Amazon’s New Sponsored Products Bid Automation Is Reshaping PPC Economics for Mid-Market Sellers

Amazon Ads pushed a significant update to its Sponsored Products bid automation engine on June 24, 2026, and the full consequences are only now becoming visible to the sellers and agencies managing the bulk of mid-market ad spend on the platform. The update — which Amazon internally describes as “Dynamic Bidding 2.0” in its Ads API documentation — introduces placement-level machine learning signals that override manual bid adjustments more aggressively than earlier versions. For sellers accustomed to tight manual control, the shift is proving disorienting. For those already leaning on automation, some early results are striking.

The change lands at a complicated moment. Amazon advertising revenue crossed $62 billion annually in Q1 2026, per the company’s earnings disclosure, and competition for top-of-search placement in high-volume categories has pushed average CPCs in home goods, supplements, and apparel to multi-year highs. Sellers who built profitable PPC architectures on manual bidding discipline are now watching Amazon’s own algorithm override their caps with increasing frequency.

What exactly changed in Amazon’s Dynamic Bidding 2.0 update?

The core change is the expansion of real-time contextual signals feeding Amazon’s bid adjustment engine. Previously, Dynamic Bidding (Down Only, Up and Down) operated primarily on conversion probability derived from historical campaign data and basic session signals. The June 2026 update incorporates basket composition data, cross-category browse history, and what Amazon describes as “purchase intent velocity” — an aggregate signal derived from how quickly a shopper is moving through a category versus browsing passively.

In practice, this means Amazon’s system is now raising bids on impressions it judges to be high-conversion opportunities, even when sellers have set conservative manual bids or bid modifiers. Agency operators report seeing effective CPCs run 30–55% above their stated maximum bids in certain placements, a behavior that was previously rare enough to be attributed to data anomalies.

“We had campaigns where we’d dialed manual bids down to $0.90 on broad match to control spend, and the system was clearing $1.40 on top-of-search. That’s not a rounding error — that’s Amazon deciding it knows better than our bid logic. Sometimes it’s right. The ROAS held. But it completely breaks our forecasting model.” — Liran Hirschkorn, founder of Incrementum Digital, speaking at a Prosper Show follow-up webinar on July 8, 2026

💡 Article Summary

Key Insights

1

What exactly changed in Amazon’s Dynamic Bidding 2.0 update?

2

Which seller segments are most exposed to the bid override behavior?

3

How are third-party PPC tools responding to the algorithm change?

4

Are there category-specific patterns sellers should know about?

5

What tactical adjustments are working right now for affected sellers?

Source: Ecommerce Times

Amazon’s Ads documentation update, published July 2, acknowledges that the system “may bid above your entered CPC when our models predict a significantly higher conversion probability,” but stops short of specifying the ceiling multiplier. Sellers in the Seller Central forums have reported bid overrides of up to 2.1x on Product Detail Page placements during peak traffic windows.

Which seller segments are most exposed to the bid override behavior?

The impact is not evenly distributed. Based on reporting from seven independent agency operators and three in-house brand teams who spoke with Ecommerce Times between July 3–10, the sellers most disrupted fall into two groups:

• Manual bidding purists — Brands that built profitability models on granular bid control, often in competitive categories like kitchen, pet, or personal care, where CPCs were already high and ACOS targets were tight (sub-20%).
• Portfolio-scale operators — Sellers running 200+ SKUs with rule-based automation from tools like Perpetua, Pacvue, or Skai (formerly Kenshoo), where bid logic is scripted and the new override behavior creates unpredictable spend variance at scale.

Sellers already using Amazon’s own “Maximize Conversions” automated bidding strategy report less disruption — largely because they’d already ceded bid control to the algorithm. For them, Dynamic Bidding 2.0 represents a refinement rather than a shock.

“If you were already running Rule-Based Bidding through Pacvue with hard spend caps, you have a buffer. The sellers getting hurt are the ones who thought manual bids plus a 20% top-of-search modifier was a fortress. It’s not anymore.” — Elizabeth Greene, founder of Junglr, in a July 9 LinkedIn post that drew over 400 comments from Amazon sellers.

How are third-party PPC tools responding to the algorithm change?

The major Amazon advertising platforms are moving quickly to adapt, though their responses vary by architecture. Perpetua pushed a patch on July 7 that increases the frequency of its bid harvesting cycles from every 6 hours to every 2 hours for accounts flagged as experiencing high bid variance, allowing its system to recalibrate faster against Amazon’s real-time overrides. Pacvue announced on July 9 that it is releasing a new “Bid Variance Alert” feature — currently in beta for enterprise accounts — that flags campaigns where Amazon’s effective CPC deviated more than 25% from the set bid over a rolling 48-hour window.

Helium 10’s Adtomic team has been quieter publicly, but three agency sources confirmed that Adtomic is testing a “defensive bid floor” feature intended to give sellers a mechanism to signal to Amazon’s API that bids should not be exceeded regardless of intent signals. Whether Amazon’s API will honor such signals under the new bidding model remains untested at scale.

• Perpetua: Faster bid harvest cycles (2hr vs 6hr), available now for Pro and above tiers
• Pacvue: Bid Variance Alert in enterprise beta as of July 9
• Skai: No public statement as of press time; agency sources report internal testing of override detection logic
• Helium 10 Adtomic: Defensive bid floor feature in private testing
• Amazon’s own console: No UI-level controls to limit override multiplier have been announced

Are there category-specific patterns sellers should know about?

Data shared by two Pacvue agency partners — covering a combined $8.2 million in Amazon ad spend in June 2026 — points to heavy override activity in specific categories. Home & Kitchen saw the highest frequency of bids exceeding stated CPCs by 40% or more, followed by Beauty & Personal Care and Sports & Outdoors. Electronics and Grocery categories showed the lowest deviation, which agency analysts attribute to Amazon’s conversion models being more conservative in high-return or perishable verticals.

Apparel sellers face a distinct version of the problem: because Amazon’s intent signals are built heavily on conversion history and apparel has structurally lower conversion rates (high browse, high return), the algorithm appears to be bidding more aggressively on apparel shoppers who have shown recent purchase behavior in adjacent categories — a pattern that inflates CPCs without proportional ROAS gains for sellers whose catalog doesn’t match the cross-category intent signal.

“We manage a $120K monthly apparel account and our blended ACOS jumped from 18% to 26% in the first two weeks of July. When we audited the impression data, we were paying top-of-search rates on shoppers who’d just bought shoes from a competitor. Amazon thinks that’s a buying signal. For a dress brand, it really isn’t.” — Mina Elias, founder of Trivium Group, in an interview with Ecommerce Times on July 10, 2026.

What tactical adjustments are working right now for affected sellers?

Agency operators who’ve been stress-testing responses over the past two weeks converge on a short list of defensive moves. None of them fully neutralize Dynamic Bidding 2.0’s override behavior, but collectively they reduce exposure:

• Switch high-CPC campaigns to “Down Only” bidding: While not immune to overrides, Down Only campaigns have shown 15–20% lower override frequency than Up-and-Down campaigns in early July data from Incrementum Digital’s account portfolio.
• Set portfolio-level daily budget caps with 10% buffer: Overspend from bid overrides can compound quickly in high-volume campaigns. Portfolio budgets give a harder ceiling than campaign-level caps under the new system.
• Segment by match type more aggressively: Broad match campaigns are experiencing the highest override rates. Isolating exact and phrase match into separate campaigns gives cleaner bid control and clearer attribution data.
• Reduce reliance on top-of-search bid modifiers above 50%: High modifiers are amplifying the override behavior — Amazon’s system layers its own multiplier on top of the seller’s modifier, compounding the effective CPC spike.
• Monitor Search Term Reports daily during transition: New intent signals mean previously low-converting terms may be eating disproportionate spend. Daily SQR audits are catching bleed faster than weekly reviews.

What should sellers expect from Amazon Ads in the second half of 2026?

The direction of travel is clear: Amazon is consolidating more bid decision-making authority into its own ML stack, and the gap between what sellers set and what Amazon spends is going to widen before it narrows. The platform’s incentive structure — Amazon earns revenue on every click regardless of conversion — creates an inherent tension with seller ROAS targets that Dynamic Bidding 2.0 makes more visible, not less.

Amazon Ads has signaled at its unBoxed conference roadmap (held April 2026 in Austin) that a unified “Performance+” campaign type is coming in Q4 2026, which would consolidate Sponsored Products, Sponsored Brands, and Sponsored Display into a single AI-managed campaign structure similar to Google’s Performance Max. If that timeline holds, the current Dynamic Bidding disruption may be a preview of a much larger structural shift in how Amazon allocates ad inventory.

For mid-market sellers running $50K–$500K in monthly Amazon ad spend, the next 90 days are a stress test for both their PPC infrastructure and their software stack. The sellers who emerge cleanest will likely be those who tighten match type segmentation now, pressure their tool vendors for override visibility features, and start modeling what “ceding bid control” scenarios actually look like for their margin structure — before Amazon makes that decision for them.