Vendors of PPC bidding apps frequently label their automated PPC bid management tools as “artificial intelligence” or AI. (E.g., AI-powered ad optimization, smart Amazon advertising AI, AI-based algorithm for Amazon advertising optimization, etc.) When compared to Amazon AI, however, these “vendor PPC AI” tools are not all that intelligent.

Vendor PPC AI tools know only about bid performance for a few products. Amazon AI knows an immense amount about all the shoppers behind the bids, all the related products and all of the associated winning and losing bids. The vendor PPC AI development effort is trivial when compared to the massive 20 year+ development experience of Amazon AI. Amazon employs thousands of the world’s smartest AI scientists. PPC AI vendors employ very few. There is simply no comparison between vendor PPC AI and Amazon AI.

To be clear, to use automated vendor PPC AI is to not use Amazon AI. They are mutually exclusive.

900%™ leverages Amazon AI by using maximal placement bid adjustment accelerators. In other words, 900%™ produces a base bid that always allows either Top of Search or Product Pages (or both) placement accelerators to be set at their respective 20X or 15X maximums. The approach allows Amazon AI to determine when to bid strongly, how much to bid and when to hold back. Amazon AI’s determinations are based on the deep learning of Amazon AI. 900%™ maximizes the role of Amazon AI on placement bidding. Every campaign can be highly profitable.

Most vendor PPC AI solutions ignore Amazon AI altogether. They back into keyword bid determinations through a process of bidding trial and error. They have no ability to replicate placement bidding accelerators. They lack any knowledge of the Amazon shopper. 

A few PPC AI vendors may add campaign placement bidding accelerators on top of their fixed keyword bids but this always leads to sub-optimization. The base bids will always be too high which results in average bids being too high overall and particularly expensive in Rest of Search. Furthermore, the Amazon AI accelerated bids will always be too low because of the mismatch between keyword-based bids and campaign-based accelerators. Campaign accelerators with PPC AI need to be capped to protect against keyword overbidding. The result is severely sub-optimal.

In short, using vendor PPC AI for Amazon sponsored products ads is expensive. The monthly charges are high, advertising dollars are squandered, and the overall opportunity cost is large.

900%™—solely based on Amazon AI—is far more profitable because it minimizes the base bid and allows the full maximum bid available to Amazon AI to be determined by campaign conversion rates and margins. 

What About All Those Ancillary PPC Bidding Features? 

Are any of the following ancillary features worth it when PPC AI bidding is sub-optimal in the first place?

1. Custom Automated Bidding Rules

Custom Bidding Rule #1: If keyword has low ACoS, increase bid or If keyword has high ACoS, decrease bid. These are both bid-by-keyword statements and only relevant if you are using PPC AI tools. The rule makes no sense when using Amazon AI. With Amazon AI, base bids and placement bidding accelerators are made at the campaign level. With 900%™ if the campaign ACoS is low, the base bids and accelerators will be adjusted upwards in a coordinated fashion. If individual keywords are found to be performing poorly they can be removed altogether or moved to a more hospitable campaign structure. Similarly, as the campaign ACoS approaches the user-defined target ACoS, the base bid and the placement bid accelerators are adjusted accordingly to keep the ACoS at or less than the specified target.

Custom Bidding Rule #2: If no orders or low conversions, pause keyword. Again, this is a bid-by-keyword rule. Nevertheless, it’s easy for this automated rule to decide in error. The decision to pause a keyword can’t be separated from the costs. If the accumulated costs to keep advertising a keyword are less than the gross margin of a sale, it may make sense to not pause it. This custom rule is of very small value, and possibly of no value at all. The rule likely only adds to the overall sub-optimized dollar waste and opportunity cost of vendor PPC AI.

Custom Bidding Rule #3: If out of budget, increase budget. What is the point of a budget if it is to be automatically increased? A budget is set for a variety of reasons. Presumably reaching the advertising budget triggers an evaluation of the underlying causes. This is the prime reason for a budget. If the logic of an automatic increase is sound, then simply set the budget much higher. This rule solves a nonexistent problem.

Custom Bidding Rule #4: If one campaign is outperforming another, rebalance the allocation of ad dollars among campaigns. One campaign will always outperform the others, but that doesn’t mean the others are poor campaigns. This rule is an automatic throttling of lower ranking campaigns. If all the campaigns are increasing ad-based revenue while lowering ACoS and staying under their target ACoS, then rebalancing may not be the wisest business decision. A manual re-evaluation of the overall budget is a superior action.

2. Day Parting

Amazon AI has natural day parting. When you know nothing about the shopper and the market environment—as is the case with all vendor PPC AI bidding apps—you need to find ways to protect against overbidding. While day parting may protect from wasting some bid dollars, it misses authentic off-hours customers. It is a sub-optimal feature caused by the PPC AI vendors lacking any data beyond keyword bidding history.

3. Keyword clustering

Keyword clustering is fancy math in search of value. Most often the valuable clusters of Amazon keywords are not based on letter similarity but rather conceptual meaning. Unless you have thousands of keywords it is easy enough to manually cluster keywords. It is also helpful to manually cluster keywords, not only because of the high cost of automation, but also the quality of the clustering. Furthermore, doing your own clustering teaches you about the market, something that you don’t learn well with algorithmic clustering. Finally, 900%’s™ use of Amazon AI indirectly reports clustering. Separation between the accelerators of Top of Search and Product Pages correlates with quality of the keyword clustering. Conversely, a weak separation means that the campaign keyword clusters should be changed.

4. Target Harvesting

Amazon sellers have been successfully harvesting keywords for years. Automated keyword harvesting describes an algorithm that automatically processes the typical keyword funnel beginning with (1) an Auto campaign, then a (2) Manual one, and then within manual, from (3) Broad to Exact. While this may be valuable in general, there is a high price to be paid for the feature when it comes as an ancillary feature of a suboptimal PPC bidding application. For most Amazon sellers, target keywords are not that difficult to develop and often the manual process produces more valuable results at far less cost. Target harvesting is expensive if it means forfeiting the use of Amazon AI.


When looking at automating PPC bidding on Amazon, the PPC AI vendors’ offerings are extremely weak in comparison to the power of 900%™ which fully leverages Amazon AI.

Under close evaluation, not only are the vendor PPC AI offerings a high price to pay for sub-optimized bidding, the other bundled features add little incremental value. If deemed important, they can be purchased ad hoc from other vendors.

900%™ is the seller’s front end to Amazon’s AI back end.  There is an optimal money-making bid for every campaign placement—900%™ helps you find it.