To Automate or Not to Automate?

Nov 8, 2021

Scaling up Amazon ads brings with it the temptation to dip into automation software. Should you give in to it, though?

A case for both


Since the rise of Amazon advertising software in the previous 3-4 years, their usage has become a source of heated discussion among advertisers and sellers. Some of the principal arguments you’ll hear are things like:

“There is no such thing as “set and forget” – software cannot think and react strategically.”

“You’re leaving money on the table if you do not automate – it is simply too much data to handle for humans.”


Subscribe to our newsletter

“Setting ACoS goals is not enough – automation may just plummet your sales to reach a lower ACoS.”

“Humans are emotional and make mistakes; this won’t happen with automation.”


And they are all right. But to what extent?

Today I want to have a closer look at those arguments. Why and when are they right? What are they not considering? Who wins?


“Automation is not strategic” (Anti Software)

Absolutely. The software cannot know your underlying goals for a particular campaign or product. 


It can’t know if a specific product is still to be pushed aggressively to significantly increase its number of reviews and potentially dominate the niche in the future. Software can’t know if you are planning to reorder many units or not. Or how desperately you need to sell out to make space for another product. It can’t know your brand awareness goals. Do you even want to build up a well-known brand, or is the brand factor not crucial in your niche? And it most certainly cannot save you from a bad campaign structure.

However, this is also not what the grand share of automation software claims to achieve. It’s you who has to set the campaign structure, budgets, and bid (+ placement) levels according to your strategic goals. 

And only AFTER that is when automation software comes in. Only after those requirements are in place, and when the software will rightly claim the ability to eliminate some of your repetitive tasks, to inch closer to those targets day by day.

Humans should be strategic and software operational.


“Too much data for humans to handle” (Pro Software)

Fair point indeed, especially for larger accounts!


But even for smaller accounts, who among us can truthfully say they will go into a daily data deep dive to analyze the perfect bid level for each target? Or the perfect campaign bid multipliers depending on the different placement performances? Or analyze the average lagging sales attribution per product to better understand how the ACoS will improve over time?

Who will look at trends across multiple timeframes to make conversion rate predictions for every single target – and then constantly analyze how good those predictions turned out to be to improve them and set even better bids in the future? If you are doing this, I’d love to learn all about your time management skills!




However, at this point (November 2021), you also have to say most software cannot consider the data that truly matters as humans can. As of now, you won’t find software that can consider your keyword rankings, the share of voice, inventory levels, and profit margins all at the same time and prioritize their importance according to your overall goals.

For now, software and humans will have to get together to make the most out of our data.


“Setting ACoS goals is not enough” (Anti Software)

Very much so!


It’s true that almost all AI software only allows you to control the automation’s bid management by setting an ACoS goal for a specific campaign or a group of them. Coincidentally, one of my recent videos is just about that topic – neither ACoS nor TACoS should be your only KPIs based on which you will increase or decrease bids.

This is again where great campaign structure and naming systems come into place so that your actual campaign goals are much clearer to you. It will also allow you to go more into rule-based automation (in contrast to AI). For instance, you may want to increase bids based on impressions or orders. Or simply turn off automation for campaigns where the goal is not focused on profitability.


“Humans are emotional and prone to mistakes” (Pro Software)

Indeed we are, aren’t we?


It’s true – we may see a very high ACoS in the past week, panic, and decrease that keyword’s bid drastically. Without checking a longer time frame and realizing the keyword is performing exceptionally well. The past week’s ACoS is only based on very little traffic and missing sales from lagging attribution. 

Or we may want to set the bid to $0.4 and end up saving it as $4 by accident. With automation, this generally doesn’t happen. AI will always consider average lagging sales and look at all relevant time frames. At least the ones for which it has data, which is why a data storage service like Ad Astronaut may be worth checking out.

However, you should also consider this:

  • When humans make mistakes, they are usually limited to one or a few targets (unless you mess up with your bulk operations).
  • When automation goes wrong, it can quickly become very costly.

So, who wins?

Should you be using automation software for your Amazon ads?


Even though the Anti front has some decent arguments and the Pro arguments aren’t air-tight either, for me, the answer is pretty clear. If you have a proper campaign structure and know the goal of each campaign and product, then make life easier for yourself by using automation.

And if you have no clue about ads and campaign structure (and don’t want to hire somebody), then look for software. It should be one that can create a reasonably decent campaign structure for you and set a defensive ACoS goal. After all, some very defensive ads optimized by the software are almost always better than no ads at all.

To start with automation, I recommend using AI bid management software on “exact keyword” campaigns of “established” products, i.e., products where you just want to hold certain profitability. Why exact keywords? Because even AI cannot predict which search terms will get the traffic from broad or phrase keywords, let alone automatic campaigns.

There are almost no reasons you would not want to have some basic rules as a security backup in place. Unless you are selling luxury products, at least set up a negative exact keyword rule for search terms with 0 orders after 40 clicks. Or 30 if you are going to regularly check on what the automation does.

For most products with decent conversion rates (10%+), you may not even want to wait that many clicks. But again, this is just a nice-to-have security backup and not meant to replace looking at your search term reports.

So yes, use automation, but use it wisely!


About the author

David Z.

David Zimmermann is an Amazon Advertising enthusiast. Having worked for one of the very first Amazon Advertising software (Sellics) and one of the largest FBA brand aggregators (Razor), he has shared his learnings on YouTube and in his online course. David is also the publisher of the Amazon Advertising News Blog and has a particular interest in all things related to software and automation.

Produced by Sellers Alley

Terms of Service and Privacy Policy

@2021 Wild PPC Bunch. All Rights Reserved