
About 300,000 newly registered domains are categorized every day. Counting them by category shows where new activity is going.
Sites that go dark
Domains that lapse or become parked between snapshots show contraction in a field.
Category changes
Sites that pivot, for example adding AI tools, show how a field is changing from within.
Delayed launches
New domains are checked again after a week and a month, so slow launches are not missed.
Keep the method fixed: if you change the category mapping or filters between snapshots, the trend will reflect your method rather than the market. Version your rules alongside your results.
Worked example: is a new field still growing?
A second invented study, this time about change. An investor wants to know whether the number of AI tool websites is still rising, and in which countries. A single total would not answer that, so the analyst builds a monthly series instead.
How the analyst approaches it
- Take the AI / LLM Tools category, plus the more detailed AI tool categories where finer splits are needed.
- Count newly registered domains entering the category each month, after their week and month rechecks.
- Count existing domains that move into the category, such as software firms adding AI products.
- Count domains that leave: lapsed, parked or switched to another topic.
- Split entries and exits by country, and by whether sites reach a popularity band.
- Plot net change per month with the same rules every time.
Reading the result: rising entries with flat popularity suggest a crowded long tail. Rising entries with more sites reaching popularity bands suggest real adoption. The difference matters more to an investor than the raw total.
Checklist before you publish a figure
Run through this list before a count goes into a report, a pitch deck or a press release. Each item catches a mistake we see often.
Parked names removed?
Commercial niches are full of parked keyword names. Leaving them in can inflate a count noticeably.
Counting rule stated?
Primary category or any category gives different totals. Readers need to know which you used.
Sample checked?
Opening a random sample of sites is the fastest way to catch a mapping that captured the wrong things.
Comparison fair?
Cross country comparisons should use shares. Over time comparisons should use the same rules.
Precision sensible?
Round to a level the method supports. "About 12,000" is more honest than "12,347".
Proxy named?
Say "active websites" or "web footprint", not market size in money, unless you combined sources.
Who uses category data for research
The same counts serve very different readers. These are the groups we see most often, from quick screening work to long running studies that repeat the same count every quarter.
Investors and analysts
Screening sectors and countries for activity before deeper due diligence, and tracking whether a hot field keeps growing.
Consultancies
Fast market maps for client work, combined with interviews and financial data.
Sales and marketing teams
Territory sizing and total addressable market estimates, see lead generation and sales intelligence.
Academic researchers
Studies of the web's structure, digital adoption by sector, and regional differences. See the website classification dataset.
Public bodies
Measuring digital presence of sectors and regions for policy work, without personal data.
Product teams
Sizing the number of potential customers that run websites in a target vertical.
Media and publishers
Data stories about how the web is changing, such as which topics grow fastest in a country, told without personal data.
Trade associations
Tracking the online presence of their sector over time, and comparing it with neighbouring countries.
Franchise and retail planners
Gauging how many local competitors run websites in each region before expanding into it.
Presenting results responsibly
Clear presentation protects both the analyst and the reader. A few habits go a long way, and they make it easy for others to repeat or challenge your numbers later.
Say what was counted
State the categories, countries, filters and counting rule in one short paragraph next to every chart.
Prefer shares and ranges
Shares travel better between countries than raw counts. Round figures to show the right level of precision.
Date every number
Web data changes monthly. A count without a date is hard to reuse or compare.
Name the proxy
Call it web footprint or number of active sites, not market value, unless you have combined it with revenue data.
Quick glossary for readers of your study
Web footprint
The number of active, categorized websites in a field and place. A proxy for how many players are present online.
Primary category
The category that describes most of a site. Counting by primary category gives each site exactly one home.
Long tail
The many small sites with few visitors each. Usually most of the count, but a small share of attention.
Popularity band
A range placing a site among the most visited globally or in one country, used to weight counts.
Snapshot
A dated copy of the data used for one count. Comparing snapshots with fixed rules shows change.
Net change
Sites entering a category minus sites leaving it over a period. The clearest single measure of growth.
IAB tiers
Levels of the IAB Content Taxonomy, from broad tier 1 topics to narrow tier 4 niches.
Persona
A plain role describing the likely audience of a site, such as a financial analyst or a home office worker.
Getting the data into your tools
Research work is easiest with a local copy. The database is delivered as files that load straight into spreadsheets for small cuts, or into data warehouses and notebooks for full counts.
CSV, Parquet, JSON
Formats that load into common analysis tools. See download formats.
Category and country cuts
Ready lists per category and country, see database of websites by category.
Lookup API included
Check single domains while you sample and validate a count.
Browse before buying
Use explore the database to see how categories and fields look.
Frequently asked questions
Can you size a market by counting websites?
You can measure its web footprint: how many sites are active in a category and country, and how that changes. It is a useful proxy for the number of players, not revenue, and works best combined with other sources.
What are the main caveats of website counts?
One company can run many domains, many small firms have no site, parked names inflate counts, and country attribution varies by market. Clean and weight counts before drawing conclusions.
Which taxonomy is better for market research?
IAB categories give fine detail down to tier 4; the 59 web filtering categories give broad, stable groups. See the IAB categories list for the topic tree.
How do I compare countries fairly?
Compare shares within each country rather than raw counts. Raw totals depend on how many sites each country has and how easily they are attributed.
Can I track a market over time?
Yes. Monthly refreshes and daily new domains let you follow how many sites enter, leave or change category. Keep your method fixed between snapshots.
Does the data contain personal information?
No. Each row describes a domain: categories, country and popularity bands. Research at website level does not need data about individual people.
Count the market, not the noise
Categories, country, personas and popularity bands for over 120 million domains, with parked and risky names labelled so you can remove them. Files plus a lookup API included.
IAB is a trademark of the Interactive Advertising Bureau. It is named here only to identify the content taxonomy discussed. URL Categorization Database is not affiliated with, sponsored or endorsed by the Interactive Advertising Bureau.