The Great Pipe Debate

The marketing industry has spent five years debating which pipes work better for finding the right people.

But the transformation isn't happening in the pipes. It's happening in what flows through them. I posit that:

When content adaptation costs $50,000, broadcast economics force you to optimise for reach efficiency. You need the right pipes.

When content adaptation approaches $0, personalisation economics force you to optimise for individual relevance. You need the right data.

First-party data advocates argue for deterministic targeting: finding Jane Smith specifically, tracking her journey precisely, measuring her response accurately. They point to measurement clarity, and independence from platform control.

Third-party data pragmatists counter with probabilistic scale: finding people like Jane Smith across massive audiences. They cite reach efficiency, lookalike modeling sophistication, and the reality that most brands lack meaningful first-party data.

Conferences dedicate entire tracks to this debate. LinkedIn is full of heated arguments about accuracy versus reach, precision versus scale, deterministic versus probabilistic.

But notice what both sides are discussing: targeting efficiency. Who can find Jane Smith fastest, cheapest, most accurately?

This is pipe thinking. They're optimising the delivery mechanism while ignoring what's being delivered.

Here's the transformation that makes the entire targeting efficiency debate obsolete: the marginal cost of content adaptation is collapsing toward zero.

Not someday. Right now.

The crucial difference: content adaptation is now driven by media signals, not creative decisions made weeks earlier.

When Jane Smith browses outdoor equipment, abandons cart, returns via email, and clicks through to product page, those aren't just targeting signals anymore. They're content signals. They tell you what message Jane needs in this specific moment.

The pipes (whether you find Jane deterministically or probabilistically) matter far less than whether you can adapt content based on what those pipes tell you about Jane.

Let me be specific about what "adaptation cost approaching $0" actually means in practice. The GenAI stack enables this:

All of this happens automatically, driven by the media signals your pipes are already collecting.

Now the strategic importance of first-party data becomes clear, and it has nothing to do with media targeting efficiency.

With third-party data:

With first-party data:

The value isn't finding Jane deterministically instead of probabilistically.

The value is knowing enough about Jane to personalise the content she sees based on the signals her behaviour provides.

When content adaptation cost $50K per variation, this distinction didn't matter. You couldn't afford to personalise anyway.

When content adaptation approaches $0, you can't afford NOT to personalise - because your competitors are.

This is why major operations are restructuring. Production capabilities are being nested within media operations because the economic model changed.

When adaptation was expensive, you produced assets first, distributed them second. When adaptation is nearly free, you distribute signals first, generate content second.

This is the shift media strategists miss because they think about audiences, not content. The pipes debate assumes content is fixed and you're optimising who sees it. The actual transformation is content becoming variable and media signals determining what each person sees.

The first-party versus third-party debate lives in media planning circles, where careers were built on targeting efficiency. These are people who spent 15 years optimising CPMs, negotiating rates, and proving their media plans reached the right audiences.

They think about finding people, not adapting content for people.

But here's the uncomfortable truth: AI isn't new to programmatic media. Machine learning has powered bidding optimisation for a decade. Finding audiences efficiently? Solved problem.

The unsolved problem was always: What do you show them when you find them?

Broadcast economics forced one answer for everyone. GenAI economics enable personalised answers driven by media signals.

The deterministic versus probabilistic debate matters when you're trying to find Jane Smith efficiently. It becomes existentially critical when you're trying to adapt content for Jane Smith specifically - because probabilistic targeting can't tell you enough about Jane to personalise effectively.

You need to actually know Jane Smith to adapt content based on her specific signals. That requires first-party data.

Not for targeting efficiency. For content adaptation capability.

1. Production Becomes a Media Function, Not a Creative Function

The traditional separation, creative agency produces assets, media agency distributes them, becomes an operational liability when competitive advantage comes from content responding to media signals in real-time.

Early evidence: Media operations building in-house production capabilities. Creative agencies embedding staff within media workflows. The organisational boundaries that defined agency structure for decades are collapsing because the economics changed.

2. "Creative Excellence" Gets Redefined

When you're producing 1,000 variations driven by media signals instead of 3 perfect executions from creative instinct, the concept of "craft" fundamentally changes.

Creative excellence is no longer: singular perfect execution refined over weeks
Creative excellence becomes: systematic personalisation responding to signals at scale

The best creative teams aren't producing the most beautiful assets. They're building frameworks that generate hundreds of contextually relevant executions automatically based on media signals.

3. First-Party Data Investment IS Production Investment

Companies currently allocating 11.2% of digital budgets to first-party data infrastructure think they're investing in targeting efficiency and measurement accuracy.

They're actually investing in content adaptation capability.

The ROI calculation isn't "Can we find Jane Smith 5% more efficiently?" It's "What's the revenue lift from showing Jane Smith personalised content based on her specific signals versus showing her the broadcast message?"

4. Platform Control Becomes Existential Threat

Meta's Advantage+ and Google's Performance Max aren't just taking over media planning. They're positioning themselves to control content adaptation.

When Meta knows your audience better than you do (because they have behavioural signals you don't), they can adapt content better than you can. Their Brand Concierge AI can personalise messaging based on signals you never see.

Without independent first-party data infrastructure, you're not just losing targeting control. You're losing the ability to personalise content based on your own audience signals.

You become dependent on platforms to adapt your content for your customers. That's not a media relationship. That's strategic surrender.

The competitive advantage is moving upstream, to the content adaptation capability that first-party data enables. Not finding Jane Smith deterministically instead of probabilistically. Knowing Jane Smith well enough to show her personalised content based on the signals her behaviour provides.

The 1PD versus 3PD debate isn't about better pipes. It's about whether you control the signals that determine what flows through those pipes.

And when content adaptation is nearly free, controlling those signals is the entire game.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 21 November 2025 in the Brandflow newsletter on LinkedIn.