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.