Choose an Optimization Model for Sponsored Products

When you set up a line item you enter a starting cost per click (CPC) bid. Commerce Max then adjusts that bid in real time, based on how likely a shopper is to take the action you care about:

• If a shopper is highly likely to act, the model may raise your bid to help you win the placement.

• If the chance is low, it may lower your bid to avoid unnecessary spend.

This means you do not need to adjust bids by hand throughout the campaign.

Your optimization model determines what the engine optimizes toward. Your bid and budget settings determine the limits it works within.

Choose the Clicks Optimizer when you want the highest click volume relative to your impressions, which is the best click-through rate (CTR).

• What it optimizes: click-through rate.

• CPC control: full. The bid you enter is what you pay for every click on the line item.

• Learning phase: a few days, the shortest of the three.

The Commerce Max engine decides which pages are most relevant for your ad based on the shopper's current behaviour and their likelihood to click. Your bid stays constant for every opportunity.

Choose the Conversions Optimizer when you want to maximise orders and reduce your cost per order (CPO).

• What it optimizes: cost per order.

• CPC control: partial. Your bid varies slightly based on your ad's likelihood to trigger a purchase.

• Learning phase: 21 days.

The CPC you enter becomes a baseline rather than a fixed price. You can also set a Max CPC Bid Limit, separate from your input bid, which acts as a hard cap on how high the adjusted bid can go.

Choose the Revenue Optimizer when you want to maximise attributed revenue, measured as return on ad spend (ROAS).

• What it optimizes: return on ad spend.

• CPC control: partial. Your bid varies based on your ad's predicted attributed sales.

• Learning phase: 21 days.

As with the Conversions Optimizer, your input CPC is a baseline, you can set a Max CPC Bid Limit as a hard cap, and the average CPC in your reports will differ from the bid you entered.

Adaptive CPC is a newer generation of the standard optimizers. Like them, it adjusts bids according to how likely your ad is to produce the action you chose. It then adds two factors the standard models do not consider:

• Pacing. How quickly the line item is spending against its daily budget.

• Traffic patterns. Observed and predicted shopper traffic on the retailer's site.

Together these allow the system to spend your budget across the full span of shopper activity rather than exhausting it early.

Adaptive CPC also introduces auction throttling: when a line item is pacing ahead of plan, it enters only a subset of the auctions available to it. Standard optimizers only change the bid amount; Adaptive CPC changes both the bid and how often you compete.

Each line item:

• Accelerates spend when it is pacing behind.

• Conserves budget when it is pacing ahead.

• Aims to stay live for at least 16 hours a day.

The optimizer reassesses every few minutes, weighing what the line item should have spent against what it has spent, alongside win rates, bidding history, your Max CPC, and the retailer's floor prices.

The trade-off is control for consistency. Standard optimizers give you more levers; Adaptive CPC takes most of them away and manages delivery for you.

Why You Might Choose It:

• Reaches traffic later in the day, which often converts better or costs less because fewer advertisers are still competing.

• Reduces bids automatically on line items pacing too fast, protecting a limited budget.

• Raises budget utilisation on line items that struggle to spend.

• Cuts manual work, because you set a Max CPC and nothing else.

• Suits volatile or high-volume traffic, where a fixed bid stops being competitive quickly.

• Suits long-tail activity, where you have too many line items to manage bids individually.

When To Stay on a Standard Optimizer Instead:

• You need granular control through product bid overrides, keyword bid overrides, or page type bid multipliers.

• You are working to a strict ROAS target. Adaptive CPC optimizes delivery as well as performance, so it may spend more to stay live.

• You are running multi-retailer line items.

• Your budget is already spread thinly across many similar line items. Adaptive CPC does not move budget between line items, so it will not correct over-allocation.

You can switch back to a standard optimizer at any time by editing the line item.

Setting Your Max CPC

On an Adaptive CPC line item, the Max CPC is the only bid control you have, so where you set it matters more than on a standard line item.

• Use the top of the suggested bid range, where one is shown. This allows the engine to bid competitively, while keeping your ceiling based on current auction data.

• If the line item is new and no useful range is available, start around 25% to 30% above your historical average CPC, while staying above the retailer's minimum bid.

We recommend starting with a modest daily budget and increasing it as pacing improves. Adaptive CPC works from your daily budget, so a large budget set before you know how the line item paces makes over-allocation harder to identify. To learn more, please see Optimize a Live Sponsored Products Campaign for how to diagnose a line item that is not spending.

Every model has a learning phase. It begins when you select the model and the line item goes live.

The Conversions and Revenue Optimizers need the full three weeks to pass their learning phase. Judging either model's performance before then tells you about an optimizer that has not finished learning and conclusions will not be fully reliable.

Two factors shorten the learning phase and one restarts it:

• Higher volume shortens learning. The more impressions and conversions the line item generates, the faster the optimizer reaches full efficiency.

• Leaving settings alone shortens learning. We recommend not changing bid settings during the learning phase.

• Switching models restarts it. Moving from Conversions to Revenue restarts learning, and moving back later restarts it again.

A new SKU can take less time to learn than the line item did, because the model is learning click and sales patterns for that product against settings it already understands.

With the Conversions and Revenue Optimizers, we recommend applying the same CPC bid across all your products and line items. This lets the engine optimize across your whole budget rather than within individual bids, and saves you the work of managing each bid separately.

There are two reasons to break that rule:

• Your minimum bids vary. Applying one CPC everywhere means setting it above the highest minimum bid across your line items. If campaign ROAS then falls below target, lower the CPC on the line items or SKUs where the minimum bid allows it.

• Your campaign has mixed objectives. If you are pushing visibility on a specific SKU, or concentrating spend on particular inventory, adjust CPC bids to control volume where it matters.

With the Clicks Optimizer, adjust bids at line item or SKU level to control click volume directly.

With Adaptive CPC, you set only the Max CPC. Bid overrides and multipliers are unavailable.