Two weeks ago, we launched our new targeting solution Full Reach, a solution utilising both an ID-based and an ID-less model. In this blog post, you can learn more about what an ‘ID-less prediction’ is and how we can create audiences without cookies.
With the launch of Full Reach, we are enabling publishers to target an audience on every single impression. This by offering a targeting solution not only relying on one predictive model but a hybrid technology; specifically, an ID-based model, where historical data in connection to a device identifier is available and an ID-less one, where just a single impression is available.
The ID-less model is designed for the future, as it enables publishers to target an audience without relying on cookies, and only by looking at a page view logline.
The ID-less model: cookieless but effective
There are six different inputs that the ID-less model uses from a page view:
- Content (e.g. sports site)
- Device (e.g. mobile)
- Web browser (e.g. Chrome)
- Operating system (e.g. iOS)
- Geolocation (e.g. city)
- Time (e.g. 12.30 pm)
Each of this information tells something about the user, and depending on circumstances, some inputs may be more relevant than others.
For impressions on pages dedicated to very niche topics, the content itself will be very important to the model. So, if we log an impression coming from a blog post dedicated to ‘First-time mums’, we will surely consider this input when defining our prediction.
On the other hand, content is not enough for impressions on more generic pages, and the model then looks at the other inputs.
For starters, the device, web browser and operating system are good indicators of age. Think of your own life experience: who do you think is more likely to navigate the web on a desktop via Internet Explorer using an old operating system like Windows XP? The older generations or Millennials?
Also, geolocation can tell us a lot about the user; both the approximate location and the internet service provider used can indicate income level or if the user is a homeowner. For example, by looking at internet service providers, the model can identify a user as employed as the IP address is recognised as a corporate one.
Finally, time is also considered by the model. For example, by looking at when during a day, a user is browsing, the model can identify if the user is employed or if she has kids.
In short, the Full Reach solution relies on a lot of different inputs and chooses the most reliable one based on the context.
Historical knowledge to ensure quality
One thing is how the ID-less model identifies a user. Another thing is how we make sure that the model delivers high-quality predictions.
The ID-less model is based on the AudienceProject panel, meaning that our technology first collects the different inputs from a page view and then validates them against the sociodemographic information from millions of panellists and billions of page views.
Only by looking at a large amount of historical data focused on the panellists’ behaviour, the ID-less model can predict with high precision.
Let’s say that you want to find out if a user is employed based on the time of the day she is surfing the web. Only if you have a large amount of data regarding users and their behaviour, you can tell that people who are employed are more online during the evening compared to those who are not employed.
How different inputs are equally good at predicting correctly
ID-less targeting is not something new. Adtech vendors have been selling solutions based on context and geolocation for years. What is different about the Full Reach ID-less model is how it utilises the most effective input for every case so that every impression benefits from the most effective targeting solution.
To understand what this means, let’s have a look at this graph where we are showing how much each input can predict age when compared to just ‘spraying and praying’ (undirected targeting).
What the Full Reach solution does is taking into consideration different ID-less inputs and matching them with data points derived from a large number of panellists and billions of page views, making it possible to apply the best solution for any given impression in real-time.
So, in this case, inputs like content, operating system and device family would be important to the model when predicting the age of the user, as each of these inputs tells more about the age of the user than the other inputs. With the content input alone, the model will be able to deliver a 57% uplift in reach for an age segment. This means that if you want to reach 25-34-year-olds in the UK, who make up 10% of the UK online population, the Full Reach solution will make sure that you instead reach 15.7% in this target group by considering the content input.
Get ready for the new digital scenario
Implement our Full Reach solution today to boost your share of addressable users and let media buyers know that you are able to offer accurate audience targeting for all impressions and video views.
And most importantly, make sure that you are future-proofed for what is yet to come, no matter what Google says.
Reach out to email@example.com or your usual contact at AudienceProject and learn how to get started targeting your cookieless audience.
Full Reach is available for publishers in the UK, Germany, France, Denmark, Sweden, Norway and Finland.