Programmatic advertising is a great way for advertisers to buy ad inventory without the hassle of dealing with individual publishers. DSPs who are serious about delivering device-related targeting and reporting need a reliable source of the most accurate and comprehensive device data that includes every web-enabled device potentially accessing the ad content.

What is a Demand Side Platform?

A Demand Side Platform (DSP) is a part of the advertising technology puzzle typically used by brands or agencies buying ad space. DSPs allow you to buy ad inventory from multiple ad exchanges (representing the publishers) at the same time using just one interface (bidding system) which is what makes it more efficient and cheaper.

The goal of DSPs is to show ads to as many users as possible based on the targeting criteria and the bid value set by the advertiser. More targeting criteria and better campaign reporting are extremely valuable to advertisers. They get the ability to target using any specific device characteristic, as well as getting a better understanding of how their ad dollars are spent.

Typical DSP use cases

Accurate and detailed data on all web-enabled devices that may access online content is an essential asset for all DSPs. With today’s device fragmentation (screen resolutions, phone makers, Android versions), no serious player in the Ad tech space can function without an excellent source of device information that can be accessed in real-time. Here’s what it’s typically used for:

• Campaign management

Device data makes it possible to populate campaign management interfaces used by advertisers to better define the campaign and enrich the targeting options with various device properties (characteristics). The more device properties in the database, the more accurate and detailed device-based targeting options can be offered in the campaign interface.

• Populating data from external sources

In some cases, ad exchanges may not populate all device-related fields in the bid request due to the lack of device data (a common problem for exchanges which don’t use a device detection solution). In this case, a DSP may use its own source of device data to enrich bid requests.

• Device information beyond the scope defined by the RTB protocol

The RTB protocol between an exchange and its bidders defines a number of data points that may be extended with device data to enable more sophisticated targeting by advertisers.

• Data Management Platform (DMP)

Device information is typically used to populate user profile fields stored at the DMP level. Its purpose: to store user profiles, and track device information against the user profile. This is matched against bid requests from the ad exchange, since the bid request carries the User-Agent string.

• Selecting the right creative for the device

With thousands of different screen sizes, software versions, and connectivity levels, it is extremely important to make sure that the ads look and work great on the user's device. If the DSP includes an ad server, a device detection solution can be used to identify the optimum size of creative according to the device characteristics to make sure it displays well on the target device (within the bounds of the ad format mandated by the bid request).

• Consistency in device naming

Device information in the User-Agent (UA) string is typically just a model name that may be difficult to understand. A good source of device data makes it possible to enrich reporting to the advertiser by providing consistent, consumer-friendly device naming. For example, all Galaxy S5 models are grouped into one bucket (each variation of a phone comes with a whole set of User-Agent strings).

Why device detection can be problematic

There are numerous reasons why building your own device detection solution for this purpose may not be such a good idea which we listed in this blog post. Here is just one example which shows the importance of getting device detection right: detecting iPhone models.

Device detection is done by analyzing User-Agent strings which typically include information on the visiting device. The information can be as detailed as the browser maker and the device manufacturer want it to be. In the case of Apple devices, iPhone model tokens are missing in the UAs which is what makes targeting particularly difficult. DeviceAtlas deals with this problem with a Client-side Component, a JavaScript file which is sent to the user’s device to retrieve some extra information such as the iPhone model.

This is what helped AOL’s ad server (formerly ADTECH), a DeviceAtlas client, to offer real-time recognition of Apple mobile and tablet devices. Christian Ruck, ADTECH Principal Product Manager:

"DeviceAtlas enables ADTECH to be first to identify requests from different Apple devices. This makes it possible to deliver the most optimal ad in real time in the right environment, increasing consumer engagement and delivering advertisers and publishers a higher return on investment. With the new technology, advertisers can precisely target advertising creatives and receive granular reporting, which was not previously possible due to Apple’s browser headers schema."