Starting July 2022, Smart Shopping campaigns have given way to Performance Max ones. The ultimate promise is to “[increase] the Conversion value by an average of 12% at the same better return-on-ad-spend”. This is a further step towards the automation of Google Ads campaigns, which has been accelerating over the past few months through the generalization of RSAs in June 2022, the promotion of the data-driven Attribution model at the end of 2021, as well as the update of match-type Phrase back in July 2021. This gradual automation comes with an increasing level of opacity for advertisers, who have no choice but to trust algorithms entirely. This implies two challenges: first, controlling their budget allocation in an environment where the Search market is growing rapidly, and second, protecting their brand image.
In this context, how can advertisers retain control over their Google Ads investments?
As long as they’re used smartly, algorithms often allow to improve campaign effectiveness. When it comes to automated bidding strategies, the Algorithm relies on a variety of signals in order to adjust bids in real-time: day, time, audience, query, context… For instance, we carried out an A/B Test for a campaign, with the existing “manual CPC” strategy on the one hand and the automated “maximizing conversions with a target ROI” strategy on the other hand: conversion rate increased by 0.2 points, and revenue went up 45%. While it was once necessary to conduct analysis in order to determine which upward or downward adjustments had to be applied to certain audiences or times of day, it is now the algorithm that performs this task, in real-time and on a case-by-case basis. The latter point is a good illustration of how automation saves operational time for agencies, which can now focus on higher value-added tasks such as in-depth analysis.
However, because automation brings about so much opacity, one needs to take a step back and consider the displayed results with caution. The data available to assess campaign effectiveness is dwindling. Reports on queries are either sampled or non-existent. Similarly, reports on ad placements as well as their exclusion potential are very limited, at the expense of Brand safety. Besides, if not carefully monitored, automation can lead to a cost increase. Let’s take the example of a client wishing to ensure the visibility of their brand. To do so, they can set up their campaign with the “Target impression share” automated strategy along with a 90% share of voice target. The algorithm will then bid in such a way as to guarantee this threshold, which can prove problematic in the context of increasing competition. If another advertiser were to adopt the same queries, the algorithm might actually increase its bids so as to guarantee the minimum required, thus strongly reducing the return on investment. For example, we carried out an A/B Test regarding the brand of one of our clients in order to compare the existing “Target impression share” strategy with a “manual CPC” strategy. The latter allowed us to bring CPC down by 60% for equal visibility and revenue. Still, on the topic of bidding strategies, we have noticed that “Maximizing conversions with a target ROI” also had a tendency to increase costs. After a certain threshold, the defined ROI target has a direct impact on the CPC increase (see diagram). In addition, this strategy can result in very high unit CPCs, sometimes even higher than 10€. We can’t help but imagine, then, that the algorithm will be prone to more expensive clicks once its target ROI is secured.
Illustration of the impact of Target ROAS on the evolution of CPC
Against this background, advertisers must be able to trust their SEA agencies to control their investment and make the most out of the automation solutions that are offered to them, or to rely on external agencies in the case of internalized media buying. Their role is threefold: strategic, tactical, and operational.
From a strategic standpoint, the agency can help the advertiser define its business objectives, prioritize them, and translate them into measurable media KPIs. This step makes it possible to determine the appropriate strategy to maximize performance on such indicators. Let’s take the example of an e-commerce advertiser that wishes to maximize their revenue while keeping control over their ROI. We can consider that the ROI will be their main performance indicator. Since the CPC level has a direct impact on costs, it can be regarded as a control KPI.
The agency can then define the tactics that will allow the advertiser to reach their goal, i.e. the account structure, keywords, match types, bidding strategies, etc… If we take up the previous example: in order to maximize revenue while guaranteeing a minimum ROI, it would be tempting to apply to campaigns the automated bidding strategy: “Maximizing revenue with a target ROI”. But how can you be certain that this is the right choice? By using an A/B Test if the platform allows it. A test & learn approach is indeed key to ensuring account optimization. It is a much more rigorous approach than the one-click application of Google Ads’ ready-to-use recommendations.
Lastly, the agency is responsible for the account’s operational excellence. The latter is achieved through technological and competitive intelligence, as well as the adoption of a continuous improvement approach. By this, we mean: going beyond performance alone, conducting further A/B tests, or questioning the architecture of certain campaigns. Beyond monitoring account performance, safeguards can be easily integrated in order to ward off the drawbacks of automation. For example, bidding portfolios can be used to avoid the impossibility of associating maximum CPC thresholds to automated bidding strategies. Besides, the first headline of an RSA ad can be frozen in order to ensure that a message is always displayed. For instance, we have clients who want their brand name to always be present in the ad. Furthermore, the granularity of Smart Shopping or Performance Max campaigns can be defined at the ad group level in order to fine-tune its optimization.
Although it raises more and more questions about the possibility of retaining control over one’s investments and brand safety, automation is a fundamental and therefore indispensable market trend. One must therefore rely on the intelligence of platforms and automate what can be automated… but in a smart way. That is to say: by establishing safeguards, by not taking anything for granted and rather adopting a test & learn approach, and by ensuring to always take a step back from mere performance thanks to the little data that is still available. In this context, advertisers should be able to trust their agencies now more than ever.