Targeting Options In Programmatic Advertising

The purpose of a programmatic purchase is to automatically buy individual ad views from people who are most inclined to become my clients. This article describes the targeting options available on this market that produce the best effects in this ad purchase model.

Programmatic Advertising – Targeting Options

There are many available options for targeting ad purchases; when running campaigns, usually at least a few of them are combined. It is important to select them for the target group sought, and it is best to try to describe each target group in several different ways. This will enable you to compare selected targeting methods.

Targeting options are available on purchasing platforms (DSP) and they are set at the level of planning the subsequent campaign elements. For example, targeting in Display & Video 360 can be set at the level of the entire campaign, at the level of individual insertion orders, or at the level of individual line items.

It should be noted that new line items inherit the targeting settings assigned to the insertion order (and a new insertion order inherits the campaign setting); however, a change made at a higher level during a campaign issue is not automatically transferred lower down the hierarchy of a given campaign.

It is impossible to discuss all available targeting methods in this article, so we will be focusing only on the most popular and interesting ones.

Geotargeting or Where Is My Client?

It is obvious that, when running a business locally, we need to reach people with advertising on this local market. This is equally true for a hairdresser’s, a pet shop, or an electronics store processing orders only within the country. For the vast majority of campaigns, some kind of geotargeting is set. This is usually targeting a given country, less often to a province or a city; it depends on the particular campaign.

Geotargeting is based on the IP addresses of computers and telephones, and it is assigned to relevant cities, provinces, and countries. Part of the available inventory (mobile applications) can be targeted more accurately, e.g. within a certain radius from a given shopping center.

Technicalities, or How We Reach the Customer

If I know that my advertising campaign converts best on desktop computers, it is practical for me to exclude users browsing the internet on phones and tablets. In this case, I can forgo advertising not only in the environment of mobile applications but simply on all devices of this kind.

If we want to advertise a mobile application that works only on one operating system, we will exclude the others. We would broadcast mobile video advertising only to users with a high-speed 4G internet connection because the others would not want to wait for the advertising spot to load. We would present a new model of a mobile phone to people who are already using similar devices from other manufacturers, for example, within the same price range.

Content and Context

In a certain way, targeting in the programmatic market works similarly to targeting ads on the GDN network – it is based on the content presented on the pages.

This solution works especially well when we want our advertising to reach some niche websites or content; for example, discussion forums on specific topics. It also works when we value the quality of users on a given website, but we do not want to appear on all its subpages – only on those whose topics are most relevant.

Contextual targeting, i.e. based on keywords of the page content, gives some approximation as to the interests of the person who will be shown the ad. A person who is reading an article about diapers may potentially be interested in advertising of diapers or other baby products.

On the other hand, user-generated content, including comments under articles, is a weak point of this kind of targeting. After all, the word, “diaper” could be used in a comment written by a person referring to the situation described in an article on automotive or political topics.

Targeting based on keywords is often used to exclude content that we do not want to appear on. For example, many advertisers use words that are commonly considered offensive, or words that refer to tragedies or natural disasters.

On the other hand, it is not practical to exclude very general words – for example the word, “politics” – because this will exclude all websites that, somewhere in their navigation menu, have a link to a thematic section of the same name.

Good Climate and Good Weather

An interesting set of data to which campaigns can be targeted is weather data. They mainly refer to weather phenomena (rain, snow, wind), temperature (also sensible temperature) and warnings against weather phenomena (e.g., heatwave).

Interestingly, they also contain information about influenza occurrences. They can be used in case of advertising portable air conditioners or winter tires. Or when we want to encourage users to stay at home and watch a TV series while enjoying a box of ice cream or a packet of chips.

Audience Data, or What We Know About Users

It was my intention to leave the most interesting targeting methods, i.e. user data, to the end. Based on this data, we are able to achieve the best results in programmatic campaigns.

Data on shoppers’ intentions are very valuable for all performance campaigns. It is especially useful to show a new phone model specifically to people who are looking for a new mobile phone right now. The same principle applies in the case of a washing machine, renovation services, a car, or a trip to Paris.

When we want to build brand recognition, it is worth considering reaching out to people who are interested in our branch of industry in the first place.

When the profiles available on the purchasing platform are not enough, it is worth considering creating one’s own. I do not mean retargeting, but creating a custom audience based on pages visited or where passwords were entered by users.

These may be specific product names, specific problems (e.g. “broken iPhone screen”) or specific pages of our competitors. After all, a person who has broken the screen on the phone may not only be willing to order service for its replacement, but may also consider buying a new phone.

Data Collected by Publishers

A separate issue is user data collected by publishers, which they can use in their sales platforms (SSP).

As a buyer, we only have access to them in the form of private deals issued by publishers. If the publisher knows that a given group of users are interested in motoring (because they are searching a lot of articles on this subject) or in buying a used car (because they are basically browsing only this kind of offer on a classifieds website), he may provide these data to buyers for targeting in the form of deals.

How to Use Programmatic Targeting Effectively?

The variety of targeting data available encourages experimenting, i.e. combining individual profiles and narrowing emissions deeper and deeper.

Although such deeper targeting allows better matching of the message to the interests of the user, it also reduces the potential for page views to buy. Therefore, it does not always translate into improved results.