On June 26, 2018, I hosted a #PPCChat discussion about using automatic optimization features in PPC (such as CPA bidding or the “optimize” ad rotation setting).
#PPCChat is a weekly Twitter chat that occurs every Tuesday at 12 PM ET, touching on a wide range of topics related to PPC.
Throughout the conversation, a number of industry folks offered some great insight into when to use automation, when to avoid it, and how to approach testing it. Here’s a recap of highlights from the discussion.
Almost Everyone Uses Automation
Out of 34 respondents, only 3% said they never use automation. While respondents leaned toward using auto-optimization features with caution, most were willing to embrace them on some level.
Q1 Do you use automation features in your PPC management? #ppcchat
— Tim Jensen (@timothyjjensen) June 26, 2018
Which Automation Features Work Well, and Which Work Poorly?
Responses covered a wide range of opinions on which automation features work well and which ones individuals were less likely to trust. Ultimately, the varying experiences here point to the fact that no two accounts are the same, and what works for one client won’t necessarily work for another. In general, there was some consensus that Facebook’s bid optimization features tend to be better than other platforms.
A2: It is interesting to me how much we all just let FB do the auto bidding pretty much without too much questioning, but generally not the same attitude on AdWords.#ppcchat
— Julie F Bacchini (@NeptuneMoon) June 26, 2018
A3: Don’t know if this counts but I’ve never had any luck with dynamic display and think they look awful! Very much for the lazy/uninformed in my opinion. I think it’s important to have a certain level of control. The “set it and forget it” mindset is ALWAYS BAD #ppcchat
— Mark Gustafson (@markpgus) June 26, 2018
A4: If it comes to running basic algorithms based on data, then all of them equally. Once we get into the predictive stuff, then I trust those with the best data. E.g.,for optimizing to demographics, I trust FB more than Google because much better data. #ppcchat
— Kirk Williams 🚴 (@PPCKirk) June 26, 2018
A4. I know this sort of a cop out answer but my trust in the platform depends on the client. There's just certain accounts I trust to hold the reigns based on historical data. Whether that's AW, FB or LI. BUT – if I had to choose one then Facebook. #ppcchat
— Tim Halloran (@timmhalloran) June 26, 2018
Accurate Data & Sufficient Data Are Crucial
In order for automation to work properly, sufficient volume is necessary. While the ideal number of conversions varies by platform, you should be aware of thresholds and be cautious about applying automated bidding features in low volume campaigns.
In addition, the accuracy of data is also vital for proper performance. Since automated bidding works from a raw conversion count, if conversions were miscounted for some reason, performance will be thrown off.
Definitely. Limited data or bad data is a kiss of death. Which is I'm skeptical of all these AI & Machine Learning tools. #ppcchat
— JD Prater (@jdprater) June 26, 2018
Cautions for Using Automation
You should never just mindlessly turn on automation in your campaigns without a background understanding of how it works in a particular platform and how it is likely to play with your particular account. Auto-optimization features should never be treated with a set-it-and-forget-it mentality. They’re here to help you do your job better as a PPC manager, but changes can happen quickly, and you still need to be monitoring performance closely. Participants had several words of warning for using automation.
I've learned to be cautious with new tools, like new bidding strategies. For example, ROAS bidding works well now, but was horrible when it was first introduced. #ppcchat https://t.co/1akG8lNSyv
— Martin Roettgerding (@bloomarty) June 26, 2018
A6: You need to know your stuff! Automation is NOT an easy solution that takes no knowledge or talent. Having a deep understandjng on PPC (manual strategies) really is important even if you end up relying on automation. You need to know WHAT the strategy does! #ppcchat
— Mark Gustafson (@markpgus) June 26, 2018
What’s Next?
We ended the discussion by looking ahead to the future of automation. General predictions included a move away from manual settings, more actions taken behind a “black box,” and smarter integration of data.
A7 I hope I'm wrong, but I see platforms replacing some manual settings with automation, they're so sure the algos can do better job. I don't dispute that, but I still need settings to divine what's going on & explain to client #ppcchat
— Glenn Schmelzle (@heyglenns) June 26, 2018
A7. I know predictions and wish list are not the same, but I HOPE that the future of automation means more time for strategy because the algos are working more effectively. I FEAR it will mean more time spent determining why performance shifts, to explain it to clients. #ppcchat
— Emma Franks (@akaEmmaLouise) June 26, 2018
A7: Data will become the new currency for optimization as AI (machine learning and predictive analytics) are readily integrated into tools and platforms. #ppcchat
— Christi Olson (@ChristiJOlson) June 26, 2018
A7: #ppcchat the future analytic optimization IMO, especially when voice abilities come along, and rules that can be used across channels (ie video, search, social, etc)
— Jon Kagan (@JonKagan) June 26, 2018
What’s Your Take on Automation?
Overall, this topic generated some in-depth discussion from PPC professionals on the benefits, hazards, and future of automation in campaign management. Ultimately, PPC managers need to embrace the fact that automation is here to stay and will become more and more prevalent while knowing when to strike the balance with manual oversight.
What are your thoughts on auto-optimization features in PPC management? Continue the discussion in the comments below!