Online Advertising Automation

The Disadvantages of Automated Google Shopping Ads Bidding

With our ever-changing technology and rapid innovation, digital agencies and in-house PPC teams are constantly trying to automate bidding and conversion processes to get ahead of the curve.

All the way back in 2007 Google Ads released their first automated bidding update with the “Conversion Optimiser” feature, which has since gone through many iterations. There are now dozens of automated bid strategies for each campaign type, from maximising conversions to maintaining set CPA targets.

Google are keen to recommend using these bidding strategies across entire accounts. But is full automation always the best idea? We don’t believe it is! With more and more ‘big’ agencies approaching automated AI (Artificial Intelligence) bidding on Google’s suggestion, do you think it’s right to trust this new technology with your campaigns? Do you expect more from an agency than simply flicking a switch and you paying them 10% of your spend?!

This article will discuss the reasons why Onefeed is the obvious choice over Automated Google Bidding for e-commerce businesses.

Automated Google Bidding – What Is It?

Automated Google Bidding is a technique that relies solely on machine learning or (AI) to maximise conversions and results based on a business’s campaign goals. What this means is that Google looks at historical data for automated bidding to work, such as location, demographics, user devices, the time of day, conversion rate and potentially other factors depending on the task.

Google states: “Automated bidding takes the heavy lifting and guesswork out of setting bids to meet your performance goals”. At first glance this sounds fantastic, another process that can be fully automated and one less thing to worry about. However, e-commerce businesses should carefully consider all the implications and dangers of automated bidding before taking this route.

Hidden Risks:

The factors that Google doesn’t take into consideration with automated bidding include a whole array of different issues. For instance, machine learning can only look at the data you’ve given it. Physical events that affect bidding and ad performance such as, flash sales, weather, seasonal trends and media coverage can all vastly affect a business’s performance. Automated bidding with Google Shopping Smart Campaigns doesn’t take into account any of these physical issues that could suddenly arise unexpectedly throughout the year.

Another danger of automated bidding is that Shopping Smart Campaigns doesn’t know your business inside out, you will still have to adjust some parameters manually to increase performance. So, the bidding itself is automated, but as a business you will still need to monitor your campaigns to evaluate if the bid strategy is meeting your desired goal. Google’s broad data might not exactly reflect the target audience or your company’s strategy.

Google’s most recent Shopping Campaign type is Smart Shopping, as discussed above this campaign type effectively takes over both your display re-marketing and the entirety of your shopping campaigns. This on the face of it seems like a good idea. This however means that you have zero visibility or options to change any of the settings for this campaign. Google is in effect a black box that just spits out clicks to your site. There is no means of ad scheduling or even adjusting device bids; it’s all left to a computer to decide. There’s a massive risk involved in passing such an important element of your advertising for your business to a machine and a load of algorithms. What happens if you’ve got a load of new stock that has never been on Google before or seasonal items are now in stock that have been gone for 9 months? How do you push the bids on these items? You can’t! What if organically you sell the latest greatest product to hit the market, but you’re getting no traffic from Shopping? Can you change the bids on this item? You can’t! Can you set negative keywords that you know are never relevant? No, you can’t! Everything that you’d perhaps want to adjust, and control is hidden in the black box.

It takes time for Google to learn a company’s campaign performance and to obviously learn from its own mistakes. Likewise, there is every chance the AI will push items that you’d never want to push, but you’ve got no control. If you were after a quick fix to gain conversions and increase site visits, then you would be mistaken. Automated bidding requires historical data to optimise a company’s goal or objective, therefore a company just starting out with Google Ads would consequently find it very difficult or nearly impossible to use automated bidding. Google does state that it requires data to make decisions, however a fresh account with nothing is going to really struggle to make any impact whatsoever.

Automated bidding does take care of a company’s bid optimisation and this means a company can use “set and forget” campaigns which can be a useful tool to save time, however, Google requires a lot of broad data to do this. Here lies another problem as Google’s vast amount of data means the client has no visibility or control over what data is being used. Which ultimately means AI based bidding never truly has the agility to maximise performance. In order to do this, you will need us.


Why choose us over Automated Bidding Systems?

Unlike Google, we’re not fully automated, meaning a lot of the dangers that were outlined above do not occur. We use a combination of bespoke management tools, creative analytical minds and machine learning suggestions to drive client performance. We have built our own platform in-house called “Client Centre” which is a sophisticated management platform specifically designed to enrich and improve client feeds, manage bids at the product level and control the entirety of the Google Shopping performance. Client Centre is constantly being developed and fine-tuned by our very own in-house UK based software developers in order to continue being a market leader within the industry. It is the combination of Client Centre and our bidding experts that analyse and adjust the bids accordingly to find the Clients sweet spot between increasing revenue and keeping Cost of Sale under control.

A Key Statistic:

“Our merchants average an increase of 63% in revenue and 47% increase in conversion rate”.

This would not be possible if it wasn’t for our expert knowledge in Google Shopping and managing campaigns. Often with automated bidding systems, the data and algorithms are not sophisticated enough to adjust the bids appropriately, so they will never achieve an average increase in revenue as stated above. With our Tracking software, we’re able to constantly view client’s sales in real-time. This means our experts can analyse live sales data and make appropriate adjustments to the bidding, something that an automated bidding system cannot deliver.

As Google Shopping management experts, we only bid accordingly based on previous bid performance and future targets, whereas automated systems will try to spend their way out of trouble by continually bidding higher per click. Everyone has seen the case whereby you increase bids on an item and all you get is the same click volume, but a massively increased cost. There are massive savings to be made by adjusting the bids in line with performance and not overbidding. Our bidding system uses intelligent data gathered from the Client Centre based on 25 different performance metrics to optimise product listings. This element of Machine Learning makes suggestions of where product-level bids need to be and can ultimately be ignored by our account managers if the suggestion isn’t acceptable. This isn’t relied upon entirely, but our AI suggestions engine does help to keep campaigns on target. However, one of the real benefits we have is our continual account manager intervention.

Our Account Managers are always using the Client Centre to adjust and reassess product-level bids based on the performance statistics. As well as looking at all the in-house data to determine their decisions, they also consider seasonal trends, flash sales, the weather and any other human-based factors. Human and physical factors can greatly influence a client’s performance, as previously mentioned when discussing automated bidding systems. Client Centre is essential in making well-informed decisions by our account managers but being able to think about all considerations for the client is where we flourish.