Organisations that invest in pay-per-click (‘PPC’) advertising will be aware that the PPC platforms have become much more complex in recent months and years and are constantly changing.
A complex auction
The display of ads to online searchers or social media users is the result of a real-time, high-speed auction. There are many competing factors that determine which ads are shown when, including:
- Ad configuration – you can specify what days and times your ads should appear. You can also specify where you want the viewers of your ads to be located (city, region, country etc). Depending on the ad platform, you can specify that searchers should have certain interests and/or demographics.
- Advertising budget – how much is each advertiser willing to pay. Obviously, an advertiser who is willing to pay the most in a given location, time will be more likely to have their ad shown. Ad budgets are often specified as daily spend amounts, so the platform needs to keep track of how much of your budget has been spent that day. Budgets can also be configured to be spent as quickly as possible within a given month.
- User experience. Whilst how much an advertiser is willing to pay is a big factor in how often their ads are shown, PPC platforms such as Google Ads also give a high priority to user experience. Google will use up more of an advertiser’s budget per click when it determines the advertiser is providing the searcher with a poorer experience compared to another advertiser. Google calculates the user experience based on many factors, including the actual behaviour of searchers e.g. the likelihood that people who click on your ad will also take another action from your landing page.
Due to the complexity, it is essential that PPC platforms use AI to determine the best ads to show, without the user experiencing any delay, but also to keep both advertisers and searchers / browsers happy. Advertisers are happy when the money they are spending on ads is defrayed by the sales or other they gain from the ads. Searchers are happy when the ads they click on help them to quickly solve their problem or meet their need.
AI is used in the delivery of both search ads (e.g. ads in the Google Search engine and Bing Search engine) as well as ads in social media platforms or web pages (e.g. Google Display Ads, ads shown in Facebook & Instagram feeds etc).
AI is used within particular bidding strategies, in order for those strategies to be achievable. For example, the Google Ad platform allows advertisers to use newer bid strategies such as “Target CPA” which targets searchers for whom the cost per acquisition is likely to be at a certain $ value. That value is set by the advertiser, but the choice of which searchers are likely to deliver that CPA is calculated by Google’s AI during the real-time auction.
AI is also employed by the ad platforms to automatically make recommendations to advertisers about how their ads can be improved – more about this in a moment.
AI is employed by ad platforms to combine ad components to create the ad on-the-fly, based on the likelihood of the various possible ad components to match the searcher’s needs, attract the searcher to click, and take the searcher to the most useful landing page. This is required for ad types where the advertiser sets up a collection of possible headlines and ad content, and the PPC platform chooses the best options to combine within the search results.
AI is employed by ad platforms to build interest groups and understanding of the current intent of a particular person with respect to a particular product / service, at any given point in time. They then use this information to provide ads to the right people at the right time.
The impact of AI in PPC
AI is required because of the complexity of delivering great results for advertisers & searchers, as explained above. But at the same time, the use of AI is causing complexity in the PPC advertising platforms. This is because the use of AI:
- makes it harder for the platform team to explain or predict exactly what is going to happen on a given day for a particular advertiser, and
- means that the results are changing all the time and therefore the best way to configure your ads changes as well, as the platform continues to learn and adjust to real world behaviour, and
- requires that advertisers keep up to date with how AI-powered configuration options work, so that they know when to use them, and when not to use them.
The Meta Advertising Platform places your ads into a “learning phase” when you first set them up, or when you edit them after they have been running for any period after the learning phase. It can take several days or weeks (depending on the size of the audience that meets the configuration criteria for showing your ad) for your ad to generate enough activity so that the platform can efficiently spend your advertising budget by showing your ad to people most likely to click on it and then act (aka “convert”). Until your ads are out of their learning phase, your cost per click will likely be higher, and the likelihood that a person who clicks will take the desired action will be lower. You need to have the appetite to spend for long enough on your ads to get through the learning phase and then, hopefully, get the desired return on investment.
Whilst not strictly an impact of AI, PPC ad platforms use the data they gather on the performance of ads to improve their features. For example, it’s been well-known in recent years that the Google “Broad Match” option could cause advertisers to waste money by showing their ads to searchers who weren’t looking for their service at all. Google undoubtedly has been very aware of advertisers avoiding using “Broad Match” except in very controlled circumstances. Hence, it was no surprise that Google recently re-launched their Broad Match feature, which anecdotally is providing much better results for some advertisers than previously.
To help combat the increased complexity, some PPC ad platforms now constantly provide recommendations for how your ads should be modified to deliver better outcomes for yourself as the advertiser or the searcher / browser. These recommendations can take many forms, such as suggested budget increases, suggested changes to keywords, bid strategy changes, opening up your ads to new audiences, or ad types, and much more.
If you don’t monitor and action the recommendations (either by dismissing them or following the suggestions) then your score on the platform will reduce, and you will pay more for your ads because you are deemed to be providing a poorer experience for the searcher. This means that your ads need to be monitored on a very regular basis, or you will be paying too much, and your ad performance will reduce.
Perhaps an even greater impact on PPC than all of the points above, is the recent launch of ChatGPT – a publicly available generative AI tool. You can read about the impact this is having on the use of AI by business. The integration of OpenAI’s functionality within the Bing search engine, which was launched in a limited capacity on the February 7, 2023 is likely a harbinger of seismic change in the realms of PPC advertising. We can only guess what changes might be … will Google lose their dominance in the extremely lucrative search market to Bing? Will users move en masse away from search engines (and long lists of search results, including ads) in favour of a single result, or a result that contains more nuanced explanations and advice on questions they have?
The Future
AI is here to stay in PPC advertising. It is impossible for the PPC ad platforms to function without AI.
For the foreseeable future, it will be imperative for the person/s managing PPC advertising for a business to stay abreast of the growing complexity of the ad platforms, for such advertising to be cost effective and to provide you with competitive advantage.
It is important that PPC advertising is regularly assessed regarding its usefulness for a particular business.
Need help starting or managing your pay-per-click marketing campaigns? We’re here to help!