It’s easy to see the appeal of a heavily automated PPC account, with machine learning promising to remove the risk of human error and dramatically reduce the number of management hours required.
There are countless third party software platforms that claim to have the ability to manage PPC accounts, as well as automation being at the forefront of Google Ads mind for the past 18 months. Google currently offers over 10 automated bidding strategies ranging from Target CPA to target Impression Share Bidding. This trend looks likely to continue with smart bidding being a prominent feature in Google’s 2019 marketing live keynote.
Too good to be true?
I have tested a number of PPC third party tools, as well as the vast majority of Google’s smart bidding capabilities, across a number of different industries and account sizes. Although some of these tools do have their place in a well-structured PPC account, I would definitely not be investing my own money into a PPC account that relied heavily on these types of automation.
Let’s look at smart bidding in isolation, these strategies allow Google to decide your keyword bids based on specific goals, e.g. to appear at the top of the search results page, or drive the maximum number of conversions. On the face of it, these opportunities sound great however, over time, I have seen that these bidding strategies can have a negative impact on accounts. These tools do not take into consideration competition, seasonality, or other external factors, they simply bid with tunnel vision towards a single metric. Google’s bidding strategies seem to have one thing in common, they all seem to result in increased cost per clicks over time. This is great news for Google, but not great for advertisers. In the paid search landscape, once a keywords bid has increased, and competitors have responded with increased bids of their own, then it becomes a mammoth task to bring the required bids down without losing volume or ad positions. The true impact will occur when both advertisers and their competitors both try to implement smart bidding at the same time. There will still only be a certain amount of market opportunity, resulting in an automated bidding war, with cost per clicks being inflated, potentially irreversibly.
You can’t automate the human touch
There are some PPC tasks that automation and machine learning just can’t cope with, and these optimisations still absolutely require human PPC specialists. Writing fresh, engaging and innovative advert copy is vital in any PPC account, especially in highly competitive markets. There have been a number of attempts to use machine learning to help in the copy creation process, but these have always been underwhelming and often led to incorrect messaging being shown to consumers. Even the basics such as search query reporting and negative keyword farming benefit from human input. When you are paying for each individual click you don’t want to risk bidding on irrelevant homonyms or questionable acronyms, a human PPC specialist would pick these up quickly, however automation may mistake these for useful terms and bid more aggressively on them. Even optimisations that can in theory benefit from machine learning still need to be checked and monitored by a PPC specialist to deliver the best results so that budget is used efficiently.
We are always looking for new ways to enhance our PPC performance, and automation is definitely going to play its part in the future. However it’s still very much a work-in-progress, and we will continue to test ourselves against both third party software and other automation solutions as they develop.