Retail Media Is Expanding Into Commerce Media: How Should Advertisers Decide Where to Invest?

 

From Commerce Signals and Transaction Proximity to Incrementality, Measurement and Business Value

Retail media has grown quickly over the last few years. For advertisers, the attraction is easy to understand: retailers sit close to the transaction and have access to signals that many other media environments simply do not have.

But the landscape is becoming much broader.

Retailers are building advertising businesses. Marketplaces have their own media ecosystems. Delivery platforms can connect advertising with transaction behaviour. Travel businesses have booking and destination data. Commerce audiences can increasingly be activated beyond the original platform across other media environments.

For a performance marketer, that creates more possibilities, but also a very practical problem:

Where should the next advertising euro actually go?

That decision cannot come from reach or ROAS alone. I want to understand the signal behind the audience, where that consumer sits in the buying journey, how close the media is to a transaction, what the partner adds to my existing media mix, how sales are measured and, ultimately, whether the investment created additional business.

That is the lens I want to use throughout this article.

First, What Do We Mean by Retail Media and Commerce Media?

The easiest way to understand retail media is through a shopping example.

Imagine someone visiting an online retailer and searching for running shoes.

Several brands sell running shoes on that retailer. One of them pays for its product to appear prominently within the search results.

That is a classic retail media placement.

The retailer has something extremely useful for advertising: it understands what people search for, which products they view, what they add to their basket and, importantly, what they eventually buy.

Retail media can include much more than sponsored products:

Sponsored Search | Product Listings | Onsite Display | Retailer Audiences | Offsite Advertising | Video | CTV

depending on the retailer and its advertising capabilities.

So what changes with commerce media?

The same principle can extend beyond traditional retailers.

1.     A marketplace sees transactions.

2.     A food-delivery platform sees ordering behaviour.

3.     A travel platform sees destination searches and bookings.

4.     A loyalty ecosystem can understand repeat purchasing.

5.     A commerce platform may have detailed product and transaction information.

These businesses can use those signals to create advertising opportunities.

So I think about the progression roughly like this:

Retail Environment → Commerce Data → Advertising Audience → Media Activation → Transaction Measurement

Commerce media broadens the number of businesses and environments capable of doing this.

That means advertisers now have a growing number of partners competing for the same media budget.

And that is where the investment decision becomes interesting.

I Would Start With the Business Problem

Imagine a German fashion retailer has €500,000 available for incremental media investment.

There are several things the business might want from that money:

1.     Acquire new customers

2.     Launch a footwear collection

3.     Increase sales in an underperforming category

4.     Support Black Friday

5.     Grow repeat purchases

6.     Enter another European market

The media strategy should look different depending on which problem we are solving.

For a new product launch, I may need meaningful category reach.

For customer acquisition, I care about identifying and converting genuinely new customers.

For an established product with strong organic demand, incrementality becomes particularly important because attributed sales could include customers who would have purchased anyway.

So before asking:

Which commerce media network should I use?

I would ask:

What exactly am I trying to change in the business?

That becomes the anchor for everything that follows.

Where Does the Partner Sit in the Customer Journey?

Customers rarely experience advertising through one neat sequence.

Someone buying a pair of €180 running shoes might see a video, encounter the brand on social, search for reviews, compare models, visit a retailer, leave, search again and eventually purchase.

Different media environments participate at different points.

Discovery → Research → Comparison → Product Evaluation → Purchase → Repeat Purchase

Commerce media can potentially operate across several of these stages.

A sponsored product may appear very close to purchase.

An offsite video campaign using commerce audiences may reach the same consumer much earlier.

That distinction matters because I would not evaluate both placements using exactly the same expectations.

A lower-funnel placement might be judged heavily on conversion and incremental sales.

An earlier-stage campaign might have a broader job within the media plan.

So one of my first questions for any commerce media partner is:

What role can this environment realistically play in the customer journey?

What Does the Commerce Signal Actually Tell Me?

This is where I would spend a lot of time.

Imagine three media partners all offer an audience called:

“Running Enthusiasts.”

That label tells me very little.

a)    Partner A may define the audience using people who read running content.

b)    Partner B may include people who recently browsed running shoes.

c)     Partner C may know that someone purchased running shoes twice during the last twelve months and recently started browsing another pair.

These are very different signals.

The same applies across commerce media.

I want to understand the actual behaviour underneath the audience:

Search History → Browsing Behaviour → Category Interaction → Product Views → Basket Activity → Purchase History → Purchase Frequency → Transaction Value → Loyalty Behaviour

Recency matters too.

Someone who bought a washing machine yesterday probably has very different immediate value to an appliance advertiser than someone currently researching one.

For me, audience evaluation therefore starts with a simple question:

What does this partner genuinely know about the consumer that is useful for this campaign?

Transaction Proximity Matters, but So Does Influence

Commerce media gives us another useful dimension: how close the consumer is to buying.

Think about the journey:

Browsing → Category Search → Product View → Comparison → Basket → Purchase

A consumer with a product already in their basket is obviously close to a transaction.

That makes the media opportunity valuable.

But there is another question I would ask:

How much opportunity does the advertising still have to influence what happens?

Suppose someone buys the same coffee brand every month.

They visit the retailer, search for that exact brand, see a sponsored placement and purchase it again.

The ad was certainly close to the transaction.

But proximity alone does not tell me whether the advertising created additional value.

Now imagine someone browsing several competing coffee brands and categories without a fixed preference.

The advertiser may have greater opportunity to influence that decision.

So I would look at:

Transaction Proximity + Opportunity to Influence

rather than treating proximity as a standalone measure of media quality.

Scale Needs Context

Reach still matters.

A brilliant audience signal with almost no usable scale is unlikely to transform a large advertiser's business.

But I also would not choose a partner because it offers the biggest number on a media plan.

Imagine:

 

 

 

 

 

	Partner A	Partner B
Addressable Audience	25M	8M
Category Relevance	Medium	High
Purchase Signal	Limited	Strong
Transaction Proximity	Medium	High
Transaction Measurement	Partial	Strong

 

25  million people sounds impressive.

Eight million may represent the more commercially useful opportunity.

Or the opposite could be true if the business objective requires substantial incremental reach.

The useful comparison is therefore closer to:

Addressable Scale × Relevance × Intent × Transaction Proximity × Measurement

And even that needs to be interpreted against the campaign objective.

There is no universal number that tells me which partner deserves the budget.

Onsite and Offsite Create Different Opportunities

Commerce data becomes even more interesting when it can be activated outside the original shopping environment.

Onsite

The consumer is already inside the retailer or commerce platform.

Advertising can appear around:

Search → Category Browsing → Product Pages → Recommendations → Checkout Journey

The proximity to shopping behaviour can be extremely useful.

Offsite

The same commerce signals may potentially be activated across environments such as:

Display | Online Video | CTV | Social | Other Addressable Media

where the partner supports those capabilities.

Now I can potentially combine:

Commerce Signal + Broader Media Reach

That is useful, but it introduces another planning question.

If I already reach these consumers through search, social, programmatic advertising or CTV, how much additional value does commerce-powered offsite activation bring?

I would want to understand:

Incremental Reach → Audience Overlap → Frequency → Media Cost → Downstream Performance

Otherwise I could end up paying several platforms to reach substantially the same people.

ROAS Is Useful. I Still Want to Know What Was Incremental.

Suppose a commerce media campaign reports:

€100,000 Spend → €600,000 Attributed Revenue → 6x ROAS

I absolutely want to know that.

But I would not stop there.

Imagine €350,000 of those purchases came from loyal customers who buy the same products regularly.

Some may have purchased regardless of the campaign.

The question then becomes:

How much additional business did the €100,000 actually create?

That is why I separate two concepts.

Attribution

Which sales were associated with advertising according to the measurement rules?

Incrementality

What happened because of the advertising compared with what would likely have happened without it?

Those questions can produce very different views of the same campaign.

A campaign can have strong attributed ROAS and modest incremental impact.

Another campaign may look less spectacular through last-touch reporting but generate meaningful new demand.

For budget allocation, I want both perspectives.

How Would I Measure Incrementality?

There is no single experiment that fits every advertiser or commerce media partner.

Depending on scale, geography, platform capability and available data, I might consider:

Randomized Holdouts | Geo Experiments | Matched Markets | Model-Based Counterfactuals | Econometric Analysis

The methodology matters.

If a partner tells me a campaign generated €2 million in incremental sales, I want to understand how the counterfactual was constructed.

a.     What happened in the control population?

b.     Were test and control groups comparable?

c.     How long did the test run?

d.     Were other campaigns running simultaneously?

e.     Could customers move between test and control environments?

f.      Was the analysis based on online sales only or total transactions?

Incrementality is useful because it helps answer a difficult business question.

The quality of the answer depends heavily on the quality of the test.

Closed-Loop Measurement Gives Commerce Media an Important Advantage

One reason commerce media attracts so much advertising investment is its potential connection between advertising and transactions.

The loop can look like:

Ad Exposure → Product Interaction → Purchase

For a retailer with both online and physical stores, that could potentially become even more valuable if the measurement environment can connect media exposure with transactions across both.

But I would still inspect the methodology.

I want to understand:

Attribution Window | Online vs Offline Coverage | New vs Existing Customers | Cross-Device Matching | Product-Level Measurement | Returns and Cancellations | Cross-Channel Exposure | Incrementality Methodology

Two platforms can both report “ROAS” while measuring it differently.

Before comparing them, I need to know what sits underneath the number.

Eventually, I Need to Reach the Economics

Media metrics help me operate campaigns.

Business economics help me decide how much those campaigns deserve.

Consider this hypothetical example:

€100,000 Media Spend → €600,000 Attributed Revenue → €250,000 Estimated Incremental Revenue → €100,000 Incremental Gross Margin

Suddenly the 6x attributed ROAS is only one part of the investment story.

Then I may need to consider:

Customer Acquisition Cost | New Customer Rate | Average Order Value | Gross Margin | Promotional Discounts | Repeat Purchase Rate | Customer Lifetime Value

For a performance marketer, the analysis gradually moves:

CPM / CPC → Conversion Rate → Attributed Revenue → Incremental Revenue → Margin → Customer Economics → Incremental Business Value

That is where media planning becomes much more closely connected to commercial decision-making.

My Commerce Media Investment Framework

If I were comparing several commerce media partners, I would put them through the same framework.

Dimension	What I Want to Understand
Business Objective	What outcome are we trying to change?
Customer Journey	Where does this environment sit in the decision process?
Commerce Signal	What behaviour does the partner genuinely observe?
Transaction Proximity	How close is the consumer to purchasing?
Addressable Scale	Is there enough relevant reach to matter?
Activation	Where can the signal actually be used?
Media Overlap	What does it add to channels already in the plan?
Measurement	Can exposure be connected reliably with transactions?
Incrementality	What additional business did the media create?
Economics	Does that incremental value justify the investment?
Scalability	What happens when I increase the budget?

 

The value is in using the dimensions together.

Strong signals with limited scale may still have an important role.

Large scale with weaker transaction visibility may solve a different problem.

Excellent attributed ROAS with little incremental impact changes how I would value the investment.

Each strength needs context.

A Hypothetical German Fashion Retailer

Let's make the framework practical.

A German multi-brand fashion retailer wants to increase new-customer footwear revenue.

It is evaluating three hypothetical commerce media opportunities.

Partner A: Large Marketplace

Large addressable audience, significant product-search activity and substantial scale.

Potential role:

Discovery → Product Comparison → Purchase

Partner B: Fashion Commerce Platform

Smaller audience but deeper fashion-specific browsing and purchasing behaviour.

Potential role:

Category Intent → Product Consideration → Acquisition

Partner C: Broader Transaction Network

Less fashion-specific but strong transaction signals and broader offsite activation.

Potential role:

Commerce Audience → Offsite Reach → Customer Acquisition

I would not immediately rank A, B and C.

First, I would test the job each partner could perform.

Then I would evaluate:

New Customers → Incremental Revenue → CAC → Margin → Audience Overlap → Marginal Performance as Spend Increases

The outcome may be that all three deserve investment for different reasons.

It may also be that one adds very little to what the retailer already gets elsewhere.

That is exactly what the testing needs to establish.

Building the Portfolio

A large advertiser could eventually have:

Search + Social + Programmatic Advertising + CTV + Retail Media + Marketplace Media + Delivery Media + Other Commerce Media

That creates substantial opportunity.

It also creates complexity.

More partners can mean:

More Signals | More Inventory | More Measurement

but also:

More Audience Duplication | More Attribution Claims | More Platforms | More Reporting | More Operational Work

So I would map every commerce media investment against its role.

       i.         Which partner helps me discover new customers?

     ii.         Which one identifies strong category intent?

    iii.         Which sits closest to the transaction?

    iv.         Which gives me useful offsite activation?

     v.         Which provides the strongest measurement?

    vi.         Which produces incremental growth?

The portfolio should have a reason for being a portfolio.

What Happens When I Increase the Budget?

This is one of the questions I care about most.

A network may perform extremely well at €20,000 per month.

That does not tell me what happens at:

€50,000 → €100,000 → €250,000

As spend increases, I may begin reaching less relevant audiences, the same consumers more frequently, more expensive inventory, or demand that becomes increasingly difficult to influence.

So I would track marginal performance, not only blended performance.

If the first €50,000 generates excellent incremental economics but the next €50,000 produces substantially weaker returns, that should affect allocation.

The practical budget question is therefore:

What does the next euro produce?

So Where Should the Next Advertising Euro Go?

As retail media expands into a broader commerce media landscape, advertisers will have access to more networks, more transaction signals and more ways to activate them.

My investment process would remain fairly disciplined:

Define the Business Objective

↓

Understand the Customer Journey

↓

Interrogate the Commerce Signal

↓

Assess Transaction Proximity

↓

Evaluate Relevant Scale

↓

Understand Onsite and Offsite Activation

↓

Measure Transactions

↓

Test Incrementality

↓

Connect Results With Commercial Economics

↓

Evaluate Marginal Returns

Only then would I decide how much more budget a partner deserves.

For me, that is the real opportunity in commerce media.

It gives performance marketers another powerful set of signals and media environments to work with, while making the investment decision increasingly dependent on something much closer to the business:

What additional value did this advertising create, and what is the next advertising euro likely to produce?

 

Closing Thoughts

Retail media is becoming a much bigger planning conversation than sponsored products and placements inside retailer websites.

As commerce media expands, advertisers can potentially work with richer shopping and transaction signals across retailers, marketplaces, delivery platforms, travel businesses and other commerce environments. Those signals can also travel further through offsite activation, creating more ways to connect commerce data with the wider paid media mix.

That also makes budget allocation harder.

I would want to know what sits behind an audience, how relevant that signal is to the business objective, where the customer is in the buying journey, how close the media sits to the transaction, what reach is genuinely incremental, and how reliably advertising exposure can be connected with business outcomes.

And I would keep coming back to incrementality.

A sale attributed to advertising is valuable information. Understanding whether advertising actually changed the outcome gives me a different level of information for investment decisions.

The same applies to scale. Strong performance at one level of spend does not automatically tell me what happens with the next €50,000 or €100,000. Marginal returns matter when deciding where additional budget should go.

Conclusion

The expansion from retail media into commerce media gives performance marketers more signals, more inventory, more activation possibilities and potentially much stronger connections between media and transactions.

For me, the useful way to evaluate that opportunity is:

Business Objective → Customer Journey → Commerce Signal → Transaction Proximity → Addressable Scale → Activation → Measurement → Incrementality → Economics → Marginal Returns

The final decision is then much simpler to frame:

What did this media investment add to the business, and what is the next advertising euro likely to produce?

That is the question I would want commerce media planning to answer.

 

Beyond the Click: What ChatGPT Sponsored Agents Could Change in Paid Media

 

From Ad Exposure to Conversation | From Customer Intent to Consideration | From Engagement to Business Outcome

ChatGPT Ads has already introduced a different context for paid media: reaching people while they are exploring, comparing and making decisions through conversation.

Sponsored Agents add another layer to that journey.

After seeing an eligible ad, a user can choose to open a clearly labelled conversation with an AI representative sponsored by that business. They can ask follow-up questions about the product or service and, when ready, continue to the advertiser's website.

That creates an interesting paid media journey:

Ad Exposure → Sponsored Conversation → Product Exploration → Consideration → Website → Business Outcome

Sponsored Agents are currently being tested with selected advertisers in the United States. So this is not a hands-on European performance case study. My interest is in what this format could mean for paid media strategy, creative, measurement and optimization if it develops more broadly.

1. What Exactly Is a Sponsored Agent?

Imagine someone researching running shoes in ChatGPT.

They see a relevant sponsored ad for a particular model. Instead of immediately leaving ChatGPT, they can choose to continue with the advertiser's Sponsored Agent.

The conversation could develop around questions such as:

“Is this suitable for overpronation?”

→ “How does the sizing compare with my current shoes?”

→ “Would this work for 20–30 km of running each week?”

→ “Which model would suit me better?”

→ Visit the advertiser's website

The Sponsored Agent conversation is clearly identified as sponsored. It is also separate from ChatGPT's independent answers and separate from the original conversation the user was having.

From a paid media perspective, that separation is important. The Sponsored Agent is an advertiser experience entered from the ad, rather than the advertiser taking over the original ChatGPT conversation.

2. A New Layer in the Customer Journey

Consider the journey of a typical paid media click:

Ad → Website → Product Exploration → Consideration → Conversion

With Sponsored Agents, part of that exploration could potentially happen before the website visit:

Ad → Sponsored Agent → Questions → Product Exploration → Website → Conversion

That middle layer is what interests me.

Take a hypothetical furniture retailer advertising a dining table.

A potential customer might want to know:

Will a 180 cm table fit comfortably in my room?

Can six people sit around it?

How should the surface be maintained?

Does it come in a lighter finish?

Those questions represent consideration.

For paid media planning, I would therefore think about Sponsored Agents as another stage in the journey rather than simply another creative format.

3. From Click Intent to Conversational Intent

Clicks give us one signal:

The user chose to engage with the ad.

A conversation can potentially contain much richer expressions of intent.

Imagine a hypothetical electronics advertiser.

User A:
“Tell me more about this laptop.”

User B:
“Does this support two external 4K monitors, and is 16 GB RAM enough for Premiere Pro?”

Both users engaged with the advertising experience, but the nature of that engagement is very different.

This opens an important paid media question:

Which conversational interactions are meaningful indicators of consideration?

I would want to understand signals such as:

Sponsored Agent Starts → Follow-Up Interactions → Product Exploration → Website Visits → Conversion Events → Revenue

Exactly which conversational signals advertisers will ultimately receive for reporting and optimization will depend on how the product develops. That measurement boundary is important because private user conversations remain protected from advertisers.

4. Creative Strategy Could Change Too

A standard ad has limited space.

The headline, description and image need to create enough relevance and interest for someone to take the next step.

Sponsored Agents potentially give the advertising experience somewhere to continue.

For a hypothetical travel advertiser, the ad could introduce a proposition around a Mediterranean holiday.

The Sponsored Agent could then help the user explore:

Destination → Travel Dates → Hotel Type → Budget → Activities → Available Options

That changes how I would think about creative architecture.

The initial ad still needs a strong proposition. But it can also serve as the entrance to a deeper product or service experience.

This creates another optimization layer:

Which ad propositions generate clicks?

and potentially:

Which ad propositions generate meaningful sponsored conversations that eventually contribute to business outcomes?

Those are different questions.

5. The Website Visitor Could Arrive With More Context

Sponsored conversations could also affect what happens after the user leaves ChatGPT.

Take a hypothetical fashion retailer.

A user might discuss:

Occasion: September wedding
Budget: €300
Style: Smart-casual
Preference: Nothing too formal
Product: Outfit and shoes

If that conversation helps the user narrow the available options before visiting the retailer, the website visit occurs at a different point in the consideration journey.

For performance marketing, I would want to examine whether that changes:

Landing Page Engagement → Product Views → Add-to-Basket Rate → Conversion Rate → Average Order Value → Customer Acquisition Cost

I would also want to understand where the Sponsored Agent contributes value and where the website remains better suited to the task.

Product discovery, detailed comparison, inventory, checkout and transaction completion do not necessarily need to happen in the same environment.

6. Measurement Becomes the Critical Question

This is where I would spend considerable attention before scaling investment.

The ideal measurement framework would help me understand the complete journey:

Stage	What I would want to understand
Ad Exposure	Reach, impressions, relevance
Ad Engagement	Clicks and engagement
Sponsored Agent	Meaningful interaction signals made available to advertisers
Website Visit	Qualified traffic and landing behaviour
Consideration	Product views, key actions, add to basket
Conversion	Leads, purchases or subscriptions
Business Outcome	Revenue, CAC, customer quality and incremental value

 

ChatGPT Ads already has conversion measurement infrastructure through browser and server-side tracking. Sponsored Agents introduce another potential interaction layer between the ad and the advertiser's website.

For me, the important question will be how clearly that layer can eventually be connected to downstream business outcomes.

A high number of Sponsored Agent conversations would be interesting.

A high number of commercially valuable conversations that contribute to profitable customer acquisition would be much more meaningful.

7. Optimization Could Gain Another Signal

Suppose two hypothetical campaigns each generate 10,000 ad engagements.

Campaign A

10,000 engagements
→ 1,500 Sponsored Agent interactions
→ 600 website visits
→ 60 purchases

Campaign B

10,000 engagements
→ 3,000 Sponsored Agent interactions
→ 1,200 website visits
→ 150 purchases

These are deliberately hypothetical numbers.

The example illustrates the optimization question I would eventually want to answer:

Can the quality of conversational engagement help us understand and optimize downstream performance?

If the platform eventually provides suitable aggregated signals, advertisers could potentially evaluate more than impression and click behaviour.

The optimization chain could become:

Media → Ad → Conversation → Site Behaviour → Conversion → Business Value

That would make Sponsored Agents particularly interesting for products where customers naturally have several questions before making a decision.

8. Where I See the Strongest Paid Media Use Cases

I would initially look at categories with meaningful consideration journeys.

Business / Vertical	Potential Sponsored Agent Role
Fashion & eCommerce	Sizing, fit, occasion, product comparison
Consumer Electronics	Specifications, compatibility, use cases
Travel	Destination, accommodation, budget and itinerary questions
Home & Furniture	Dimensions, materials, suitability and care
Beauty	Product characteristics, routines and comparisons within appropriate advertising boundaries
Automotive	Features, configurations and product research
Subscription Services	Plans, features and suitability
Technology	Product capabilities, integrations and requirements

 

The common characteristic is a customer who benefits from additional information before taking the next step.

That is where conversational advertising becomes particularly interesting to me.

9. The Advertiser Workflow Is Becoming Conversational Too

There is another development happening on the opposite side of the advertising experience.

Advertisers can now use natural-language instructions through ChatGPT Work to work with Ads Manager.

That can support activities such as:

Website / Brief

↓

Campaign Creation

↓

Campaign Updates

↓

Performance Analysis

↓

Recommendations and Next Actions

Ads Manager is also adding AI assistance for creative development, including suggested copy and imagery based on the landing page and campaign objective. Advertisers can review and edit those suggestions before using them.

AI-powered text customization can go further by adapting existing headlines and descriptions to conversational context and translating copy into the user's preferred language when enabled.

So there are potentially two conversational layers developing around the same advertising ecosystem.

Consumer side

User → Ad → Sponsored Agent → Business

Advertiser side

Marketer → Natural-Language Instruction → Advertising Platform → Campaign

For paid media operations, that combination could become significant.

10. CRM and Commerce Integrations Add Another Layer

The connection with existing marketing systems is equally interesting.

CRM integration can bring advertising closer to:

Campaign → Lead → CRM → Follow-Up → Customer

Commerce integration can bring advertising closer to:

Product Catalog → Campaign → Product Discovery → Website → Purchase

For example, ChatGPT Ads can now connect with CRM workflows that support campaign creation, lead capture, follow-up and performance measurement.

Commerce integrations can connect advertising workflows more closely with merchant product data and campaign management.

For me, this matters because paid media rarely operates as an isolated platform.

Its value ultimately needs to connect with the systems that contain:

Products → Leads → Customers → Transactions → Revenue

11. What I Would Test First

If Sponsored Agents become broadly available, I would begin with a controlled test built around a product where customer questions are already an important part of the buying journey.

For example:

Hypothetical German Consumer Electronics Retailer

Product: Premium laptop range

Step 1: Identify common pre-purchase considerations.

Performance → Battery → Compatibility → Display → Portability → Price

Step 2: Build advertising around commercially relevant contexts.

Step 3: Use the Sponsored Agent to help users explore those considerations.

Step 4: Connect website conversion measurement.

Step 5: Evaluate the complete journey.

Ad Exposure → Sponsored Conversation → Website → Product Behaviour → Purchase → Revenue

I would keep the initial investment controlled until enough data exists to understand the quality of the resulting traffic and customers.

12. The Questions I Would Want Answered

Sponsored Agents are still in limited testing, so several questions matter before making broader performance conclusions.

Measurement: Which Sponsored Agent interactions will advertisers be able to measure?

Optimization: Can meaningful conversational engagement eventually become an optimization signal?

Reporting: How much aggregated insight will advertisers receive about sponsored interactions while preserving user privacy?

Creative: How much control will advertisers have over the knowledge, tone and behaviour of their Sponsored Agent?

Product Data: How dynamically can product availability, attributes and other relevant information be incorporated?

Attribution: How will Sponsored Agent engagement appear within the wider conversion journey?

Incrementality: Are Sponsored Agents creating additional conversions, assisting existing demand, or doing both?

Scale: How does performance behave as advertiser adoption and available inventory increase?

Those questions require testing and evidence.

13. What This Could Mean for Europe

There is an important geographic distinction today.

ChatGPT Ads is already available across European markets, while Sponsored Agents are currently being tested with selected advertisers in the United States.

So I would not treat Sponsored Agents as an existing European campaign capability.

But their development is highly relevant to European paid media because it gives us an early view of where the ChatGPT advertising experience could evolve.

If the format eventually expands into Europe, the regional environment will also matter.

Privacy requirements, personalization capabilities, measurement, audience availability and market-specific product features will all influence how I would build the strategy.

The U.S. test gives us the concept.

A future European implementation would need to be evaluated on its own capabilities and constraints.

14. My Paid Media View

Sponsored Agents introduce an interesting additional layer between advertising exposure and the advertiser's website.

For products with meaningful consideration journeys, that layer could allow consumers to explore questions that traditionally happen across landing pages, product pages, comparison tools, FAQs, reviews or sales interactions.

From a paid media perspective, I would evaluate the opportunity through four connected areas:

Media Strategy

Where does the sponsored conversation belong in the customer journey?

Creative Strategy

Which proposition should start the conversation, and what should the Sponsored Agent help the customer explore?

Measurement

Can interaction with the Sponsored Agent be connected meaningfully to downstream customer behaviour and business outcomes?

Optimization

Can those signals eventually help advertisers allocate budget toward conversations that create genuine commercial value?

And alongside that consumer journey, the advertiser workflow itself is becoming increasingly conversational through AI-assisted campaign creation, creative development, analysis and integrations with CRM and commerce systems.

That creates a broader paid media model worth watching:

Customer Intent → Ad Exposure → Sponsored Conversation → Consideration → Website → Conversion → Business Outcome

with an advertiser workflow increasingly moving toward:

Strategy → Prompt → Campaign → Measurement → Analysis → Optimization

For me, the most interesting part of Sponsored Agents is what happens between the impression and the business outcome.

That is where the next set of paid media questions begins.