Evaluating Automated Bids
The previous post handled the topic of automated bidding with Google Ads for online stores. This post examines how automated biddings can be evaluated in terms of their performance which after all is the most essential and interesting part of any online campaign.
Such an evaluation can be done with the two tools of campaign draft and experiment which are not the same and have to be separated from each other. Multiple changes to a campaign can be prepared with a draft. Following that it is either possible to apply draft's changes back to the original campaign or to use a draft to create an experiment. Experiments help measuring results to understand the impact of changes before applying them to a campaign.
When it comes to campaign evaluations, Google itself recommends to keep things focused on simple which means to stay concentrated on automated bidding and the related KPIs (Key Performance Indicators) during a specific test phase. More precisely this means to resist testing other types of advertisements, landing pages etc. at the same time. How long should such a test phase last? Typical recommendations indicate a time frame of at least 30 days. The topic Google Ads and time frames will be handled in a separate post.
A way to analyze success is a comparison of current bidding strategies with new strategies, e.g. automated bidding VS manual bidding. Such testing is also summarized under the beforementioned term campaign drafts and experiments. Google Ads offers such testing under its page section “All Experiments”. The related campaign drafts can be adjusted and changed but should always be saved in order to have the best performing drafts available at the end of the testing phase of approximately 30 days.
At the beginning, campaign drafts are elaborated which according to Google are clones of an existing campaign that do not serve their own ads but can be used to stage changes without modifying the original. Once having worked out this new draft, it is now possible to either replace the initial draft with the new draft or to let the initial draft run in competition against the new draft for some time. At any point it is possible to return the first draft, of course.
In case of an experiment, now the traffic share between the running campaign and the experiment must be decided, as this determines the share of activity at the auction after a customer has performed a search on Google. After this it is possible to monitor performances and either cancel experiments early or replace old campaigns with more successful experiments.
It is important to note that drafts and experiments are only available for Search and Display campaigns while Video, App, or Shopping campaigns are not supported. Furthermore, campaign drafts do not support Ad schedule report category & search terms, auction Insights, display placements report, scheduled email reports, bid landscapes, keyword diagnosis and shared budgets. Experiments generally support the same features as campaigns, the exemption of Ad customizers that use "Target campaign" or "Target ad group".