Google Ads Automation, Smart Bidding, Smart Creatives – terms that you hear more and more often, but maybe you haven’t used them yourself yet? Today I want to take away your fear of the unknown and explain what you should look out for when creating your automated Google Ads campaign to make it successful.
Why you should use Google Ads Automation?
The automation of Google Ads works through a highly intelligent algorithm based on machine learning. It uses information from users to get the most out of your ads. It uses all the information you provide in your Google Ads account, the website’s content, and data from user behavior analysis.
Thus, the Google algorithm uses much more information about the consumer than you or I could ever do in manual campaign optimization. It would be best if you used this advantage to tailor your campaigns to customer needs and their search behavior. Your ads will be displayed individually for each user. Also, you can evaluate successful content afterward and further optimize your campaigns.
How to automate campaigns
Today I will focus on Google Ads Automation for existing accounts with already active campaigns, emphasizing Smart Bidding, and Smart Creatives. I won’t deal with automated rules and smart campaigns in this blog post. Everything about intelligent shopping campaigns can be found here.
Like all Google Ads automation, Smart Bidding, and Smart Creatives are based on an algorithm. This algorithm is much more complicated compared to automated rules and therefore requires more information.
It is, therefore, advantageous if standard campaigns are already active. Besides, you also have better comparison options for the evaluation. Today I am focusing accordingly on Smart Bidding and Smart Creatives. You can combine both areas of automation perfectly in your campaigns and thus increase your success.
Google understands “Smart Bidding” as automatic bidding settings. Using machine learning, the Google algorithm selects a bid. In doing so, it analyses user behavior (also about the chosen objective) and increases the request if it is more likely to take the desired action. Depending on the campaign objective, a corresponding bidding strategy must, therefore, be selected. Google offers a suitable selection for all target actions.
You can choose smart bidding strategies for the following target:
- More brand awareness
- Increase visibility
- Achieve more conversions or higher conversion value with the given budget
- Achieve more conversions or higher conversion value with a specific target CPA or target ROAS
Smart Creatives, like Smart Bidding, are also based on an algorithm. This algorithm does not use static ad texts or banners but combines texts, images, and videos made available to it according to the modular principle, tailored to each consumer. Thus your ads are individually adapted to the user and his search behavior.
Depending on the previous user behavior, the elements are put together in such a way that, from the algorithm, it leads to your goal. Smart Creatives, therefore, not only offer a time-saving way to create ads, but also provide you with additional information to understand the customer’s intention better. You have the following options: dynamic ads, responsive search ads & responsive display ads.
1. Check the current status
Before you start using Google Ads Automation, you should first check your account’s current status because not every company has automation working for them.
The decisive factor for the evaluation is a look at which campaigns have already run so far; and, above all, the conversion figures. The algorithm needs enough information to work. So there should be a website with existing traffic, and a certain number of conversions should also be available.
To ensure this, Google requires a minimum number of conversions for some bidding strategies. Therefore, it is best to take a close look at how many conversions you achieve on average within 30 days and how much traffic you have on the website.
Google now recommends the use of Smart Bidding even for lower conversion rates, but the algorithm takes longer to learn and sometimes does not work as well. In our experience, you might need 30 conversions per campaign using Smart Bidding – but we would be happy to hear about your experiences.
2. Avoid standard mistakes
Campaigns can fail for many reasons. Typical standard errors can happen with automated and regular drives. You should avoid these if possible and take them into account when creating your automatic ads. You can also find many starting points on the recommendation page of Google. You can find them in your account in the top left corner of your account directly under “Campaigns.”
You will find many useful optimization options for your account (and unfortunately also many non-sensical ones – do not blindly accept them, but evaluate them according to their usefulness). Think about this:
- Do the recommendations support the goals of my campaigns? (e.g., bidding strategy)
- Do the suggestions fit my product/service? (e.g., keywords or ad extensions)
- Can or do I want to implement this change? (e.g., budget increase)
The percentage of recommendations influences the level of the optimization factor. According to Google, the level of this factor has a decisive effect on the playout of your campaigns (whether this is the case is difficult to understand). However, you do not necessarily have to implement the recommendations.
It is possible that they do not make sense for your goals. You can simply reject them using the three dots in the top right corner of the recommendation box. Click on the dots and then “Delete all.”
3. Choose conversions and bidding strategy correctly
Smart Bidding is based on the behavior of users about your conversions. If the wrong conversions are selected (or if conversions are mismeasured), they may not support your overall goal with Google Ads.
Here you should think very carefully about what the goal of your website or campaign is. If you are in e-commerce, transactions are probably the most relevant conversion for you. Of course, several conversion targets can also make sense for you.
Accordingly, you should also define these targets. But in your conversion settings, consider which ones should be included and which ones should not.
In addition to conversions that are not ideally set, the selected bid settings can also lead to the success or failure of a campaign. Finally, Google optimizes in line with the set strategy. Sometimes there are also different bidding strategies for a target.
You have to decide which bidding strategy best suits your goal. For example, if your goal is to generate a lot of revenue, you can choose between target CPA, target ROAS, and maximize conversion value. For brand awareness, “maximize clicks” and “target share of possible impressions” are proper bidding settings.
You can use “maximize conversions” for lead generation, but you should be careful here. The Google algorithm will then try to maximize your budget to generate as many conversions as possible. This could lead to a sharp increase in expenditure.
4. Provide the algorithm with as much information as possible
The algorithm works based on the information supplied to it. If too little information is provided, this can harm your campaigns.
The algorithm can only use the data provided to it to analyze, evaluate, and react accordingly. However, if it receives too little information, it cannot collect enough useful data.
This can delay the learning phase and cause Google Ads Automation to fail or perform worse than it would with manual optimization.
So what can you do?
Utilize the number of texts and characters:
Use the number of text available to you and the number of characters in the individual text elements. For example, in responsive text ads, you have 15 headlines and four descriptions available.
That gives you a total of 19 text elements per responsive text ad. Wrap as much information as you can in these short text elements. The user is already used to finding information directly on Google and no longer has to click on websites. Therefore, your goal must be to arouse curiosity via the short ad text / the Responsive Ad in such a way that the user clicks.
Make it clear in the book where the ad is going and what to expect – Provide an incentive to click on it.
In terms of text, this means that if you can insert five headlines, then insert five. And if the number of characters in the headline is 30, try to get as close as possible.
- Remember, you never know which elements will be played in the Smart Creatives.
- No matter which combination is shown, it must always make sense and not depend on another text element.
- The more elements the algorithm has, the better it can find the best combinations.
- Depending on user behavior, the ideal combination can vary.
Therefore the content of the texts should also be different and offer options.
This approach for text & character count is also valid for the images of the responsive display ads. You can upload many different models, which can be played in various combinations with text and logo. Ideally, the photos should be added in landscape & square format, this allows the algorithm to vary as much as possible.
Targeted use of target group settings:
Another way to provide information to the algorithm is to set target groups. There you can adjust the campaign to specific target groups or set it to watch. If you focus a drive on a target group, your ads can only be shown to the people in this group. On the other hand, your reach is not limited when you use the watch function. However, this allows you to monitor the performance of your ads to this target group.
Here it would help if you stored the appropriate target groups on observation and not on alignment. An exception would be a remarketing campaign, where you must, of course, choose an adjustment. Through machine learning, these observed target groups are more strongly observed, and it tends to be played out to them more often.
At the same time, you don’t cut off any traffic, as with alignment. You tell the algorithm in advance which users are more relevant for it, and it aligns the campaign accordingly.
Google also offers you more and more possibilities with the target groups. New target groups are continually being added. You can promote all this additionally by creating remarketing lists yourself.
You can also set these to observation, and Google offers you the possibility to select a target group “similar” to the users on your remarketing list. This means: They have not visited your website yet, but they show the same interest and user behavior as your previous visitors.
Use information already collected
You should also use your experience from other ads here. Look at your most successful ads and use them as a guide when creating copy for your automated campaign. Also, note the keywords of these selected ads. In general, brand keywords convert better. Here the success of the ad may not be due to the text but to the already higher interest of the customer.
Specific to dynamic search ads is the following tip. Instead of directly specifying URLs, you should choose the option “Target pages from the default ad groups.” The mechanism will use the URLs from the ad groups already booked in that campaign.
This has the following indirect positive effect: With this setting, not only the landing pages are taken over, but also a lot of other information. The keywords, ad texts, and target group of the standard ad groups are now taken into account during machine learning. You tell the algorithm in principle what kind of user you want to have on the target pages, namely, similar to the standard ad groups.
For these groups, the algorithm already knows what kind of user behavior it should play out for and uses it for the dynamic ads.
5. Use your budget correctly
For the algorithm to have the necessary leeway, it needs an additional budget for learning. Therefore you should plan on 20% more budget. Be aware that the algorithm is testing a lot in this phase, and the costs do not necessarily bring the expected output.
Only when the learning phase is finished, do the numbers become representative and you can adjust the budget to the planned expenditure. Again, you should not change more than 20% per week.
In retrospect, the playout works the better, and the input brings the corresponding output.
You should still check the campaign regularly. If the learning phase gets out of control, you can always stop. But you should only do this if it is essential. First, check if all settings are correct and other errors (e.g., too few conversions in your account) can be excluded.
6. Let the algorithm learn
If it is now time, and you have created everything right, you can start the campaign. From then on, the algorithm is in the learning phase. During this time, the status of “learning” is indicated in brackets.
For the learning phase you should know:
- The duration of this phase varies from campaign to campaign and depends on the amount of information provided.
- The more information provided, the faster the algorithm learns.
- On average, you can expect a phase of 1 to 2 weeks.
- In the case of low traffic or low budget, it can be extended.
If “learning” is no longer in the status, the critical learning phase is completed. Once the learning phase is over, you should ideally not change anything for another two weeks.
The main phase is finished, but the algorithm is still susceptible and needs more time to learn. During this time, you should not change anything at all about the campaign. Every algorithm has to learn in the beginning and should never be disturbed.
Otherwise, the results will be distorted, and the campaign performance will deteriorate. It can lead to faulty playout, the learning phase is prolonged, and in the worst case, the whole learning phase starts all over again. Instead of incorrect information, the algorithm can also be confused by inappropriately supplemented data. This does not mean that this information has no relevance; only the time and amount of change is too large.
During this time, you should monitor your campaign regularly. Machine learning can work very well, but there are always exceptions. However, you should only intervene if necessary, and you should check possible disturbing factors before.
So, after finishing the learning phase plus two weeks, you can theoretically make changes. However, even if the algorithm is no longer in its learning phase, significant changes can lead to a return to the learning phase.
For example, this happens if a lot of information is added (or even missing at once – e.g., if conversion numbers collapse due to a newly installed cookie opt-in banner). Then the algorithm learns wholly new and makes already sounded errors again.
Therefore you should always make only small changes. For the budget, the change should not be more than 10-20% per week. The whole thing varies depending on how much information the mechanism already has and how sensitive it still is. With keywords & texts, it is difficult to give a comparison value, here I can only recommend you to keep the change small.
7. Correct evaluation
It is also possible that your campaign is thriving, but you do not notice it at all. How does that happen? By stopping the drive too early because of insufficient or too early analysis.
- The success of the campaign can only be determined with the right analysis.
- Depending on the budget and size of the campaign, the ideal time for evaluation can vary.
- With automated Google Ads campaigns, the algorithm needs two more weeks to learn after the learning phase. If this time is over, it makes sense to check weekly and look for positive trends.
- Of course, you should also check the campaign regularly the weeks before, but you should not draw any conclusions, because, in this period, the numbers are not meaningful.
A correct evaluation is then, in most cases, useful 4 to 6 weeks after the end of the learning phase. Until then, you should not change anything about the campaign.
Small budget adjustments of max. 20% is okay, but only after the end of both learning phases. If the campaign is still not successful after this time and has little or no conversions, you should start troubleshooting.
If there are no glaring mistakes in the campaign, you should look at other aspects.
The most common problems are often: poor usability of the website, price-performance of your products is not right, or the tracking is not set up correctly. Take a close look at everything before you deactivate the campaign.
Also, be sure to keep disruptive factors, such as double keywords and general fluctuations in mind. Above all, variations are straightforward to compare with the performance of other campaigns.
8. Prepare yourself in time for seasonality
The Google algorithm can handle seasonality very well, and you don’t have to intervene automatically just because you know a seasonality. For many B2C companies, however, events such as Christmas, Easter, or the start of school are particularly relevant to turnover, and they can invest more budget or run new campaigns.
When adjusting the budget, you should not make significant changes all at once. Therefore, it is vital to plan your budget well in advance and increase it by 10-20% only once a week.
If you create an entirely new campaign for the Christmas season, you should do this in October. Then your campaign will have enough time for the whole learning phase, and in early December, the algorithm will work as desired. This means that you do not have to adjust the budget weekly if you already have the desired amount during the learning phase.
Conclusion and outlook for the future:
To sum it all up: Create a good basis, provide a lot of information, let the algorithm learn and then stop interfering. That’s how you become successful with Google Ads Automation!
Whether you want to take the risk of using Google Ads, Automation is your decision. I can only recommend you give it a try. But only if your account has enough traffic and conversions for Automation. In the future, Google Ads will work more and more automated.
Of course, the current situation also has an impact on Google Ads. We currently see changes in search and purchase behavior in all Google Ads accounts, especially increasing CPCs and CPOs.
Some industries are more affected, others less. From Google’s point of view, the Google algorithm is so good that it can adjust well to a changing environment after a few days. Nevertheless, we recommend much more intensive monitoring – and of course, more intensive campaign management.
Useful Links For You: