-
Automation Vs Non-Automation
Just curious as to when you guys would actually run manual campaigns these days. I’m sure they have their value – particularly for brand new accounts with no data – but in general there seems to be far too many data points to effectively manage manually, and performance is going to be significantly better almost all of the time when allowing for automation because of this.
And just to clarify, by automation I don’t mean just letting Google do everything, I’m talking mainly with regards to bidding.
What are your guys thoughts?
-
1 Reply
-
I run manual cpc for clients in two instances:
1) Brand new accounts
2) Niche businesses or small budgets that have very little conversion data per month.
On a case by case basis I’ll also run manual cpc for extremely expensive CPCs, $40+ average non-brand cpc, and I’m usually running it as a side by side experiment with a smart bid strategy.
-
I tend to use tCPA since it offers a blend of automation and manual control.
I have a report I run each week which checks performance across close to 20 different dimensions of data – the stuff Google claims is being handled by their machine learning – and for any given dimension (ie performance by time of day or location), there are usually about 50% or so of accounts that have significant discrepancies & optimization opportunities without manual intervention.
So with tCPA I can still use things like bid modifiers to push the automation in the right direction when needed.
If you’re using Max Clicks/Conversions, another good “health check” is Lost IS to Budget. If that’s high, it means those bid policies aren’t doing their job, as literally their entire purpose is to bid based on your daily budget.
-
– Net new accounts
– Low conversion volume accounts
– Instances where manual control over specific keywords / phrases provides better return vs. automated bid strategies. These usually get tested quarterly to verify that manual is still the winner.
– In scenarios where I want very granular control over very specific KPIs and Google doesn’t have an automated bid strategy to align with that yet (i.e. I need to drive a specific amount of profit per month and the margins between products / product groups vary too much to successfully segment them within a larger feed).
– In scenarios where I’m running campaigns explicitly to get auction insights information (i.e. very limited brand spend just to see who is surfacing). I actively want to pay the cheapest possible, and don’t care about clicks, conversions, or anything other than max impressions for lowest cost.
– On the occasional campaign that I want to ‘set and forget’ that’s very low volume, but high intent and high conversion. I can just key in high manual bid and not have to worry about the algos shifting bids over time. I still usually test these against automated strategies eventually, but more often than not there isn’t enough data to properly train the algos. -
If you are just running automated bidding campaigns, what actual value are you providing to your clients that they can’t do themselves? What, filtering negative keywords lol.
If you are a performance agency, you run manual.
-
I would never run a campaign on Manual CPC for several reasons.
i. Your competitors are and their automation will always optimize bids around your static target.
ii. Your manual CPC will require constant updating based on the market which can differ from region, to audience, to location, to demographic, to keyword, and you have far better things to do with your time than automate tens, hundreds, or thousands of things regularly.
iii. Your manual CPC won’t directly reflect your business goals. Do you need a certain CPA to be profitable? Do you have a ROAS goal? CPC’s are a very complicated and diverse fraction of what ultimately achieves your marketing – and business – objectives.
At the end of the day you should be focused on strategy and not incredibly complicated and time-consuming minutiae and that is why automation is critical.
The only time I see Manual CPC work well these days is for an extremely limited number of Brand campaign issues, EXTREMELY niche companies where data points are hard to find (specific kinds of cancer research/applicants for example) where even competitor info is going to be spare for the system, or .. well that’s it. Otherwise I just see it when agencies try to prove value by putting in a thousand changes a day with out-dated third-party automation.
Best,
Google FTE
-
I steal accounts from agencies all the time with my manual bidding strategy.
Log in to reply.