$ZETA: Winning a Negative-Sum Game
Data is the prize
The content of this analysis is for entertainment and informational purposes only and should not be considered financial or investment advice. Please conduct your own thorough research and due diligence before making any investment decisions and consult with a professional if needed.
Marketing is not just competitive.
It is becoming negative-sum.
Not because “everyone loses money.”
But because the real scarce resource is data, and data cannot enrich everyone at the same time.
Think about it like oil fields.
If one company signs the contract to extract from the best field, competitors do not get to extract from that same field. The resource still exists, but access is exclusive.
That is the game Zeta is playing.
And the reason I’m spending time on $ZETA is simple:
Zeta is not trying to win by having “better software.”
Zeta is trying to win by locking up the best data relationships.
If they succeed, competitors do not just fall behind.
Competitors get starved.
1) What I mean by “negative-sum”
In a normal market, many players can improve together. You can build a great product, and competitors can build something similar. The pie grows.
In Zeta’s market, the key input is:
first-party enterprise data
identity data
transactional and loyalty signals
omnichannel behavior
And the key word is limited.
Enterprises do not do “deep partnerships” with ten vendors at the same time. They pick one system, integrate it, feed it, trust it, and then build workflows on top.
So the game becomes:
Zeta signs a big enterprise.
Zeta plugs into that company’s first-party data.
Zeta improves outcomes for that company.
That same first-party data does not enrich Zeta’s competitors in the same way.
So every enterprise that joins Zeta is not only a customer gained.
It is also a training and tuning opportunity competitors do not get.
That is negative-sum.
Because the “best data pie” does not expand fast enough to give everyone the same quality of learning loops.
2) AI makes this more brutal: data access compounds
In the old world, martech competed on features:
better UI
more connectors
nicer dashboards
more automation rules
Now software is easier to replicate.
AI changes what matters.
When outcomes depend on models and algorithms, the company with the best data relationships gets:
better identity resolution
better predictions
better personalization
better measurement
better ROAS
And that loops back into sales.
Zeta’s CEO framed it directly as ROI:
Every dollar that is spent through the Zeta Marketing Platform returns a 6x return in revenue to our clients… on par with pretty much any walled garden.
If that is even approximately true, it creates a dangerous flywheel.
3) First-party data is the difference between “AI copilots” and real ROI
Here is my blunt take:
Most “AI marketing apps” without a strong proprietary dataset are just copilots.
Nice UI. Nice text generation. Nice automation.
But without deep data, they are generic.
They cannot “know.” They can only “suggest.”
For AI applications in marketing, data is the critical element. Without it, AI is mostly a layer on top of weak signals.
$ZETA stands out because it does not start from “AI features.”
It starts from the foundation: identity + first-party data + contextual signals, at scale.
And then AI becomes useful.
4) Why Zeta’s identity layer matters (and why they brag about match rates)
Zeta tells a story that is not random marketing.
A CMO expects a 70% to 75% match rate.
80% is “gold standard.”
Zeta plugs the company’s first-party data into their graph and says:
first run: 92%
second run: 94%
The CMO says “impossible.”
This is not just a flex.
This is the whole business.
Because if you can match more identities, you can:
target better
measure better
retarget better
personalize better
waste less spend
It is like fishing.
If you can see 94 fish out of 100 in the lake, you fish differently than someone who sees 70.
In a world where cookies die and signals get restricted, match rate becomes a moat.
5) “Age of answers” vs “age of data”
There is a broader shift happening.
We are moving from:
data-driven → answers-driven
Marketers do not want a dashboard with 300 metrics.
They want:
“What is working?”
“What is not working?”
“What should I do next?”
“Where do I allocate budget tomorrow?”
In other words: they want actionable answers that produce ROI.
This is not only true in marketing.
It is true for AI in general.
If AI has to “pay for itself,” it cannot just aggregate data.
It must produce answers and decisions that create measurable outcomes.
Zeta is positioned well here, because with the right data foundation, “answers” are not just guesses. They can be grounded in real customer behavior.
6) The platform is not “vibe-codeable” (and this is underestimated)
A lot of people today think: “AI can build software, so moats will disappear.”
True for some categories.
Not true when the core value is a complex, deterministic system + data + routing + context.
Zeta’s platform is hard to replicate with casual vibe-coding because:
It is built on a deterministic foundation, built in-house over years.
It requires deep expertise in routing technology through LLMs and MCPs to keep functionality reliable, not just “cool demos.”
And the real crown jewel is contextual intelligence: memory, personalization, and non-generic agents that feel like extensions of human users, not basic tools.
In short: you can copy UI.
You can copy AI features.
You cannot easily copy a system that combines deterministic infrastructure with contextual intelligence and data access at scale.
7) Switching costs are weakening, so Zeta is attacking migration friction directly
One of the most interesting points from management is about AI agents that help migration from legacy systems.
They basically say:
Zeta’s agents can look at the actual code and things a company built in systems like Salesforce, and recreate them inside Zeta, removing friction.
This matters because it highlights the bigger trend:
pure software is no longer a defensible moat.
Switching barriers get weaker.
So the real moats become:
data
ecosystem
distribution
outcomes (ROI)
Zeta is playing this correctly. They reduce migration pain, then they try to keep customers with outcomes and data advantages, not with lock-in UI.
8) Acquisitions: M&A as a data weapon (17 acquisitions in 17 years)
Zeta has done 17 acquisitions in 17 years.
That matters because it signals something very specific: this is not a company that occasionally buys growth. This is a company that treats acquisitions as a core operating strategy.
In a negative-sum market where data is scarce, M&A becomes a way to buy exclusive access to signal.
Two examples:
LiveIntent (2024): hashed email scale and publisher-side engagement signals. This strengthens identity resolution where cookies are weak and attribution is messy.
Marigold enterprise business (2025): loyalty, transactional, retention, and omnichannel behavioral data. This is the “real economy” layer that makes models smarter because it reflects what people actually buy and repeat.
Put simply: Zeta is building its moat the hard way, one dataset at a time. And if they keep doing what they have done for 17 straight years, the message is clear: they plan to keep consolidating scarce data sources faster than competitors can replicate them.
9) The flywheel (and why the market becomes “winner takes more”)
Zeta’s moat is reinforced by a powerful flywheel:
More first-party data and identity signal
Better personalization and measurement
Better ROI for clients
More clients join
Even more data and signal
Repeat
And here is the negative-sum twist:
Customers that partner deeply with Zeta and contribute data are unlikely to do the same with Zeta competitors.
So as Zeta scales, competitors do not just lose market share.
They lose access to the best learning loops.
That is how leaders get disproportionately larger.
This is why I call it negative-sum.
10) Important nuance: “Zeta aggregates first-party data across its customer base”
This is core to the value proposition.
Zeta connects client first-party data to its identity graph and uses it to drive better outcomes. In practice, the platform gets stronger as it sees more real-world behavior across more clients and channels.
Even if client data is handled with privacy constraints and processing rules, the strategic point remains:
Zeta becomes the system sitting in the middle of high-quality first-party relationships.
And in a world where first-party relationships are the scarce input, that position is extremely valuable.
Conclusion: Zeta is winning a negative-sum game by capturing scarce data relationships
Zeta is interesting because it may be building a moat that looks like:
identity scale
proprietary datasets (plus acquisitions)
exclusive enterprise first-party data relationships
an answers-driven platform that proves ROI
and a flywheel that makes the leader stronger while starving competitors
In a market where data is scarce, that is how you win.
Because every enterprise that chooses Zeta is:
revenue for Zeta,
a “no” for competitors,
and one less source of premium learning loops for everyone else.
That is compounding.
That is the bet.
This was my first deep dive into the qualitative side of Zeta. It has sparked real interest, so I’ll be going deeper with a follow-up piece focused on quantitative validation of the thesis. Earnings season sits right in the middle, so expect it in a few weeks.
In the meantime, please challenge anything I’ve said — I read every comment and will address the strongest ones in the next article.