I can’t really agree with your article…
In general what I tell my clients is that smart bidding is more time efficient. If I bill them 150€ per hour of exec work from my team then it might be better to use smart bidding and reduce the time spent on the account by a part.
Also smart bidding is not always worse in performance – in my experience especially smart shopping vastly outperforms manual shopping + display rm. By that I don’t mean it’s a 100%. It’s more like we try it with 20 clients and it increases performance for 5, doesn’t change performance for 10 and decreases performance for another 5. *BUT* this means that for the 10 where it is even it’s a productivity plus on our side since the shopping maintanance is slashed.
My recommendation for anyone inclined to try smart shopping is always to just try it and stick with what works best.
Whether you try it via experiments, replicating campaigns or simply swapping in settings has no statistically significant up or downside in our books.
What I find important as well is to ask why is works or doesn’t. How clean is the data we transfer to Google Ads. How is our attribution influencing campaign statistics. Why does tCPA work incredibly well while tROAS is utter trash in this campaign and why is it the complete opposite in another.
In my opinion smart bidding is great if it works since you save a lot of time and ressources and nothing really changes if it doesn’t work since you just return to where you were before. Therefore there is a net positive of trying it out.