How to NOT Automate Your SEO & PPC with AI

Automation is on the rise, with artificial intelligence (AI) tools able to handle simple, mundane tasks that would have taken large amounts of human time in years past. Automation can save time in both search engine optimization (SEO) and online advertising (PPC) if it is guided by human strategy correctly.

However, if automation processes are implemented without any human insight and strategy, it can actually hurt your PPC and SEO performance more than it helps! Here are some things to keep in mind when implementing automation to save time with PPC and SEO tasks.

Performance Metrics Reporting Automation

Automated dashboards and tools can sync performance data from multiple sources, such as Google Analytics and Google Search Console for SEO or Google Ads, Microsoft Ads, and Facebook Ads for PPC. This saves the time of manually downloading and organizing data from each platform separately and then adding them together for an account-wide view. It’s been one of the most helpful uses of automation to save time in SEO and PPC reporting!

However, metrics from different platforms cannot always just be meshed together in a meaningful way. For instance, the various ad platforms use different attribution when counting conversions, which may lead to under or overestimating the impact of one platform compared to another.

To understand the true performance of each platform, campaign, and keyword or audience, you must understand how the metrics are measured and make rules that help you compare and contrast them equally. Or have a single analytical “source of truth”, like Google Analytics, that gives each platform equal footing.

While automating reporting can save you a lot of time, marketers still need to spend time analyzing the data and applying meaningful learnings to direct strategy. The recommendations and insights from Google Analytics 4, Google Ads, and other third-party tools tend to be very general, and have no knowledge of how it fits your specific business strategy.

If you are only using automated reporting to present performance, you are missing out on the key reason for the reporting in the first place—to find those nuggets of insight that will help develop your future strategy. Marketers should still be spending time digging into the reporting and filtering and segmenting those dashboards to find valuable intel.

 

Writing Website Content with AI

The amount of fresh content needed for the average website to rank highly for a portfolio of keywords can be daunting for most companies. They need teams of writers to keep up with demand and to hope to rank for what their potential customers are searching for. ChatGPT has been making news recently since it is able to conversationally generate detailed responses to queries where it has good training data to learn from.

While AI tools can definitely generate high volumes of content on simple topics, it is not ready to replace human writers yet. There are many niche topics that it does not have training on, it can generate wrong information, and the writing style is still not quite human. It needs human editing and review to be useful.

And, while Google says helpful AI content is not something they consider spam—the company also warns against using it inappropriately to make low-quality content at scale to manipulate results. So, it is not a good idea to use an AI tool to generate low-quality content at scale just so you can rank in the search engines.

However, AI content producers can still help with mundane content tasks to save time, like helping to write ad copy, social media posts, email copy, metadata, etc. with human review and editing. It can help content creators generate fresh ideas for topics and be a useful tool in automating some content tasks.

Automated Metadata Creation

For websites with hundreds or even thousands of pages, it can be daunting to do keyword research and write SEO-optimized page title tags, meta descriptions, and headers for each page.

Many website developers will lean on formulas to create titles from the topic of each page and generic meta descriptions that can fit categories of pages to make sure that all pages have a title and meta description. This way, the site metadata is automated by a formula—no writer has to go page by page through the entire site to write unique, researched metadata. While this can be useful in the short term with a new site, in the long run, it is not helpful for SEO rankings or helping users or Google understand each page of the site.

Formulas are not able to tailor both the metadata and the content of each page of the site to certain keywords that you want that page to rank for organically. Formulas also are not able to write specific meta descriptions to the intent and information on each page to convince the user to click on your result. They only allow for generic, and often duplicated, results that have lower keyword rankings and lower click-through rates than pages where metadata has been carefully researched and written by an SEO analyst. 

Smart Bidding and Smart Campaigns

Google has led the way in developing various AI-driven bidding strategies, including Maximize Conversions, Target ROAS, Target Impression Share, and others.

Microsoft Ads has adopted some of these recently as well. In addition to this automated bidding, Google has been promoting Smart campaigns—like Performance Max, Dynamic Search Campaigns, Smart Display, and Smart Shopping—where the AI targets the best audience and, in some types of campaigns, also creates the ads, with the goal of getting the best results for companies.

These can save marketers time in creating detailed campaigns, keeping track of performance to make bid adjustments, and manipulating bids on thousands of keywords manually every day. Google’s AI has detailed search history and online behavior data on each searcher and knowledge of other competitor bids that allow it to tailor bids to each search.

Manual bidding is based on trends of many searches, not each individual searcher. With advertising data becoming more protected due to privacy laws and policies, it is harder to manually bid on keywords than it ever has been. 

However, it is not smart to let your Smart campaigns or AI bidding run without human strategy and knowledge. While the AI may be able to get you conversions based on past performance, it often struggles in new accounts or those with fewer conversions and also tends to get less qualified leads if not directed carefully.

The AI does not know which audiences and keywords are the best for your individual business. It still needs manual steering through adding negative keywords, placement exclusions, secondary bid adjustments (in some cases), and budget adjustments between different campaigns based on keyword type and impression share. 

Unfortunately, one of the main drawbacks of using Smart campaigns is that it often hides the targeting and other data that might help us learn where it gets the best results and does not allow for manual adjustments that could help to train it to learn better. Because of this, marketers should analyze the performance of Smart campaigns critically and determine if they are actually getting better performance than more manual campaigns.

Broad Audiences and Broad Keywords

Both Google and Meta have been promoting using broad-match keywords and large audiences as targeting in campaigns and letting the algorithm determine which people are most likely to convert to improve performance.

This works similarly to Smart campaigns, where the algorithm uses its vast knowledge of each searcher or Facebook/Instagram user to show them the right ad at the right time to get the conversion. The detailed targeting provided by marketers does not get in the way of the algorithm.

However, as we discussed above, the algorithm often has no knowledge of the quality of conversions, especially when it comes to lead generation or the individual audience and market of each business. Because of this, it may favor lower-quality conversions over more detailed targeting controlled by marketers.

And, similar to Smart campaigns, marketers cannot get any insight into which parts of the audience are performing better or worse than others or have any control over where the algorithm targets.

Auto-Applied Campaign Recommendations

Both Google and Microsoft have been promoting their auto-applied campaign recommendations. These recommendations vary from removing conflicting negative keywords to having their AI write new ad copy for your campaigns based on your website language, keywords, and other headlines and descriptions that perform well.

It is a little ironic that Google wants to use AI to write your ad copy since they are very against using AI to write website content. Recommendations have been a part of the ad platforms for a while but marketers had to review them manually and decide whether or not to apply them to the campaigns. Now, the platforms with apply them automatically to your campaigns, after a period of time.

Recommendations can help marketers clean up their campaigns based on best practices and catch issues that may have been missed during campaign creation. They can find new opportunities to improve performance. However, they don’t take into account each business’ unique preferences and audience. They can also be based on Google’s best practices which may not match what actually improves performance for marketers. For instance, one of the top recommendations is often to raise budgets on campaigns. However, many businesses have limited ad spend budgets and cannot max out spend on their targeted keywords.

Recommendations work best when reviewed by a human, as with most of the automated tools we have been discussing. A marketer who knows their campaigns and audiences well will be able to decide which ones make sense for the account and which to dismiss or modify before they go live.

Responsive or Dynamic Ads

Automation has become the standard practice when it comes to writing ad copy as well. Responsive search ads and responsive display ads have taken over standard ads on Google. Microsoft has adopted responsive search ads as well. Meta is using dynamic ads for Facebook and Instagram.

Basically, in responsive and dynamic ads, the marketer will write multiple headlines and descriptions and upload multiple images and videos. The AI will mix and match different combinations of these assets to show ads to users based on performance (the combinations with the best click-through rate) and user preferences (the types of ads that each user has clicked on in the past).

These responsive ads can save marketers time in writing and uploading many different ads. The many combinations can make sure that users are served ads that match their preferences and increase the click-through rate on ads. 

However, responsive ads do not show marketers the metrics behind each combination, or even each headline, description, image, or video, meaning that there is no easy way to test and learn what is performing best to improve performance over time. It is even difficult to identify weak headlines and descriptions to pause to improve performance.

Responsive ads also seem to mostly focus on getting the user to click—which seems like a good thing. However, the conversion rate can decrease if the message with the best click-through rate does not match the landing page and conversion well. With traditional ad testing, marketers were able to review both the click-through rate and conversion rate to determine the best ad copy. 

Conclusion

Intelligently using automation tools provided by the platforms and third-party tools can save marketers time on more mundane tasks with reporting and management of campaigns.

However, it cannot replace professional marketers who know how to calibrate automation to make sure that the results fit your business and audience correctly. Automation is just not smart enough, or trainable enough, to fit each individual business need. It is also not sophisticated enough to pull insights from the data and create a strategy individualized to each company. 

You want to make sure that you are working with an agency that knows how to set up and monitor automation and is not still manually bidding on every keyword, fighting all of the advancements in the field with AI, or blindly setting up Smart campaigns without closely monitoring them.

Perfect Search uses a mix of industry-leading AI tools and human strategy and control to pinpoint the right audience for each of our clients. Contact us to learn more!