With all the talk of AI being the best way to do digital marketing, a lot of brands are relying on automation to handle their PPC campaigns.
When compared to manually managed campaigns, can artificial intelligence outperform manually managed paid search efforts?
What Are The Automated PPC Bidding Options in Google Ads?
To understand bidding on paid search campaigns in Google Ads, it helps to have an understanding of the different options that exist. This is a simple guide, but it is handy to understand the different outcomes to expect when selecting different bidding behaviors in Google Ads.
We'll start with one that is fairly self-explanatory. Maximize conversions is one bid strategy available when you set up a paid search campaign. The AI or machine learning in Google Ads is trained to find the most likely converters based on your conversion actions, and then those users are supposed to convert. Simple, right?
There are two major problems with the maximize conversions options in Google Ads. The first is especially important for new campaigns, as Google has no data to act on. When you set up a campaign with maximize conversions as the bid strategy, Google has no idea what a converter looks like. It will show ads indiscriminately and attempt to learn from the behaviors in real time. The problem with this is the cost it takes to do this. Because PPC doesn't stand for "Pay Per Conversion" every time someone clicks your ad, you are charged. For some industries, these clicks can come at a cost as high as $5.00-$10.00. If you have been running a campaign manually for a couple of weeks and see conversions, then switching to maximize conversions may not be a bad idea, especially if you are managing several campaigns at once. However, even this is potentially dangerous, and this brings us to the second problem.
The other major problem is also cost-related, but does not necessarily limited to new campaigns or accounts. The maximize conversions PPC bidding strategy does not do a great job of cost control. If it is showing you can get conversions at $75.00 each, then it will continue to bid in this fashion. For some industries, this is acceptable, as roofers and lawyers will make at least 100x on a conversion. However, for those selling skin care products at $30.00 per item, a $75.00 cost per conversion is unacceptable.
Manually bidding on a campaign can both help you maximize the number of conversions and also control the overall cost per conversion. This is how The Mitten Law Firm obtains real phone calls to their office for less than $40.00 per call, translating into hundreds of thousands of dollars in billable hours.
Target Impression Share
Smaller brands operating in a competitive space might be tempted to use target impression share as a bidding option. Because huge brands are able to spend more money on their campaigns, they often drown out smaller competitors simply by outspending them. This leads many companies to think that if they use target impression share and bid on the top position, they will be able to cut through the noise.
Theoretically, this should work. However, much like maximize conversions, this option is fraught with potential issues, mostly related to cost. Because the campaigns target keywords are all bid to be in the top position, your ad may display in the top spot for a low-intent search term that leads to a click, with no conversion.
Think about it this way. If a brand like Victoria's Secret is bidding on the term "buy lingerie" and then "best lingerie companies", it might make sense for a smaller company like Lover's Lane to bid on those terms as well. However, the "buy lingerie" term has purchase intent, while the "best lingerie companies" is more of a research keyword. If done manually, the research keyword would have a lower bid, but when you select target impression share, it will spend whatever it takes to get you on page 1 and even at the top for any keyword in the campaign.
You can see how this will lead to wasted spend on any target impression share bidding strategy. A campaign using manual bidding from an expert would be able to make these distinctions and bid accordingly, saving your brand money while also helping you achieve top position on the SERP. This is how Lover's Lane outperforms larger adult retailers like Adam & Eve and achieves a 5x ROAS monthly.
Given what we have said about the first two automated bidding options in Google Ads, you might think that maximize clicks suffers from the same problems. However, because of the very nature of the goal, Google's machine learning has optimized to not indiscriminately spend money on clicks. In fact, this is one of the better automated bidding options, as it does a good job of controlling for cost.
The biggest issue with running a campaign with a maximize clicks PPC bidding strategy is the intent. Unless your website converts at a high percentage (anywhere from 5% on up), this option will eventually not give you much of a return on ad spend. The reason? The behavior model is set to get your ad the most clicks possible to drive traffic to your website. However, that traffic could very easily be someone just coming home from the bar and idly playing with their phone. They have absolutely zero purchase intent, and will therefore fall off quickly.
Some websites are built to be conversion machines. However, even a great website like that of Maureen McDermut's, a top selling Montecito Realtor, still benefit from manual bidding, as the intent of the searcher is taken into account. For example, most real estate agents do not handle rentals. So being able to control bids and keyword management for terms like "houses for rent" or "I want to lease a house" can help drive only the most valuable traffic to a website.
The e-commerce industry is regularly focused on ROAS or "return on ad spend" so targeting a specific ROAS sounds like a great idea. However, these campaigns rely heavily on previous conversion data to run at optimal levels, and the vast majority of accounts don't have the right data to make them effective.
When thinking about ROAS, the e-commerce industry standard is 3.5x. That means that for every $100.00 spent in ads, you will make $350.00 in revenue. The big problem with this type of bidding strategy again comes down to a lack of data. Newer accounts won't have the conversion data needed to inform the machine learning of how to achieve the targeted ROAS, so the AI will attempt to do this in real time, usually to less than stellar results.
Usually, target ROAS campaigns won't overspend, but the opposite. They will flounder and you will see far fewer impressions or clicks, and then also a low spend. Reducing spend is always a goal, but if you are not selling anything, then the reduction in spend simply equates to giving up.
Intent is something that machine learning has not had the best record of modeling, and this is why manual bidding still outperforms all of the AI Google Ads bid strategies.