Top PPC experts view artificial intelligence (AI) as among the top trends to watch out for in paid search marketing this year.
While AI offers a wide range of exciting developments in the search industry, it also brings about new changes that some practitioners may find intimidating.
In this Search Engine Nerds episode, I interviewed Frederick Vallaeys, co-founder of Optmyzr, to discuss the role that AI plays in paid search.
Vallaeys, who previously worked at Google for 10 years and served as an AdWords Evangelist, also explains how paid search automation can help PPC managers efficiently improve their campaigns today and beyond.
Why do you think automation is growing at the rate it is?
Frederick Vallaeys (FV): I think the underlying thing is the technology, right?
You have very simple automation like an “if this, then that” type of rule but what’s really the automation that people like to talk about is the machine learning, the automated bidding.
We’re just at this pivotal point in history where all of a sudden the base of technological improvement is so fast that we can’t even really comprehend as humans how quickly this is improving.
So if you look at PPC bid management systems five years ago, they were OK, but they often kind of missed the boat, and then things fell off and you weren’t so happy.
Nowadays, it’s much better. It’s good enough that it can often replace humans.
But even if you still think that you’re better at it today, well six months from now because of the pace of innovation and because of Moore’s law, the capacity of computers doubles every 18 months, when we’re doubling on something that was already really good, all of a sudden we’re leapfrogging that technology.
That’s why it’s becoming so important – because even if you think you’re good as a human today, you should be ready for the scenario where six months from now, you’re just blown away by how good the technology is.
You had said that one thing people should be doing right now is getting ready and preparing for the growth of automation and AI. What does it mean overall, for those that currently are in the marketplace?
FV: First, as humans, don’t try to compete with something that a machine can do more efficiently and certainly much cheaper.
What is it that you, as a human, can still add a tremendous amount of value to? That’s typically the creative stuff, but it’s also dealing with automation.
Automation nowadays is really based on expert systems. There’s still no automation that you say, “Hey, go and advertise on AdWords. Here’s a check, go and do it.” That’s not how it works.
You have to figure out which budget management system to use, which keyword automations to use, how you’re gonna deal with ad text testing. So you, in a sense, become like the quarterback.
You have to figure out who are the players that you want on your team, and how are you gonna execute on every play based on what you’re seeing in the market – putting those pieces together. I think that’s going to be very necessary for a while.
The final component is now that you’ve hopefully won the game for your customer, your client, your boss, it’s still about talking to them and understanding what is the fundamental business and how does that impact the next round of testing that we have to do.
If results aren’t amazing, having the empathy to talk to the person and explaining why maybe things didn’t go the way they should’ve gone.
With AI being such an important part of Google as a whole, what else do you currently see them doing on the PPC side of things for us to prepare for?
FV: I find it so exciting that Google actually makes artificial intelligence available in the cloud. They used to have a Google prediction API and now it’s the Google Cloud machine learning API.
Why this is so exciting to me is it enables any business to leverage the same sorts of technologies that big companies like Google have, but in their business context.
Google is pushing very hard for advertisers to give up a lot of the control and the micromanagement because they know computers can do a lot of that stuff better, more efficiently, and at a lower cost ultimately.
But the one thing that’s often being missed is there is still something unique about many businesses. And you as the business owner or the advertising agency, you know that stuff.
You know the types of things that a machine maybe should be looking at. If you think about machine learning, it’s basically pattern recognition, right? Now, the patterns that the system is recognizing is based on what the vendor has access to.
In Google’s world, they have access to who is the person, what was the search that was being done, what was the time of day, those types of aspects.
But what if you as the business owner, you sell event tickets, and you know that there’s a really strong probable correlation between whether it’s raining and how many baseball tickets you sell.
Here you have access to their machine learning system and you can actually plug all of that data into it. You can come up with a better starting point, then Google can still do its magic on top of that, but at least you’re kind of guiding them closer to what is the right solution.
Do you have any other examples of how PPC pros can utilize this AI data or automation to make their work easier and more targeted?
FV: I think the AI is really on the high-end of the sophistication spectrum and one of my recent posts on was really to kind of tell people it doesn’t have to be that comprehensive or that advanced.
In Free AdWords Scripts, a website that we own, there are free scripts you can pick up and install and 15 minutes later you’re up and running.
You really don’t have to do a lot of coding, you just copy-paste, change one or two variables, and that’s it. So that’s really where I want people to start.
On the high end of the spectrum, I think it really is about bringing in your own business data into the mix and starting to look for those patterns.
The beauty about machine learning is that it’s not on us, the humans, to figure out “this is what’s going to be meaningful.”
It’s about us throwing a lot of data at the machine and letting the machine figure out what is important.
If you think anything could have an impact, test it. Ask the question, and see what comes back.
What are some new scripts you’re currently working on and that you’ve released lately that could help out PPC managers across the board?
FV: We put out a script when Google changed the overdelivery mechanism. We wrote a simple script that brings back your 20 percent budget cap, or any cap in between 0 and 100 percent.
The beauty of this is that it took 30 lines of code to undo a change that some people didn’t like from Google.
That’s to me, the power of scripts. It doesn’t take a very long time to get things to work the way that you’re used to.
Now, I actually do recommend people leverage what Google has built, because we have another script that’s really popular, which sets the daily budgets based on how much budget is left in the month, and it accounts for day-of-week patterns. We look historically to know that Mondays maybe have 50 percent more volume than Sundays. And so based on that we can over-allocate budgets on those days.
We’ve also written scripts that are really popular to build entire ad campaigns so, mixing and matching a Google spreadsheet of all your products, all your services, with a template, and then the script just runs and basically updates your inventory on the fly. The whole campaign is inventory-driven – a good alternative to DSA type campaigns.
The way we think about scripts, it’s mostly really binary type decisions. If it’s something where you don’t have to do a lot of thinking about what is the action you will take as a result of something, that is really good to script.
To listen to this Search Engine Nerds Podcast with Frederick Vallaeys:
Think you have what it takes to be a Search Engine Nerd? If so, message Loren Baker on Twitter, or email him at loren [at] searchenginejournal.com. You can also email Brent Csutoras at brent [at] alphabrandmedia.com.
Visit our Search Engine Nerds archive to listen to other Search Engine Nerds podcasts!
Image Credits
Featured Image: Paulo Bobita
