5 Programmatic Trends You May Have Missed
Facebook and GDPR have been dominating the headlines recently, but important topics and trends have slipped through the cracks. Keep up with the fast-moving programmatic advertising world with these 5 topics you may have missed.
1. Understanding How DSPs See Your Inventory
In advertising, many factors need to be considered to build and maintain brand reputation. Brand safety, viewability, fraud detection, blacklists, whitelists… the list goes on!
These are all ways in which ads that publishers distribute are filtered. Proactively managing inventory is pivotal if demand-side platforms and buyers are to be kept happy and spending money on publishers’ campaigns.
How can this be resolved?
Don’t always rely on third-parties for analyses. Publishers are encouraged to run their own analysis on metrics such as viewability and fraud detection. Advertisers can use that data instead.
2. A Masterclass in Private Marketplace (PMP)
High fill rates, the amount of demand that can be met by using immediate stock, with favorable yields for publishers and excellent outcomes for buyers is, of course, the desired result for all parties. All of which takes place in the Private Marketplace (PMP).
Right now, there seems to be a very common problem when it comes to advertisers and the Private Marketplace: advertisers aren’t able to buy the traffic they want, and publishers cannot negotiate and execute deals with PMPs efficiently enough – and no-one knows why. However, there does seem to be a solution.
To combat the difficulties that advertisers and PMPs face, cooperation down the whole line of Demand-Side Platforms, Supply-Side Platforms, Data Management Platforms, and publishers will be needed to solve PMP issues. This could be quite a stretch to achieve easily, but it would begin to solve the problems faced by all parties.
3. The Hidden Secrets Of Your Log Files
Log files in advertising, the data points collected from when a programmatic advert appears on your screen to when it leaves, are extremely important to both publishers and advertisers. This can be things such as viewability and estimated pricing of the ad.
Currently, reporting tools that give valuable insights to bid landscapes and trading patterns, for example, are quite limited. To benefit from the raw data available, the log files need to be downloaded from the platform and analyzed using differing tools.
If you manage to get through the process of analyzing the information, it’s still not in a presentable way to easily digest. The most relevant reports require some creativity but can add a lot of value.
4. Best Practices For Audience Extension
Audience extension, a behavioral targeting technique also known as “Lookalike Modeling,” is an efficient way for publishers to create incremental revenue by using their audience data.
Say your company sells clothing online. By adding a piece of tracking code to your sales confirmation page, you gain information on that user’s demographics and interests – your perfect lead!
But how can you maximize the insights you’ve now obtained? By using that data, you can target people with similar demographics and interests. In some cases, the lookalike audiences actually perform better than the original ones when measuring metrics like conversion rates.
By targeting users outside of their own media properties, publishers can create additional inventory or then they can model a lookalike audience using a data management platform. Advertisers can then reach the audience across other media properties, which are most commonly used for retargeting campaigns. It has now become more of a top priority as advertisers want to follow users across these other media properties, making it a combined cross-device retargeting and audience extension process.
This is something that has been in the works for quite a long time, but more recently we have seen issues with publishers performing audience extension, but not really making it clear for programmatic environments.
5. AI Essentials
AI became a major talking point in 2017 and is gaining huge traction in the programmatic world. It’s estimated that over 80% of digital ad spending will be done programmatically in 2018.
The key point here is that the first step in any AI project should be to conduct a thorough analysis of the data that is available, what the quality of that data is like, how it’s collected, and what data-related problems can you expect to solve with AI.
Building AI algorithms can take a substantial amount of time. Each year the process gets easier and faster, or then companies provide tools to hasten the process, yet without analyzing what you hope to achieve with the data you hope to obtain, it will ultimately delay your overall process.
What’s our take?
The programmatic world is moving faster and faster as more parties start to trade programmatically, either by taking it in-house or by working with specific agencies to achieve their goals. Remaining agile is an absolute must as well as keeping up-to-date with new methods.
The in-app environment is gaining traction, but by using multiple DSP-specific machine learning practices, all of the above points can be easily executed within the in-app space.