Here are some ways you can get automation to play nice and to drive your KPIs in a controlled and accountable way. Learn how to automate yourself to success in PPC.
Since the main purpose of Paid Search (like all marketing) is to communicate and persuade humans to choose a business, manual management is unlikely to disappear entirely.
However, Google, in particular, have been investing massive sums in machine learning so as to reduce marketers’ reliance on human management of campaigns.
Whilst we’re in this hybrid stage (a little like where we’re at with self-driving cars) it pays to not entirely reject automation.
Also, with the gradual takeover of Google Ads’ new UI, some legacy bid rules setups will be changed or removed. So getting accustomed to the more AI-style automation will be important in the coming months.
Here are some ways you can get automation to play nice and to driving your KPIs in a controlled and accountable way.
What’s great about automation?
I am the Head of Account Management at ESV Digital, and we have always employed automation in a number of ways so we’re no strangers to letting computer power take the reins.
However, we always monitor the effect on performance and the decisions made by our silicon-hearted friends.
There are currently two broad types of automation right now:
- Bid automation, which is designed to help you achieve a specific goal for your campaigns.
- Content and keyword creation (e.g. ad messaging, extensions etc.), which can help you show your ads to the right audience. Your keywords should match the terms your potential customers would use to search for your products or services.
Screenshot taken by the author from Google Ads.
The simplest and also most risky element is the bid automation. It’s simple because in Google Ads it boils down to simply setting a performance target and seeing what happens. The risk is in the fact that…you merely set a performance target and see what happens!
There are more rule-based ways to bid but they are somewhat limited both in control and in the fact that they cannot take advantage of the more target-oriented bidding schemes (called Smart Bidding) which bid per query (based on all the context of that query, such as user history, browser, location etc.).
To simplify your decision on which option for bidding to go for, look at the data levels in your account. If it’s dealing with many hundreds of conversions per day, you’re likely going to see a pretty good performance with Smart Bidding. If you’re talking a couple hundred or less per day, you’re going to need to set it on only the highest-converting parts and generally, the rules will work better.
The reason is Smart Bidding is entirely algorithm-based, which needs data. The less data, the less accurate. So scale is extremely important if the algorithms are going to do the right thing more often than not.
- Ensure your campaign structure is set up to make the most of the bid automation. If you’re dealing with mainly Broad traffic, try first to get it to majority Exact traffic. This will ensure less can go wrong.
- Get negative keywords in place to reduce the likelihood of an explosion of useless traffic when bidding commences.
- If performance and volumes are wildly different across devices, think about splitting them out so the algorithms can concentrate on the most consistent data. The more “reliable” the data set, the less the algorithm will introduce quirks. Essentially, if you know statistics, it’s getting the groundwork done to maximise the R2 ratio.
- Start any of this with the highest volume campaigns to test what target is working for you before rolling out to the rest of the account.
Also read: 3 ways you can win at PPC in 2019