Google, which provides a number of free services, relies on online advertising as its main source of income, and provides a framework called attribution model to measure the contribution of advertising to conversion, which is the advertising end goal of user product purchase and service registration. Google has announced that it will use a data-driven attribution model as a basis for evaluating the contribution of users according to various processes leading to conversion.
In recent years, it has become difficult to use the collected user data for advertisements, so protecting user personal information and enhancing advertisement effectiveness has become a challenge. An attribution model that measures online advertising contribution is an important means for advertisers who post advertisements at a cost to know which advertisements have the greatest impact on users.
With the existing attribution model, there is a last click that assigns credit to the last clicked ad or the corresponding keyword on the conversion path. However, users who purchase products or apply for services on a website are likely to be affected by multiple advertisements they have seen before, not necessarily the last advertisement they clicked on.
For example, when you see an ad on Facebook, you access the website and then remember the website you were curious about that the ad was shown on Instagram. If you go to the website from a Google search, only the Google search term is evaluated as a last click . Ads previously viewed by users on Facebook and Instagram are not rated, so it cannot be said that users can rate the exact factors that led to conversions.
So, Google has built a data-driven attribution model that comprehensively evaluates all relevant data that users convert. The data-driven attribution model uses advanced machine learning tools to ensure accurate ad evaluation based on ad format and user action conversion time while respecting user privacy.
Google claims that advertisers can use a data-driven attribution model to reduce the cost of advertising and earn more conversions for the same cost. A pharmaceutical telecom seller said it was able to cut operating costs by 18% compared to last click through automated bidding and a data-driven attribution model. The bank’s online head also said that after switching their Google ad search campaigns to a data-driven attribution model, they increased overall conversions by 8% while reducing costs by 8%.
A data-based attribution model is already available as a model for evaluating Google search and YouTube display ads, but so far, the condition has been that advertisers are getting more than a certain amount of ad interactions and conversions. In a new blog post, Google describes using a data-driven attribution model to remove data requirements for advertisers and make it available to all advertisers.
Specifically, after October 2021, all new ad conversions will be expanded to a data-driven attribution model by default, and will be rolled out to all Google ad accounts by early 2022. It also continues the option to manually switch to data-driven attribution models, including last click. Related information can be found here.