Amazon DSP Automated Bidding: What Changed

From 17 August 2026, Amazon removed the Base Bid setting from all new Amazon DSP line items. If you build a fresh line item today, you no longer set the maximum bid yourself. Amazon DSP automated bidding now decides it for you, using an AI-driven model that reads signals in real time.

For anyone who has spent years fine-tuning bids by placement, audience and daypart, this is a meaningful shift in control. It is also entirely in keeping with the direction Amazon has been travelling for some time.

The instinct among many advertisers will be to worry. We would frame it differently. The lever has moved, not disappeared. What matters now is understanding which levers remain in your hands, and pulling them harder.

What Amazon actually changed

The change is narrow but important. On new DSP line items, the manual Base Bid field is gone. In its place sits an automated bidding system that determines the maximum bid on your behalf, optimising towards the outcome the line item is set up to deliver.

This does not mean you have lost all influence over spend. It means the specific act of capping a bid manually is no longer part of the setup. Amazon's model now interprets performance signals and adjusts bids far faster than any human sitting in the console could.

It is part of a broader, deliberate push towards automation and AI-driven advertising across Amazon's ad products. The Base Bid removal is one visible expression of a much larger strategy.

Why Amazon is doing this

Automated bidding is not a cost-cutting exercise dressed up as innovation. Machine-led bidding genuinely can respond to auction dynamics at a speed and granularity that manual management cannot match.

Amazon holds an enormous volume of first-party shopping and conversion data. Handing bid decisions to a model that can act on that data continuously is, on paper, a strong proposition. The question is never whether the technology is capable. It is whether the inputs feeding it are good enough.

The trade-off brands need to understand

Here is the part that gets missed in the noise. Less manual control over bids raises the importance of everything else.

When you could shape outcomes by nudging bids, weaker inputs elsewhere could be partly compensated for at the auction. That safety net is thinner now. The automated model can only optimise as well as the strategy and signals you give it.

That puts real weight on the inputs still firmly in your control:

• Audience selection and targeting: who the model is allowed to bid on in the first place
• Creative: the asset doing the actual persuading once you win the impression
• Budgets and flighting: how much the model has to work with, and when
• Measurement: how you define and read success, which shapes what the model optimises towards

Get these right and automated bidding has excellent raw material to work with. Get them wrong and no amount of clever bidding will rescue the outcome.

Where automated bidding struggles: new product launches

Automated bidding is not inherently good or bad. Its performance depends heavily on context, and one context deserves particular caution: brand-new product launches.

At launch, a product has little or no conversion history. The automated model has nothing meaningful to optimise against, because the data it needs simply does not exist yet. Leaning on automation in that vacuum can burn through budget while the system tries to learn on your money.

This is where a considered strategy matters most. A launch is precisely the moment to think hard about audience construction, creative strength and how you sequence spend, rather than trusting a model to find its feet unaided.

If you are still weighing where DSP fits alongside your search activity, our guide to Amazon DSP vs Sponsored Ads is a useful starting point before you commit launch budget.

What this means for brands

The practical takeaway is not to fear automation. It is to earn the most from it by tightening everything around it.

Treat your inputs as the strategy. With bids automated, audience, creative, budget and measurement are no longer supporting acts. They are the performance. Audit each one before you assume the model is underperforming.

Handle launches differently. Do not point automated bidding at a product with no conversion data and hope. Build audiences deliberately, invest in creative that can carry the message, and be prepared to feed the system before you scale it.

Sharpen your measurement. Automated bidding optimises towards the goal you set. If your measurement framework is loose, the model will happily optimise towards the wrong thing. Amazon Marketing Cloud is central to this, and the recent decision by Amazon to remove the paywall on its most powerful AMC datasets makes rigorous measurement more accessible than ever.

Keep strategic control. Automation changes the job, it does not remove the need for expertise. The brands that win will be those treating the model as a fast executor of a strong plan, not a substitute for having one.

The wider DSP opportunity is expanding at the same time, from premium inventory such as the Amazon DSP integration with Channel 4 CTV to richer measurement. Automated bidding is one piece of a broader canvas that rewards brands who plan across all of it.

FAQ

Does automated bidding apply to my existing Amazon DSP line items?

The change removes the Base Bid setting from new line items created from 17 August 2026 onwards. Any new line item you build now uses Amazon's automated, AI-driven bidding model to determine the maximum bid.

Is Amazon DSP automated bidding worse than manual bidding?

Not inherently. A machine-led model can respond to auction signals faster and more granularly than manual management. Its results depend on the quality of your audiences, creative, budgets and measurement, and on the amount of conversion data available. It struggles most when that data is thin, such as at launch.

How do I stop automated bidding wasting budget on a new product?

Avoid relying on automation alone at launch, when Amazon has no conversion data to optimise against. Focus on deliberate audience selection, strong creative and controlled budget sequencing, then let automated bidding scale once meaningful signals exist. Working with a specialist team helps you structure this properly.

Where do brands still have control under automated bidding?

In everything that feeds the model: audience selection and targeting, creative, budgets, flighting and measurement. These inputs now carry more weight than ever, because there is less manual bid adjustment available to compensate for weaknesses elsewhere.

Make automation work for you, not against you

Amazon's move towards automated bidding rewards brands with strong strategy and punishes those relying on the platform to think for them. The advertisers who thrive will be the ones who master the inputs that remain in their control.

As an Amazon Ads Advanced Partner, our Advertising team manages Amazon DSP every day and helps brands get the most from automated bidding while keeping firm strategic control. If you want to know that partner status is more than a badge, read what Amazon Advanced Partner status really means.

Talk to our Advertising team about your Amazon DSP strategy.