How Tracking Works in Digital Out-of-Home (DOOH) Advertising

And How Impression Multipliers Actually Work

Digital Out-of-Home (DOOH) is still treated like a black box.

Most marketers assume:
→ “You can’t really track it”
→ “It’s just awareness”

That hasn’t been true for a while.

In Europe especially, DOOH is already data-driven, programmatic, and measurable. The difference is that it works on modeled exposure, not user-level tracking.

Here’s how it actually works.

 

1. What “Tracking” Means in DOOH

DOOH doesn’t track individuals.

There are:
→ No cookies
→ No user IDs
→ No click-level attribution

Instead, everything is based on opportunity to see.

We estimate how many people could have seen an ad based on:

• Location data
• Traffic and mobility patterns
• Screen visibility
• Time of day

So unlike paid social or search:

→ DOOH = probabilistic measurement
→ Not deterministic tracking

 

2. What Counts as an “Impression”

In digital:
→ 1 impression = 1 ad served to 1 user

In DOOH:
→ 1 impression = 1 estimated opportunity to see

This is calculated using:

• Footfall (pedestrian or vehicle traffic)
• Dwell time (how long people stay near the screen)
• Loop duration (how often your ad appears)
• Screen position and visibility

So already, impressions here are modeled, not counted one-to-one.

 

3. How DOOH Tracking Works in Practice

Step 1: Screen-Level Data

Media owners like JCDecaux, Ströer, and Clear Channel collect data using:

→ Camera-based sensors (anonymized)
→ Mobile data partnerships (aggregated, GDPR-compliant)
→ SDK-based location data
→ Public transport and city datasets

Example:

A screen at Alexanderplatz (Berlin):

• Daily footfall: ~120,000
• Avg dwell time: ~3 minutes

 

Step 2: Ad Play Logging

Every time your ad is shown, it is logged.

Example:

• Loop duration: 60 seconds
• Your ad: 10 seconds
• Share of voice: 1/6

→ ~1,440 plays per day

 

Step 3: Visibility Adjustment

Not everyone passing actually sees the screen.

So vendors apply visibility factors based on:

• Distance
• Angle
• Obstructions
• Brightness
• Time of day

Example:

• 120,000 footfall
• 35% visibility

→ 42,000 estimated visible audience

 

4. Impression Multiplier Explained

This is where DOOH becomes commercially usable.

An impression multiplier adjusts raw exposure to reflect:

→ Repeated exposure
→ Audience density
→ Time spent near the screen

If someone stays near a screen for a few minutes, they may see your ad more than once. That repetition is captured through the multiplier.

 

Example (Simple)

• 10,000 people pass a screen
• Avg dwell time allows ~2 exposures

→ 20,000 impressions
→ Multiplier = 2x

 

Example (Transit Environment)

Location: Munich Hauptbahnhof

• Daily footfall: 300,000
• Visibility rate: 30% → 90,000
• Avg dwell time: 5 minutes
• Loop: 60 seconds
• Share of voice: 1/6

A commuter standing for a few minutes is likely to pass through multiple loops and see the ad more than once.

Typical multiplier:
→ 1.5x to 3x

→ 90,000 × 2 = 180,000 impressions

 

5. How This Works in Programmatic DOOH

This is where the multiplier directly impacts pricing.

In programmatic DOOH:

→ A DSP receives a bid request for a single ad play
→ That play represents multiple impressions

This is often handled through what’s effectively a bid multiplier in the OpenRTB logic.

Example:

• 1 screen play
• Estimated audience: 50
• Multiplier: 2x

→ DSP treats it as 100 impressions

So if you bid €10 CPM:

→ You are paying based on 100 impressions
→ Not one physical play

This is what makes DOOH inventory comparable to other digital channels from a buying perspective.

 

6. Attribution Using Mobile Data (Privacy-First)

DOOH doesn’t track individuals, but it can still measure outcomes.

Important for Europe:

→ Everything is anonymized
→ Everything is aggregated
→ No individual tracking

We are not tracking a person.

We are analyzing movement patterns of cohorts.

 

Mobile Data Matching (GDPR-compliant)

Platforms like Adsquare, Hivestack, and VIOOH enable:

→ Exposure zones around screens
→ Aggregated device movement
→ Post-exposure behavioral trends

 

Example: Store Visit Measurement

Campaign in Hamburg

1. Devices observed in exposure zone (aggregated cohort)
2. Same cohort later appears near store location
3. Compared against a control group

Result:

→ +18% uplift in store visits

 

Example: Cross-Channel Activation

DOOH → Mobile → Paid Media

→ Build geo-based audience segments
→ Activate via programmatic or social

Common across Germany, UK, Nordics

 

6.5 The Reality Check: Not All Multipliers Are Equal

This is one of the biggest blind spots in DOOH.

There is no universal standard.

Different vendors rely on different data inputs:

→ Telco data
→ Camera-based sensors
→ SDK-based mobility panels
→ MAID-based datasets (aggregated mobile IDs)

So in the same city, for similar inventory, you might see:

→ Vendor A → multiplier 2.5x
→ Vendor B → multiplier 1.8x

The difference is not small. It directly impacts reported performance and CPM calculations.

 

Where This Gets More Complex

Even the underlying methodology differs:

→ Some vendors use historical traffic models
→ Others use near real-time sensor data
→ Some rely heavily on mobile SDK panels

There is still ongoing work in Europe to standardize this across markets.

Industry bodies and frameworks exist, but full alignment is not there yet.

 

What You Should Always Ask

→ What is the primary data source?
→ Is the multiplier based on real-time or modeled data?
→ Is there third-party validation?

Because:

→ Vendor math is not always market reality

 

7. Creative Context and Attention

One of the most overlooked parts of DOOH measurement is creative.

Impressions don’t automatically mean attention.

 

Dwell Time vs Creative Length

Example:

• Roadside screen
• Dwell time: ~3 seconds
• Creative: 15 seconds

→ Most users only see part of the message

So while impressions may be high, effective communication is limited.

 

Why This Matters More Now

There is a growing shift toward attention-based measurement.

This includes:

→ Time-in-view
→ Exposure duration
→ Creative-environment fit

In many cases:

→ A short, high-impact message outperforms longer video creatives

Especially in transit-heavy environments.

 

8. From Measurement to Business Impact

Tracking only matters if it drives decisions.

 

Brand Lift Studies

Since DOOH has no clicks, brand lift becomes critical.

→ Compare exposed vs control groups

Measure:

• Awareness
• Recall
• Purchase intent

 

Example

Campaign in Paris:

→ Metro screens vs roadside billboards

Result:

→ Metro screens delivered higher recall
→ Roadside delivered higher reach

 

Practical Use Case

DOOH allows:

→ A/B testing of locations
→ Measuring which screens drive better outcomes
→ Allocating budget based on real-world performance

Example:

→ Screen A drives higher store visits than Screen B
→ Budget shifts accordingly

 

9. DOOH vs Digital Tracking

Factor	Digital (Meta/Google)	DOOH
Tracking	Deterministic	Probabilistic
User-level data	Yes	No
Cookies	Yes	No
Privacy	Lower	High (GDPR-first)
Measurement	Clicks, conversions	Exposure, modeled attribution

---

10. What Actually Matters When Running DOOH

→ High dwell-time locations
→ Frequency, not just reach
→ Understanding how multipliers are calculated
→ Creative adapted to environment
→ Cross-channel integration

 

Final Takeaway

DOOH is not untrackable.

It is:

→ Modeled
→ Privacy-first
→ Programmatic-ready
→ Context-driven

And if you understand how impressions, multipliers, attribution, and attention work:

→ You stop buying screens
→ And start buying measurable audience impact

 

 

The Real Challenge in CTV Advertising: Transparency, Not Complexity

 

Connected TV is growing fast, but the ad supply chain behind it is often misunderstood.

A lot of people say the ecosystem is simply too complex. But after looking more closely at how it actually works, I think the bigger issue is transparency. Buyers often don’t have a clear view of who is selling inventory and how different partners fit into the process.

This article looks at how the CTV supply chain really operates and why improving transparency may matter more than trying to simplify the ecosystem.

 

Understanding the CTV Ad Supply Chain: Why Transparency Matters More Than Simplicity

Connected TV advertising has grown rapidly as audiences shift from traditional television to streaming platforms such as Netflix, Amazon Prime Video, Disney+ and device ecosystems like Roku.

As viewers move toward streaming, advertisers are following. This has made CTV one of the fastest growing areas in digital advertising.

But with this growth has come an ongoing industry conversation: how transparent is the CTV advertising supply chain?

Many people assume the system is simply too complicated and that the solution is to reduce the number of intermediaries involved in selling ad inventory.

However, the reality is a bit different.

The ecosystem is complex for a reason. The real challenge is that buyers often cannot clearly see how the different partners in the ecosystem work together.

 

Why CTV Advertising Involves Multiple Partners

Streaming content requires significant investment to produce and distribute. To make that model sustainable, publishers often collaborate with several technology and distribution partners.

For example, a streaming platform like Disney+ might distribute its app through device platforms such as Roku or Amazon Fire TV so viewers can access content on their televisions.

Behind the scenes, the publisher may use an ad server such as Google Ad Manager to manage advertising inventory.

Once inventory becomes available, it can be sold through supply-side platforms like Magnite or PubMatic, which connect publishers with advertisers.

On the advertiser side, media buyers frequently use demand-side platforms such as The Trade Desk or Google Display & Video 360 to discover and bid on available inventory.

A single ad impression may therefore move through several platforms before appearing on a viewer’s screen.

This layered structure often makes the ecosystem look complicated, but each participant performs a specific role that supports distribution, technology, or monetization.

 

Where the Confusion Starts

Complex systems can work well if the relationships within them are easy to understand.

In the CTV ecosystem, however, buyers often struggle to determine who is actually selling a piece of inventory and whether that seller is authorized to do so.

For instance, imagine a viewer opening a streaming app on a Roku device. The publisher distributes the app through Roku, manages ad inventory through Google Ad Manager, and allows a supply-side platform like Magnite to sell some of the available ad placements.

An advertiser might then purchase that inventory through a demand-side platform such as The Trade Desk or Google Display & Video 360.

From the advertiser’s perspective, the ad travels through multiple platforms before reaching the viewer.

If those relationships are not clearly documented, it becomes difficult for buyers to determine whether a supply path is legitimate or unnecessarily complicated.

 

How ads.txt Helps Improve Transparency

To address this challenge, the industry introduced ads.txt, a transparency tool designed to help buyers verify authorized sellers.

In simple terms, ads.txt is a public file that publishers place on their domain. The file lists the companies that are allowed to sell that publisher’s advertising inventory.

For example, if a publisher authorizes platforms like Magnite or PubMatic to sell its inventory, those companies appear in the publisher’s ads.txt file.

When advertisers evaluate inventory through platforms like The Trade Desk or Google Display & Video 360, the buying platform can check the publisher’s ads.txt file to confirm that the seller is authorized.

This has helped reduce unauthorized reselling and improve accountability across the programmatic advertising ecosystem.

 

Why ads.txt Needs to Evolve for CTV

While ads.txt has been valuable for improving transparency, the Connected TV ecosystem has evolved significantly since it was first introduced.

Publishers now operate across multiple streaming platforms and distribution environments. In many cases, the same inventory can be sold through multiple partners at the same time under different types of agreements.

For example:

• Premium placements might be sold directly by the publisher’s sales team.
• Additional inventory could be sold programmatically through platforms like Magnite or PubMatic.
• Device platforms such as Roku may also participate in monetization depending on distribution agreements.

These arrangements are common and legitimate. However, the current structure of ads.txt does not always capture the differences between these relationships.

Because of this:

• Different types of partnerships may appear identical in ads.txt
• Buyers may interpret supply paths differently
• Platforms sometimes apply inconsistent validation logic
• The same inventory may be accepted by one buying platform but rejected by another

This can create confusion even when the underlying relationships are valid.

 

Improving Transparency Without Removing Complexity

Instead of trying to eliminate complexity from the ecosystem, the industry may benefit more from representing that complexity more clearly.

Several improvements could help.

Clearer partner roles

Publishers should be able to describe the role each partner plays in the supply chain, whether that partner manages technology, distributes content, or operates a marketplace.

Support for multiple authorized sellers

In the CTV ecosystem, multiple partners often have the right to sell the same inventory. Transparency standards should reflect this reality rather than assuming a single selling path.

Signals for trusted partnerships

Publishers could also highlight trusted partners to give buyers additional context when evaluating supply paths.

 

Looking Ahead

Connected TV will likely remain a complex ecosystem because collaboration across multiple platforms helps publishers scale distribution and generate sustainable revenue.

But complexity does not have to lead to confusion.

When the industry improves how supply chain relationships are represented and understood, buyers gain clearer visibility into how their advertising reaches viewers.

The goal should not be to remove participants from the ecosystem. The goal should be to make the roles and relationships within that ecosystem easier to understand.

Greater transparency ultimately leads to stronger trust, and that trust will be essential as Connected TV advertising continues to grow.

 

 

Audience Expansion vs Lookalike Audiences

Understanding Two Core Targeting Approaches in Modern Advertising Platforms

Audience Expansion and Lookalike Audiences often get mentioned together, but they actually work in two very different ways. Knowing the difference makes it much easier to scale campaigns without sacrificing performance.

Many marketers treat them as interchangeable targeting features. In reality, they are based on two different logics that platforms use to discover new audiences.

Understanding when to use each approach is key to building scalable performance campaigns.

👥 Lookalike Audiences: Scaling Through Similarity

Lookalike Audiences start with a seed dataset. This could be:

• CRM customer lists
• Website visitors
• Leads or form submissions
• Purchasers
• App users or subscribers

The platform analyzes patterns inside this group and then finds other users who behave similarly.

In simple terms, the system is asking:

👥 “Who else looks like my existing customers?”

The algorithm studies signals such as:

• Demographics
• Behavioral patterns
• Engagement signals
• Content consumption
• Device usage
• Historical conversions

Because the audience is built from known users, marketers maintain greater control over the targeting logic.

Why Lookalike Audiences Work Well

Lookalikes perform well because they are built from real performance data.

Instead of guessing interests or demographics, the system learns directly from existing customers or leads.

Key benefits include:

✔ Higher probability of conversions
✔ Faster campaign learning
✔ Controlled scaling
✔ Strong alignment with existing customer profiles

Lookalikes are often the first step when scaling beyond warm audiences.

🚀 Audience Expansion: Scaling Through Algorithmic Discovery

Audience Expansion works differently.

You start with a defined audience or targeting setup, but the platform is allowed to go beyond those boundaries if it predicts better performance.

Instead of strict similarity modelling, the system focuses on conversion probability.

The question changes to:

🚀 “Who is most likely to convert, even if they don’t match the original targeting?”

Platforms evaluate signals such as:

• Real-time conversion behavior
• Platform engagement patterns
• Historical campaign data
• Machine learning predictions

If the system believes users outside the original targeting may perform better, it expands the reach automatically.

Why Platforms Encourage Expansion

Advertising platforms increasingly promote expansion features because they allow algorithms to operate with fewer constraints.

A broader audience pool helps machine learning models optimize more effectively.

Benefits typically include:

✔ Increased reach
✔ Discovery of new audience segments
✔ Better algorithmic optimization
✔ Reduced audience saturation

However, expansion also means less manual control for advertisers.

⚖️ The Core Difference

The fundamental difference lies in how the platform decides who to target.

👥 Lookalike Audiences
Similarity-based targeting
Find users who resemble existing customers.

🚀 Audience Expansion
Performance-based targeting
Find users the algorithm predicts will convert.

One approach replicates known customer profiles.
The other focuses on predicting future conversions.

🌐 Similar Targeting Concepts Across Platforms

Many advertising platforms implement variations of these systems.

Some common examples include:

🔁 Similar Audiences (Google)
Modeled audiences based on behavioral similarity to existing lists.

🧠 Predictive Audiences (LinkedIn)
Uses platform signals to estimate which users are most likely to convert.

⚡ Advantage+ Targeting (Meta)
Algorithm-led targeting that dynamically expands beyond manual audience definitions.

🎯 Optimized Targeting (Google Display & YouTube)
Allows Google to expand beyond manual targeting if the system predicts stronger results.

Although terminology varies, most platforms are moving toward algorithmic audience discovery.

🧩 How Advanced Marketers Combine Both

The strongest performance setups rarely rely on just one targeting approach.

Instead, marketers often combine both systems strategically.

A common structure looks like this:

1️⃣ Start with high-quality seed audiences
2️⃣ Build Lookalike audiences to scale similarity
3️⃣ Introduce Audience Expansion once conversion signals stabilize

This layered approach allows campaigns to scale while maintaining efficiency.

💡 Practical Takeaway

Think of it like this:

👥 Lookalike → “Find people like these.”
🚀 Audience Expansion → “Find whoever will convert.”

Both approaches play an important role in modern performance marketing.

Lookalikes help maintain quality and targeting control, while expansion features help platforms discover new pockets of demand that traditional targeting might miss.

When used together, they create a more scalable and adaptive targeting system across platforms like Meta, LinkedIn, and Google Ads.