With Google Ads and Microsoft Ads increasing in automation (whether or not the advertiser appreciates it), it is advisable not to simply accept and use automation, but also to optimize it.
Wait, how do you optimize automation?
Don’t you just press the button to switch to Target ROAS or Smart Shopping campaigns and then sit back and open a cold while the machine does the heavy lifting?
Automation is not about fire and forgetting.
Indeed, with troubleshooting and running machine learning algorithms, your job as a babysitter can be more hectic than you expected at times.
In this article, we take a look at the bigger picture of automation management and analyze the theory behind making decisions so that the tactical process is better controlled.
Here are tips on building the right framework to help you make informed automation decisions.
When it comes to making automation optimization decisions, we are increasingly using what is known as the “magic wheel of automation” in this cutting-edge and expensive diagram (most likely millions in NFT) (see figure below).
While many people think automation is something to be set and ignored, barring a little tweak from time to time, I encourage you to think beyond that.
Read on below
When it comes to how automation works, it’s important to know that machine learning (ML) is based on the data input it receives.
This means that if the data entries are incorrect, incorrect, unclear, or incomplete, no matter how intelligent that particular algorithm is, the ML algorithm is likely to miss the mark.
Of course, once the automation stops working, we should examine our data entries to make sure it is complete (is your conversion tracking code placed correctly)?
However, I have found that the last two issues that can arise with ML (unclear or incomplete data) are the ones that have the greatest opportunity to realign the machine.
This is where human-led automation comes in.
If your data is unclear or incomplete, it’s unfair to blame the machine. Neither man nor machine can make good decisions based on incorrect data (although I suspect that sometimes people are luckier).
Where the magic wheel of automation comes into play
First, it’s not magic at all.
Read on below
This is my semi-dry sense of humor that comes into play as much as I like to make fun of “hacks” that aren’t really hacks – they’re just the best course of action.
There’s nothing magical about the wheel. You, the human manager, are simply working with PPC automation in a never-ending process that includes these four (five if you count the bonus) steps:
- Discuss strategically.
- Develop a plan.
- Create the plan.
- Analyze the data.
- Repeat endlessly.
If you have a specific problem that you want automation to solve, the first thing you need to do is discuss what it is and how you are going to solve that problem.
This involves communicating with your client (or others who can inform this decision) and examining the various aspects of what you want to achieve.
What type of automated bid solution works best in this particular situation based on your data, your customer’s goals, and the goals you’re trying to achieve?
You then need to break your strategy down into actual steps that can accomplish what you have created, and then you need to create that plan.
As soon as everything is in order with automation, many people stop here: at the end of step three.
You pat yourself on the back while the automation works now.
But what if your strategy is missing a key component?
What if the automation doesn’t work?
Now, instead of stopping here, you need to analyze the data you’re getting from your plan (give yourself plenty of time, of course) and most importantly, be ready to have another strategic discussion to rethink and rebuild your automation.
A lot of people think they did their job because they are using ROAS targeting in this DSA campaign.
What happens if you rethink the structure of the DSA campaign?
What if you choose to separate this set of pages and instead target them with a CPA bid strategy?
Read on below
If you manage automation well, not only will you never leave your automated campaigns, you will constantly rethink them and test new structures, audiences, bid strategies, and more.
Automation doesn’t take the job away from you. it Changes your work.
Let’s play this through practically. Sometimes it’s easier to see it in a real-life scenario.
You want to test Smart Shopping on this account, but it doesn’t work.
You have tested your conversion data, uploaded complete and accurate audience lists, and achieved enough conversions for these products over time. What are you doing now?
Time to start step 1.
They connect to the client or run an internal audit to see what has to do with these products in order to provide a better subset of the products to be targeted.
Or you can get in touch with the customer to re-discuss the target audience for the various product groups they offer.
This is a tricky step because it’s so important, but it can be poorly done without the advertiser realizing they are not asking the right questions.
Read on below
Let’s say you plan a strategy with the customer and discover that too many products are being promoted based on their budget.
They had set up a smart shopping campaign targeting 5,000 products with a $ 10,000 monthly budget and in this case it didn’t work. You take this knowledge and develop a plan.
They’ll take their top 100 products and carefully group them into two separate smart shopping campaigns based on categories (since audiences for these two categories are very different).
You then create these two campaigns and push them live, ready to crack a cold one.
After you’ve got things going live, you’ll make minor adjustments to budget and ROAS goals, but give the campaigns a few weeks to see what happens.
You start by analyzing the data and learn a few more things.
You will learn that some of the products in that one category are doing exceptionally well, and in the end, you will be pushing this campaign harder.
Read on below
You have now learned from this new test that the second category you selected is not a great line of products for shopping campaigns.
Perhaps they are too expensive or not visually appealing enough to get clicks with the dynamic ads in Smart Shopping campaigns.
Regardless of what you learn, choose to test a different set of categories while your first campaign is running. You then speak to the customer and start planning the strategy for the next test … and you are back at the wheel.
Do you see how that works?
Unlike Thanos who takes a break on his farm, your job is never done. Even with automation, you can’t rest (and hey, I’m excited about that!).
Human-guided automation is the future, and this is one way we can manage it well in a process-oriented manner.
I hope this has been helpful to you and may the auction ever work in your favor!
Read on below
Screenshot by the author, March 2021.