Cookies are disappearing. DMPs no longer deliver. Tools are limited. Businesses continue to search for new approaches, yet they still run the same old “added to cart → showed a banner” playbook.

That’s a problem. And it’s a systemic one.

That’s exactly how Maksym, GMP Lead at newage. — the first certified Google Marketing Platform agency in Ukraine — opened his talk at ROMI 2025. His session wasn’t about “how to set up analytics.” It was about how to build a full-fledged data infrastructure that:

• replaces a DMP without any custom backend;
• builds audiences not in a UI, but through SQL-based behavioral logic;
• pushes those audiences to ad platforms not “eventually,” but within one hour of the trigger;
• and operates entirely independently of cookies.

“If you’re still building audiences through the GA4 interface, you’re leaving half your potential on the table. In reality, it’s even more,” Maksym said from the stage.

This infrastructure is already running across newage. campaigns — powering both local brands and global markets. And the best part? It’s open. Built entirely on Google tools that are available to everyone. No million-dollar dev budgets are required.

ROMI 2025 became the stage where marketing shifted back to effectiveness over hype. And this talk was a wake-up call: the solutions are here. They’re working. You just need to start using them.

In this article, we break it all down — how a modern data infrastructure looks, how it powers advanced remarketing, and why it’s exactly what you need to navigate the cookie-less future without panic.

Context: How Digital Analytics Is Evolving

Back in 2012, things were simple — and, to be honest, primitive.

Marketers hardcoded tracking tags directly into the site code. Developers tweaked analytics setups between bug fixes. Campaign logic boiled down to “added to cart → showed a banner” — and that was considered enough.

Universal Analytics was the go-to tool, Google Tag Manager had just entered the scene, and the idea of a centralized data infrastructure felt like something out of a sci-fi movie.

Thirteen years later, everything has changed.

Today, most teams use GTM by default. Events are configured in hours, GA4 streams data into BigQuery, and segments are activated in ads directly via Cloud Functions.

The entire pipeline — from tracking to remarketing — has evolved into a flexible, modular architecture.

No manual CSVs. No hacked-together APIs. No waiting.

“Infrastructure is about flexibility, speed, and control. If you need to change a segment, you don’t have to wait a week for a developer. You tweak a SQL query and launch the campaign within an hour,” Maksym explained from the ROMI stage.

And that’s the core shift: analytics is no longer just reporting. It has become a real-time decision-making system.

Where “setting up analytics” once meant sending events to GA, today it means building your own Customer Data Platform (CDP) — without a paid DMP, without a custom backend, and with:

• custom segmentation,
• push audiences directly into media platforms,
• analytics that measures real engagement, not just clicks.

That’s the transformation Maksym mapped out step by step — how marketing moved from “everything via UI” to “everything through data.”

Google Tools as the Foundation: What You Need to Launch

“We don’t need expensive software. We need logic — and the right architecture. Google already gives us everything.”

To build an infrastructure that truly works with audiences — in real time, with flexible logic, and no cookie dependency — there’s no need to reinvent anything. Everything you need already exists — in the form of publicly available Google tools.

Here’s the core stack behind a modern data infrastructure:

Google Tag Manager (GTM)

Web or server-side version.

Collects events, clicks, and behavioral interactions. GTM gives marketers full control over tracking, eliminating the need for developers. The web version is fast to implement, while the server-side version offers stronger resistance to blockers and allows for deeper control over data flow.

GA4 or GA360

The analytical core.

GA4 is great for basic scenarios and small to mid-sized businesses. GA360 is essential for those handling large volumes of events, needing frequent data updates, and requiring real-time integration with BigQuery.

BigQuery

Raw data warehouse.

This is where custom segments are built — the kind you simply can’t create in the GA interface: behavioral sequences, time-based logic, CRM joins, RFM, LTV, and more. Data is stored losslessly and can be queried with any level of SQL complexity.

Cloud Functions

Audience pushes automation.

These functions connect BigQuery with ad platforms, transferring segments (user_id / device_id) without any manual work. Audiences update automatically, even hourly if needed.

DV360 (Display & Video 360)

The main DSP for activation.

Ingests custom audiences from BigQuery and lets you combine them with in-market, affinity, and 3rd-party segments. All campaign types — are managed from a single interface.

Campaign Manager 360 (CM360)

The control center.

This is where complete analysis happens: how many impressions were viewable, what the frequency was, how creatives and audiences performed together, and what drove conversions, including post-view interactions.

How does it all work together?

Here’s a quick overview of the architecture in action:

1. GTM collects on-site events (web or server-side).
2. GA (4 or 360) aggregates the data and sends it to BigQuery.
3. A segment is built in BigQuery using SQL.
4. A Cloud Function pushes that segment to DV360 or Google Ads.
5. CM360 tracks performance, enabling full-funnel optimization.

This is a complete CDP-level architecture, but without custom development, or proprietary software, and fully adaptable to your business goals.

GTM + GA4/GA360: The Foundation of Event Collection

Every modern marketing infrastructure starts with events. If you’re not capturing user behavior, you simply can’t act on it. No remarketing, no personalization, no analytics.

That’s why the first critical component of this system is Google Tag Manager (GTM).

GTM: Web or Server-Side?

In 2025, most businesses still run on the web version of GTM — it’s easy to implement, fast, and widely supported. But there’s a second option: server-side GTM, which:

• runs through your own server (e.g., Cloud Run or App Engine),
• gives you full control over the data sent to analytics and third-party platforms,
• bypasses AdBlock and similar tracker-blocking tools.

When should you go server-side?

• If you work with sensitive data (e.g., banking, e-commerce).
• If you want to minimize data loss from AdBlockers, Safari, or iOS restrictions.
• If you need centralized control over all your data flows.

GA4 vs GA360: Which One Fits Your Business?

The next layer is analytics, and here, you have two options: GA4 (free) and GA360 (paid).

At first glance, they seem similar: both offer tracking, audience building, and BigQuery integration.

But the real difference lies in scale, speed, and flexibility.

“Without GA360, you simply won’t see everyone on your site — and you won’t be able to react in time,” Maksym emphasized during his ROMI talk.

When does GA4 stop being enough?

• You’re handling over 1 million events per day
• You need to update dashboards or audiences hourly
• You work across multiple markets, websites, or apps
• You’re integrating analytics with media platforms, CRM, or a data warehouse
• You need full control over the user journey in a complex conversion flow

How to Structure Your Event Tracking

The most common mistake we see? Collecting events “for reporting” — not for action.

What matters is this:

• Don’t think “What happened?” — think “How can I use this?”
• Don’t duplicate GA4 events in the UI if you plan to segment in BigQuery
• Define your core: key events, parameters, and priorities (e.g., video_start, cart_add, scroll, session_depth)

Pro tip: Even during the GTM setup stage, flag the events that you’ll later use for segmentation. This shortens the path to custom logic and speeds up audience activation.

So in the end, GTM + GA4/GA360 aren’t just about “analytics.” They’re signal collection tools — the foundation for segmentation, remarketing, creative targeting, and campaign optimization.

Next, let’s look at how this data comes to life in BigQuery.

BigQuery: The Kind of Segmentation You’ll Never Achieve in GA

GA4 is about events. BigQuery is about logic.

And that’s where the true power of audience building unfolds — when you’re no longer limited by the UI, stuck with 5 basic conditions, or waiting for delayed updates.

How GA Connects to BigQuery

Both GA4 and GA360 support exporting raw events into BigQuery — a complete event-level log of every user interaction with your website or app, including parameters, attributes, user_id, timestamps, traffic sources, and more.

• In GA4 (free), you get daily exports, with a volume limit (~1M events/day).
• In GA360, you get streaming exports, near real-time, with no volume cap.

Why does raw data = total freedom?

• You see all events, not just what’s surfaced in reports or the UI.
• You can access time-based sequences (via SQL window functions).
• You can join GA data with CRM, e-commerce, or even Excel-based sources.
• You can write any custom logic in SQL — no interface limitations.

For example, you can’t build an audience in GA4 UI like: “User clicked a banner → then added a product to cart 2 hours later → but didn’t complete a purchase within 3 days.” In BigQuery, that’s just a SQL query.

Examples of Custom Segments Only Possible in BigQuery

Behavioral Scenarios:

• 2+ visits in a week from different devices
• Scrolled 80% of a page + started video + exited without purchase

Time-Based Logic:

• Performed action A but not action B within N hours
• Returned after seeing a banner, but didn’t enter a session

RFM Models:

• Frequent buyers (10+ transactions in 90 days)
• Users with high revenue but low frequency

CRM-Integrated Models:

• Clients marked as “churn risk” or “high priority”
• Users from specific segments (e.g., B2B, VIP, test group)

BigQuery is your segmentation lab.

In GA4, you see what happened. In BigQuery, you decide who to act on — and how.

“The more behavioral scenarios you model, the more points of impact you unlock. That means less budget burned on ‘everyone’ and more impact on ‘your people’.” — Maksym, newage.

GA4 Audiences vs BigQuery Segments: What’s Better — and Why?

One of the most common questions we hear: “Why do we need BigQuery if we can already build audiences in GA4?”

The short answer? Because the UI is just the surface, segmentation is about depth.

So what’s the difference? Let’s break it down.

The Risks of Relying Only on GA4:

• Data Loss — In free GA4, after ~1 million daily events, data gets sampled or truncated.
• Delays — You launch a Flash Sale campaign, but the audience won’t update until the next day.
• Lower Precision — It’s hard to account for behavioral nuance like return visits from different devices or cross-platform interactions, especially if Google Signals aren’t active or user IDs aren’t consistent.
• No CRM Link — You can’t isolate VIPs or users who haven’t paid in 60 days, for example.

When Is Custom Segmentation Critical?

Flash Sales / Black Friday / Time-Sensitive Promotions

• Success depends on hours, not days.
• GA4 often can’t refresh audiences fast enough.

High CAC Products

• You can’t afford to burn your budget on inaccurate targeting.
• You need deep filtering by actions, traffic sources, and history.

B2B / Enterprise Campaigns

• The decision-making journey is long and complex.
• Behavioral patterns matter more than single actions.

“You don’t need to be a data scientist in BigQuery. You just need a scenario and one SQL query. That’s all it takes to build a segment that works. Not just ‘visited the site,’” — Maksym, newage.

GA4 is a great starting point. But if you want flexibility, speed, and precision, BigQuery transforms analytics from a passive dashboard into an active performance tool.

Audience Activation: How to Push Segments into Ads

Collecting data is only half the battle. Segmenting it — another crucial step.

But the real game-changer is activation: How quickly, accurately, and seamlessly can you deliver the right message to the right audience? That’s where real performance begins.

How It Works: Automatically, No Hacking or CSVs

Once your custom audience is built in BigQuery, the next step is handled by a Cloud Function — a serverless function in Google Cloud that automates the process. It:

1. Takes the segment (a table with user_id, device_id, or hashed_email)
2. Converts it into the correct format (e.g., JSON or in-memory CSV)
3. Pushes the audience directly to:
   
   • DV360 — for display or video campaigns
   • Google Ads — for search or remarketing
   • Other DSPs — via API or direct upload

It All Runs as an Automated Pipeline. No manual updates. No export–import hustle. No delays.

Sample Scenario:

1. A user visits a product page twice, from two different devices
2. They scroll to the technical specs but don’t add to the cart
3. BigQuery detects this behavior and flags it as “hesitating.”
4. A Cloud Function pushes this audience into DV360 within 1 hour
5. A tailored creative launches: “Still have questions? We’re here to help!”

“We don’t show ads just because someone ‘added to cart’. We show them because they came back twice, on different devices, and something stopped them. And that’s what we respond to,” — Maksym, newage.

Why This Matters:

• You’re not waiting 24 hours for a standard GA4 audience to update
• You react to behavior at the moment, not “after the report.”
• You’re not targeting a broad segment — you’re hitting a specific intent or pain point
• You’re not limited to Google Ads — DV360 lets you combine 1st-party, 3rd-party, in-market, affinity, and CRM segments all at once

BigQuery + Cloud Functions = a reactive remarketing model. That doesn’t kick in after the fact — it activates right when the decision is happening.

Campaign Manager 360: How to analyze what works

When it comes to display advertising, clicks are not an accurate indicator. On average, less than 1% of conversions in display campaigns happen after a click. The rest come from post-view impact — something standard analytics doesn’t show.

That’s why Campaign Manager 360 (CM360) is essential — not as just another platform, but as the single source of truth for analyzing creatives, frequency, reach, and user interaction.

Why do you need CM360?

• Attribution: the only model that connects all touchpoints — including post-view, not just clicks.
• Viewability: see which impressions were viewable (≥1 second on ≥50% of the ad surface).
• Frequency and reach: when, where, and how often a user saw an ad, and when it became too much.
• Audience overlap: who saw the ad on YouTube and then encountered a banner on mobile.
• Cross-channel visibility: unified reporting across all platforms in a single dashboard.

“Without CM360, you’re only seeing clicks. But in the display, a click ≠ performance. We see where the user got overwhelmed, where the creative underperformed, and where the audience responded without clicking,” — Maksym, GMP Lead, newage.

What is Comprehensive Analysis (Comprehensive Analysis Framework)?

It’s not just “another dashboard.” It’s a full methodology developed by the newage. team and applied in every project that leverages CM360.

This framework is built around three core components:

1. Frequency and delivery

• How many impressions did each user receive?
• How does performance change at 3+, 5+, and 10+ frequency levels?
• Is the audience being oversaturated?

2. Viewability and reach

• Was the ad viewable?
• Where did repeated exposure occur, and where were the blind spots?
• What was the cross-channel reach?

3. User response (post-view)

• Did users interact after seeing the ad?
• Which creative triggered the conversion?
• Was there overexposure — or perhaps not enough traffic?

What you gain with CM360:

• True transparency across all your advertising activity
• The ability to optimize based on real insights, not guesswork
• A platform that connects audiences, creatives, and media into one cohesive view
• And most importantly, clear answers to critical questions: What’s working? Where are we oversaturated? What should be paused, and what should be scaled?

Without CM360, you’re only seeing the surface: clicks and basic conversions. With CM360, you uncover the cause-and-effect: Who saw the ad → how often → on which platform →, and what came of it.

That’s what infrastructure for action looks like — not just “we launched something,” but “we know what worked — and what to do next.”

A Cookieless Future: Why BigQuery Is a Resilient Choice

BigQuery is your marketing insurance policy for 2025 and beyond.

The gradual disappearance of third-party cookies is no longer a theory — it’s a fact. Safari, Firefox, and now Chrome are moving in the same direction: traditional cookie-based remarketing is becoming obsolete.

• GA4 remarketing can no longer “see” the user if the cookie isn’t preserved.
• Audiences become incomplete.
• Targeting loses precision.
• The cost of reaching users increases.

“In two years, most advertisers will lose remarketing. We won’t — because our segments aren’t built on cookies, but on behavior,” — Maksym, GMP Lead at newage.

Why BigQuery Is Cookie-Independent

BigQuery segments aren’t built on client_id or third-party cookies — they rely on identifiers you control, such as:

• user_id — from login/authenticated sessions
• device_id — for apps or mobile traffic
• hashed email or phone — from CRM, transactions, or lead forms
• Any custom logic you can track or import

These identifiers are stored in BigQuery and used to build segmentation, regardless of whether cookies still exist.

Privacy-First Marketing: Not a Requirement — a Competitive Advantage

• You build your segments instead of relying on the browser.
• You control where and how identifiers are stored.
• You can combine behavioral data with CRM, without violating privacy laws (when done correctly).
• And most importantly, you don’t lose audience access when cookies disappear.

GA remarketing, as we knew it, is fading. DMP solutions are expensive, bulky, and still dependent on the same logic. BigQuery is your first-party CDP — one that:

• runs on your data,
• scales with your business needs,
• and doesn’t disappear with the next browser update.

Summary: What This System Delivers for Your Business

Everyone talks about data, but few truly use it as infrastructure.

Maksym’s talk at ROMI 2025 proved one thing: building your own CDP is no longer a fantasy — it’s a reality.

Here’s what this model gives your business:

• Full control over audiences — from data collection to ad activation
• Real-time responsiveness — segment → push → campaign in under an hour
• Smarter targeting = higher CTR, lower CPA
• Cookie independence — your system works tomorrow, and three years from now
• Holistic analytics via CM360 — measuring impact, not just clicks

Who needs this most?

• E-commerce — for dynamic remarketing, RFM logic, and time-sensitive promos
• B2B — for personalization, CRM integration, and understanding complex user journeys
• Banking / Fintech — for working with authenticated users, offline signals, and multi-step funnels
• SaaS / Digital Products — for user-based analytics, NPS, retention modeling, and activation through ads

When to start implementing it?

If you already have:

• GTM (event tracking set up)
• GA4 or GA360 (analytics in place)

→ Then your next step is BigQuery.

From there — Cloud Functions + DV360/CM360.

All within the Google ecosystem. No custom backend. No API hacking. No data loss.

Want to see how it could work for your business?

We’re ready to:

• Audit your current analytics setup
• Propose a scalable architecture
• Show you how it works

Just reach out, and we’ll walk you through it in action.