Browsing various PPC blogs (including mine) can create an impression that automation in the PPC/digital marketing world can solve all major advertisers’ problems. But is it really the case? Unfortunately not!

In this article I will look back at the last 7 years when I was heavily automating anything I could when running (mainly) PPC search campaigns. I will look at various types of available automation tool/approaches as well as into not-so-positive aspects of PPC automation.

But before we start, let’s answer the two most basic questions:

1. What is PPC automation? It’s a set of processes designed to remove manual tasks from daily campaign management.
2. What are available automation tools? Answer to this question deserves a separate section below.

Available Automation Tools

A few years ago, when someone said “automation”, my first association was “scripts”. Something very geeky which only a few people build. However, there are quite a few automation “tools” available, not just the scripts. The order below is set based on the ease of use from the easiest to the hardest. Hard ones usually means that some sort of coding knowledge is required. Here we go:

1. Automated rules – you can use automated rules to e. g. pause certain KWs or increase certain bids. These rules are quite easy to set up directly in Google Ads or in Bing Ads.
2. Automated bid strategies – you can use automated bidding to let Google, Bing, or Meta (long live Facebook!) set your bids based on various incoming signals.
3. Automated campaigns – Google has been making huge leaps forward in the recent years introducing more and more Smart and Max “whatever” campaigns which can be set up with minimum inputs.
4. Feed management platforms – if you are heavy on shopping and you are managing lots of feeds which need adjustments, you will likely end up using a feed management platform such as Mergado.
5. Creative management platforms – similar as above but for display creatives.
6. Bizarre tool #1 Power Query – I am mentioning this because I was quite heavy on Power Query, which is an Excel add-on which allow you to manipulate tables and their data in bulk with no heavy coding knowledge required. Very useful for search campaigns. Look into my older articles.
7. Bizarre tool #2 BigQuery – If you know some SQL, you can use it to manipulate massive tables and manage your search campaigns through SQL in BigQuery. The good thing about BigQuery is the fact is that you can sync your accounts from Google Ads to BigQuery (only 1 way though). Getting manipulated data from BigQuery back to Google Ads is on you!
8. Scripts – everyone knows Google Ads scripts but did you know there are also scripts in Microsoft Advertising (aka “Bing Ads Scripts”)?
9. APIs – if scripts are not enough for you, you can start making API calls to Google Ads, to Microsoft, or even to Meta. API integrations are the most demanding ones – you need to write a piece of code, run it on a server, and hope the code does what you envisioned.
10. Reporting tools – there is a ton of tools which allow you to put all your marketing numbers together so you know whether your marketing dollars are spent wisely.

We can debate whether “automated rules” are really “automation”. Similarly, we can also question “automated bidding” when you literally do nothing and all other competitors can use the same thing…

Okay, with all the automation tools being out of the way, let’s now focus on automation challenges and issues.

#1 Automated Campaigns (Smart and Max “whatever”)

I am not going to lie – I hate these. My main issue is lack of control and transparency. As a bonus, you are forced to buy traffic which you would not normally buy because… guess what… you have very little control about where your ads are being shown. It’s like giving Google a blank check and hoping for the best.

#2 Automations Do Not Generate More Volume

I’ve never seen an automation which would directly generate more volume (e.g. more leads). Usually, the benefits for generating more volume are indirect. You save time so you can spend the saved time on value creation. You save cost so you can allocate the savings to something else (e.g. buying more leads).

#3 Benefits Are Very Hard to Measure

So you’ve just spend 20 hours on writing a fancy script or you may have even paid someone to write for you. What is the ROI of this exercise? Unfortunately, it’s hard to tell. You will have to run the campaign with and then without the script a few times to find out. Do you have time for this testing? Probably not. It’s also possible that landscape will change you will learn nothing from your testing anyway.

#4 Automations Are Only Good for Very Stable Environments

Let’s say you are negating bad keywords on ad group level with a script. You are about to change the campaign structure. This will likely create a need for your script to be edited or it might be just faster to write a new script from scratch. If you are frequently making major changes to your campaigns, then you will also have to keep your automations updated every time you make the changes to your campaigns. Frequent changes to your existing automations always increase chances of making a mistake during your updates.

In other words, unless the problem which I am about to automate is going to be exactly the same for the next 6 months or more, I usually don’t recommend to even start thinking about automating if time required for building a solution is more than 20 hours.

#5 Automations Are Making You Less Flexible

Imagine, you’ve just deployed a super complicated solution to a massive problem in your campaigns. Two weeks later, you are realizing you need to make certain adjustments to you campaigns for various edge cases or outliers. Well, guess what – edge cases and outliers are very difficult to account for in your automations. It might simply be impossible.

So a human can address the edge case with a single click in web interface of Google Ads in 20 seconds, but adjusting an underlying script for the edge case may take hours or days.

You (or other members of your team) will also be afraid to make any manual changes to the campaigns because they don’t know what’s going to happen when the automation runs again. Maybe the changes will be reverted back by the automation so the human inputs will sooner or later simply negated.

#6 Scripts Are Always Tied to Their Creator

If you’ve hired someone to write a script for you, you may get stuck with the person forever. Editing an existing script written by someone else is usually very difficult. It might be just faster to write a new script from scratch, or you will need go back to the creator for every little adjustment.

#7 Your Campaigns Actually Need Human Touch

You may not realize this but your campaigns actually need humans to make decisions or choices which simply cannot be scripted. It may not be happening every day when human input is needed but imagine there is a script potentially interfering with these decisions.

What Is Automation Good for Then?

So far, this article has been very negative on automation. The question is whether there is still room for any custom automations beyond what Google/Microsoft/Meta are already offering us?

I do believe that future of advertising will be about feeding the right data to the networks. The “right data” can be:

1. Shopping feeds – building the right feed from messy inputs can be quite a challenge requiring lots of custom development.
2. DSA feeds – similar as above.
3. Targeting specs – e.g. changing targeted locations every day based on external location lists.
4. Audiences – e.g. feeding contacts from your CRM to Google Ads.
5. Offline conversions – e.g. feeding offline conversions in cookieless era back to networks.
6. Highly standardized and predictable workflows tailored for one account – of course, you can still automate any workflow which repeats frequently enough to justify time spent on a developing an automated solution assuming this solution will not interfere with changes done by humans.

I also cannot forget the reporting piece – no matter what fancy automated campaigns you run, you will always need a custom report. To build the report, you will need tools or scripts to automate your ETL data process especially when:

1. When you advertise on more than one network.
2. When not all conversion events can be measured online with pixels (which are dying anyway).

Conclusion

When I started to take on automation projects, I was very eager to automate whatever I was asked for while disregarding all earlier mentioned negatives. I simply took all projects as challenges which had to be solved no matter how much business value was generated. Nowadays, I am still automating BUT I am very pragmatic in choosing or recommending what to automate and what not.

And that’s it! Agree or disagree? let me know in the comment section below!

Do you need help with automation?
HIRE ME