# URL classification for contextual advertising

Canonical: https://classify.ai/learn/url-classification-for-contextual-advertising

URL classification describes the subject of the content at an individual web
address. For contextual advertising, it helps you judge whether the page an ad
would appear beside fits the campaign.

A placement decision made for a whole domain treats every page on that site
the same way. The pages on a site are rarely the same.

## Pages on one domain serve different purposes

Take a regional publication. On the same domain you might find:

- A winter driving guide about tire chains and road conditions
- A guide to local buses and park and ride lots
- A roundup of restaurants open over a holiday weekend
- A report on a highway accident

An outdoor gear retailer, a transit service and a restaurant group might each
find one of these pages relevant and the others less so. An advertiser may
choose to avoid the accident report altogether, even though it shares a
publisher with the rest.

A label for the domain cannot separate these pages. A classification of each
URL can.

## What page level context adds

| Question                  | Domain label                 | URL classification                |
|---------------------------|------------------------------|-----------------------------------|
| What is this page about   | Assumed from the website     | Read from the content at that URL |
| Does it fit the campaign  | One answer for every page    | An answer for each page           |
| Should it be avoided      | The whole site or none of it | Only the pages that do not fit    |

Classify describes pages with categories from the IAB Content Taxonomy, a
shared vocabulary for content. A category such as Running and Jogging then
means the same thing to the advertiser, the publisher and any platform in
between.

## Page context is not visitor identity

Classifying a URL describes what is on the page. It does not tell you who is
reading it: not their identity, their demographics, their browsing history or
what they plan to buy.

Someone reading a winter driving guide may be planning a trip, researching an
article or simply curious. The page tells you what they are reading at that
moment, and that is what contextual advertising relies on. Treat it as a
measure of relevance, not as a prediction of a visitor’s intent.

## Using page context to choose placements

- Start from the campaign: the subjects your ads belong beside and the ones
  they should avoid
- Express those subjects as taxonomy categories, so the same rule applies
  across publishers
- Decide page by page, so one unsuitable page does not rule out a publisher
  and one relevant page does not approve all of it
- Decide what to do with a URL that has no classification yet; an unread or
  unmatched page is unknown, not unsuitable
- Review a sample of URLs and their content before applying a decision widely

The guide to page level contextual intelligence covers what else to check in a
page classification system:
https://classify.ai/learn/page-level-contextual-intelligence

## Explore contextual segments

In Classify Atlas, describe the audience you want to reach and browse related
contextual segments as a starting point. Segments, curation and partnerships
are arranged directly with Classify.

- [Classify Atlas](https://atlas.classify.ai)
- [Page level contextual intelligence](https://classify.ai/learn/page-level-contextual-intelligence)