Zeta Global
Holdings Corp.

Core Business

Zeta Global operates an AI-powered marketing cloud — the Zeta Marketing Platform (ZMP) — that helps large enterprises acquire, grow, and retain customers across every addressable channel: email, connected TV, SMS, social media, display, and web. Founded in 2007 by David Steinberg and John Sculley, the company's founding premise was that proprietary consumer data, not creative, would become the decisive variable in enterprise marketing. This thesis has proved remarkably prescient.

The platform solves a fragmented vendor landscape problem. Most large enterprises historically pieced together five to fifteen disparate point solutions — a data management platform, an email service provider, a loyalty engine, a CTV activation tool, an identity resolution layer — all generating incompatible data. Zeta's "One Zeta" model collapses this stack into a unified platform. A marquee apparel retailer, cited in Q1 2026 earnings, consolidated from four vendors down to Zeta alone. One Fortune 100 telco is forecasted to increase Zeta spending by 18x in 2026. This consolidation dynamic is the engine of Zeta's ARPU expansion.

The SuperGraph — The Core Proprietary Asset

Zeta's proprietary SuperGraph database processes approximately 1 trillion consumer signals per month across roughly 245 million U.S. consumer profiles. This database underpins every product and creates the data flywheel: more customers generate more first-party data, which improves targeting, which attracts more customers. Importantly, Zeta is one of the few marketing technology platforms that owns both the data layer and the execution layer simultaneously — meaning it does not depend on third parties for either consumer identity or channel activation.

Revenue Segments

Zeta reports revenue primarily as a single "Integrated Platform" segment. However, operationally the business can be bucketed into three functional areas:

• Platform Software/SaaS (~70%+ of revenue, high margin): Subscription and usage-based contracts for access to ZMP, including analytics, AI-driven audience targeting, cross-channel campaign execution, and loyalty management (post-Marigold). This is the highest-margin, most recurring bucket.
• Agency/Performance Marketing (~20-25%, moderate margin): Managed services where Zeta deploys campaigns on behalf of clients. Less recurring but creates stickiness as clients rely on Zeta's data-fueled outcomes.
• Political Candidate Revenue (~low single digits %, lumpy): A wildcard segment that contributes meaningfully in even-year election cycles. FY2026 guidance assumes ~$15M from political revenue. This is 100% incremental contribution margin but creates noise in Y/Y comparisons.

The Marigold acquisition (closed Nov 2025, ~$325M consideration) adds a heavily subscription-dominated revenue stream: Marigold Enterprise reported greater than 90% subscription revenue with cost of revenue below 30% — structurally accretive to Zeta's blended quality profile.

Revenue Quality & Unit Economics

Revenue quality is above average for a company at this growth rate. Super-Scaled Customers (those spending over $1M annually) numbered 189 as of Q1 2026, up 19% year-over-year. ARPU for this cohort reached $1.7M, up 21% YoY — meaning both the number and depth of enterprise relationships are expanding simultaneously. Forrester documented a 600% average marketing ROI for Zeta customers and a 23% YoY net promoter score improvement. Net revenue retention has not been explicitly disclosed at a single number, but the ARPU expansion trajectory implies strong negative churn dynamics. Customer acquisition costs are not disclosed, but the company's go-to-market motion relies heavily on an enterprise direct sales force, which carries high CAC offset by multi-year contract structures with meaningful switching costs.

Revenue Scale & Growth Trajectory

2026E = midpoint of management guidance as of Q1 2026 earnings (April 30, 2026)

Geographic & Customer Concentration

Revenue is primarily U.S.-centric (roughly 85-90%). Marigold expanded the European and nascent APAC footprint, but international revenue remains a small minority. No single customer has been disclosed at above 10% of revenue, which is consistent with Zeta's stated enterprise diversification. The business serves approximately half of the Fortune 100, which reduces concentration risk but also signals increasing penetration potential within the existing base.

Profitability

The headline narrative is consistent margin expansion alongside accelerating revenue growth — an unusual and genuinely attractive combination. Gross margins are expanding as higher-value platform software displaces lower-margin managed services and as integration synergies from recent acquisitions flow through. Adjusted EBITDA margins are guided to reach 22.3% for FY2026, up roughly 800 basis points over the past four years. The path to full GAAP profitability in FY2026 (guided EPS $0.02–$0.04) represents a meaningful threshold, as the company will transition from "loss-making growth company" to "profitable, high-growth compounder" in institutional classification frameworks, which could expand the buyer universe.

Cash Flow Quality — The Critical Caveat

The gap between Adjusted EBITDA and GAAP net income is wide and deserves scrutiny. Stock-based compensation (SBC) is substantial — running at roughly $150–200M annually on a TTM basis — and is excluded from Adjusted EBITDA but represents real economic dilution to shareholders. FCF is reported after cash taxes and capex but before SBC, meaning the "true" owner earnings picture is softer than the FCF headline suggests. For FY2026, if FCF guidance of $235M is achieved against roughly $150M in SBC, the FCF after adjusting for SBC dilution is approximately $85M — implying a price-to-owner-earnings multiple of ~67x at current market cap. This is the most honest single valuation data point and it is not cheap. However, if the company delivers on 2028 targets ($371M+ FCF, $573M+ adj EBITDA), the current price is very reasonable.

Balance Sheet

The balance sheet is in sound condition. Zeta is net cash positive following the Marigold acquisition, which was funded partially via stock issuance. Interest coverage is approximately 23x on EBIT, making the debt load essentially irrelevant from a financial risk perspective. Operating leases and deferred revenue are manageable. The Marigold deal structure (partial stock consideration) preserved liquidity while adding $190M+ in annual revenue. The decision to announce a $100M share repurchase program in Q1 2026 signals management confidence in the balance sheet and near-term cash generation.

Capital Intensity & ROIC

Software businesses by nature have low maintenance capex relative to revenue. Zeta's total capex runs roughly 2-3% of revenue, with the majority directed toward platform development and data infrastructure rather than physical assets. ROIC remains negative on a GAAP basis due to the accumulated goodwill from acquisitions and ongoing GAAP net losses, but on an adjusted basis the platform-level returns are attractive and improving. The real watch item is whether the cumulative acquisition spend (Disqus, Boomtrain, LiveIntent $250M, Marigold $325M, and dozens of smaller deals) ultimately generates visible returns that justify the combined investment. Evidence is encouraging but not yet conclusive at the consolidated level.

CEO: David A. Steinberg

Steinberg is a career founder-entrepreneur who co-founded Zeta in 2007 alongside John Sculley (former CEO of both Apple and Pepsi-Cola) after a prior career that included co-founding Sterling Cellular and InPhonic (a leading online wireless retailer, which grew to a billion-dollar business). He has led Zeta for 19 years, taken it from concept to a $1.3B private unicorn, and then to a NYSE-listed company. His background is pure operator and entrepreneur — not a financial engineer or hired executive, which is relevant because his incentives are structurally aligned with long-run value creation rather than short-run optionality extraction.

Steinberg's stated thesis from Zeta's founding — that data would transform marketing before "big data" was even a mainstream term — has proven accurate over 18 years. The 19 consecutive quarters of beating and raising guidance through Q1 2026 is an exceptional track record of execution in a sector notorious for guidance uncertainty. He tends to set a conservative 2–5% buffer above internal forecasts and has consistently outperformed. The one significant blemish on his tenure is the November 2024 Culper Research short-seller report, which caused a 37% single-day stock price collapse. Steinberg's rapid, detailed rebuttal and the subsequent recovery to near all-time highs suggests the core business allegations were either overstated or immaterial — but the episode raised legitimate questions about data sourcing practices that have not fully faded.

Skin in the Game

Steinberg owns approximately 3.5 million shares valued at roughly $56M at current prices, and retains a 10% economic stake plus over 50% of total voting power through Class B shares (10 votes per share). The dual-class structure means public shareholders have limited ability to constrain management decisions, including capital allocation. This is a genuine governance risk but also means Steinberg faces large personal financial consequences from bad decisions, which somewhat mitigates the agency risk. Notably, he made an open-market purchase of ~53,676 shares worth ~$1M in November 2024 — during the short-seller panic — which is a high-conviction vote of confidence with personal capital at exactly the worst optics moment.

Key Lieutenants

• CFO: Chris Greiner — Joined in 2018 and has been instrumental in professionalizing Zeta's financial reporting, guiding the IPO process, and implementing a disciplined acquisition framework. Known for conservative guidance cadence that has enabled 19 consecutive beats.
• John Sculley — Co-founder and former Vice Chairman, still on the board. Brings credibility, enterprise relationships, and a consumer marketing perspective shaped by his decades at Pepsi and Apple.

Capital Allocation Track Record

Zeta's M&A record is mixed-to-positive. The acquisitions of Boomtrain (AI capabilities), Disqus (data assets), LiveIntent ($250M, 2024 — email-based identity resolution), and Marigold Enterprise ($325M, 2025 — loyalty/CRM) have each added strategic capabilities and cross-sell surface area. All were described as EBITDA-accretive in year one or shortly after, and guidance raises following completion have been credible. However, the company has made 35+ acquisitions since its founding — a pace that carries material integration risk. The Marigold integration appears ahead of schedule as of Q1 2026. The $100M buyback authorization is a new development and its execution deserves monitoring — whether Steinberg actually reduces share count or treats it as optics will be revealing.

Zeta's moat case is real but narrower than bulls claim. The company possesses genuine and defensible competitive advantages, but also operates in a sector where better-capitalized incumbents are actively investing and where AI commoditization threatens specific elements of Zeta's edge.

Moat Type 1: Proprietary Data Asset (SuperGraph)

The SuperGraph — 245M+ U.S. consumer profiles, 1 trillion signals per month — is the foundation. The critical question is sourcing quality, and the November 2024 Culper report directly challenged this by alleging that a meaningful portion of Zeta's data was gathered through "consent farms" (sham sweepstakes and job application sites). Zeta's rebuttal characterized the flagged data providers (Apptness, ArcaMax) as representing less than 3% of YTD revenue and less than 1% of data assets, with declining contributions. This response appears to have satisfied institutional investors — the stock has more than recovered. Still, the allegations have never been fully adjudicated, regulatory exposure under GDPR and FTC frameworks is real, and the FTC's 2023 action against Fluent Inc. for analogous practices is a directly relevant precedent. This is the single most important unresolved risk in the thesis.

Setting aside the data sourcing question, the scale of the SuperGraph does create genuine first-mover advantages in AI model training, predictive identity matching, and lookalike audience construction. Scale in data, unlike scale in many other assets, is self-reinforcing in a way that makes replication expensive and time-consuming.

Moat Type 2: Switching Costs (Platform Integration Depth)

Once an enterprise has migrated its customer data, campaign management, loyalty infrastructure, and identity resolution to ZMP, the switching cost is very high — multiple years of integration work, retraining, and data migration. The "One Zeta" consolidation strategy deliberately creates deep organizational embedding that would be painful and expensive to reverse. This is the clearest and least contested element of the moat.

Moat Type 3: Network Effects (Partial)

More consumer data improves targeting quality, which improves campaign ROI, which attracts more enterprise clients, which generates more first-party behavioral data — a genuine, if modest, data network effect. This is not as powerful as a true marketplace or communication network, but it is non-trivial at SuperGraph scale.

Moat Trends & Disruption Risk

The moat faces two structural threats. First, AI commoditization: as foundation models become increasingly capable of performing audience segmentation, campaign orchestration, and customer journey optimization at low marginal cost, the value of proprietary AI tooling (a portion of what Zeta sells) may compress. However, the data layer is less commoditizable than the model layer — you cannot prompt-engineer your way to 245M proprietary consumer profiles. Second, large incumbents expanding: Salesforce (Marketing Cloud + Data Cloud), Adobe (Experience Platform), and Google (DV360 + first-party data) all have substantially greater resources and are investing aggressively in AI-native marketing capabilities. None of these have successfully dislodged Zeta's growth trajectory to date, but the competitive intensity is structurally elevated.

Total Addressable Market

The marketing technology (martech) and marketing data market is large and growing. Industry analysts have estimated the global martech TAM at approximately $500–700B, growing at mid-to-high single digits annually. The AI-native subset — which is what Zeta is positioning for — is growing faster: estimates for AI-powered marketing automation suggest a 25–35% CAGR from 2024 to 2030. Zeta's $1.79B 2026 guidance represents less than 0.3% of a $600B TAM, underscoring the runway even if market share gains are modest.

Secular Tailwinds

• Third-party cookie deprecation: Google's eventual removal of third-party tracking cookies has been a multi-year catalyst for companies with first-party data (Zeta) at the expense of those dependent on third-party audience networks. This structural shift is still playing out and benefits Zeta's SuperGraph directly.
• AI-native marketing adoption: Enterprises are aggressively consolidating marketing technology stacks onto unified, AI-capable platforms. The "vendor consolidation" trend is Zeta's primary growth mechanism in 2025–2026.
• Identity resolution scarcity: As consumer privacy frameworks tighten (GDPR, CCPA, FTC enforcement), building a lawfully-sourced, consent-based identity graph at scale becomes increasingly expensive for new entrants. Zeta's incumbency becomes more valuable, not less.

Competitors

Cyclicality

Marketing spend is moderately cyclical. In the 2020 COVID shock, Zeta's revenue continued growing (from $368M vs $306M in 2019) as brands pivoted to digital, suggesting a positive secular offset. The 2022–2023 macro softness saw growth decelerate to ~24% (2023 vs 2022) but not reverse. The business is not immune to a severe advertising recession, but its platform software component — as opposed to media buying — should demonstrate more resilience than pure-play ad networks.

Headline Multiples (Based on FY2026E)

DCF Sanity Check (Conservative)

Assumptions: Revenue grows at 25% in 2026E, decelerates to 20% by 2028, then 15% through 2031, and 8% terminal. EBITDA margins expand to 25% by 2028, consistent with management targets. FCF conversion at 65%. WACC 11%. Derived intrinsic value: approximately $28–34 per share. This implies modest upside from current prices (~$23) but not a screaming bargain. Upside/downside asymmetry is positive but requires you to believe in Zeta's 2028 target credibility.

Why the Stock Is at Its Current Price

ZETA peaked near $34 in late 2024 before the Culper short-seller report caused a 37% collapse to the low-$17 range in November 2024. The stock bottomed around $12 in early 2025 as institutional investors reassessed data sourcing practices and regulatory risk. The recovery to ~$23 has been driven by 19 consecutive beats and raises, the Marigold acquisition (accretive from day one), the Athena GA launch, and the OpenAI partnership announcement in January 2026. At $23, the stock is recovering from a short-seller-induced discounting of the data asset value, but the consent farm overhang is not fully resolved.

Margin of Safety Assessment

At $23, the stock is roughly 15–25% below the consensus analyst target of ~$28. The 52-week low of $12.10 represents 47% downside in a worst-case stress scenario (renewed regulatory action, growth deceleration). The 52-week high of $24.90 is essentially the current ceiling, meaning the entry point is less attractive than it was three to six months ago. A more favorable entry would be $17–19, where the bear-case risk/reward improves substantially and the 2028 targets still generate a compelling IRR.

Dividends

Zeta pays no dividend and is unlikely to initiate one given the growth and acquisition pipeline. This is appropriate capital allocation for a platform at this stage of development. FCF is better deployed in organic R&D and accretive M&A than distributed to shareholders at these multiples.

Share Buybacks

The $100M buyback authorization announced in Q1 2026 is notable and directionally positive — Steinberg has historically focused on growth rather than capital return. However, net share count reduction will require buybacks to substantially outpace SBC issuance, which is a high bar given ~$150–200M SBC per year versus a $100M authorization. The buyback is better read as a signal of balance sheet confidence than as a meaningful earnings-per-share driver in the near term.

M&A Track Record

Zeta has completed over 35 acquisitions since 2007. The strategic logic has been consistently coherent: acquire data assets, technology capabilities, or customer bases that are additive to the One Zeta platform thesis. LiveIntent ($250M, 2024) brought email-based identity resolution — critical as cookies disappear. Marigold Enterprise ($325M, 2025) added loyalty, CRM, and subscription revenue quality while opening EMEA. Both transactions were funded with a mix of cash and stock and were described as immediately EBITDA-accretive. The Marigold integration being "ahead of schedule" as of Q1 2026 is an encouraging data point. The risk in M&A-heavy stories is always integration distraction, and Zeta's 35+ deal history means Steinberg has had ample practice, but it also means the balance sheet has absorbed substantial goodwill.

R&D / Organic Growth

Zeta invests heavily in platform development, particularly the AI layer underpinning Athena and the SuperGraph enrichment pipeline. R&D spend as a percentage of revenue has historically been in the 15–20% range, above most traditional marketing software peers and consistent with a technology-first positioning. The Athena launch (GA in Q1 2026) and early adoption data — 7x more agentic interactions year-over-year — represent the early return on this R&D investment.

Strategic Priorities (2026–2028)

• Athena GA Rollout & Monetization: Transform Athena from a differentiation feature to a primary ARPU expansion driver. Management explicitly expects Athena to lower the barrier to multi-use-case adoption, meaning customers start with one module (e.g., email acquisition) and expand to CTV, loyalty, and analytics within the same contract.
• Marigold Integration — "One Zeta" Expansion: Cross-sell Zeta's Acquire/Grow capabilities to Marigold's predominantly Retain-focused customer base of 100+ new global enterprise brands. Offering loyalty products from Marigold to Zeta's 567 scaled customers is the primary inbound cross-sell.
• Super-Scaled Customer Expansion: Grow from 189 super-scaled customers toward 250+ by 2028 while expanding ARPU from $1.7M toward $2.5M+. The Fortune 100 telco 18x expansion story is the archetypal case of this playbook.
• AI/Data Standard Positioning: The Snowflake Open Semantic Interchange (OSI) initiative represents Zeta's ambition to define data standards for AI-powered marketing across the entire MarTech ecosystem — a long-term positioning play that could make Zeta's data format the default.
• GAAP Profitability Threshold: Management is guiding $0.02–$0.04 in GAAP EPS for FY2026, transitioning from growth-stage to profitable-growth classification, which expands the institutional buyer pool.

Management Credibility on Guidance

The 19-quarter consecutive beat-and-raise streak is one of the strongest execution track records in mid-cap software. Management has consistently maintained a deliberate 2–5% guidance buffer, meaning they build revenue estimates with internal targets that already exceed the public number. This conservative cadence has trained the analyst community to expect upside and has resulted in a positive guidance revision cycle that reinforces institutional interest. The key risk to this track record is the larger absolute revenue base: beating $1.79B by 5% requires $89M of upside versus the $13-15M upside that characterized earlier quarters.

Catalysts (Next 12–24 Months)

• Q2 2026 earnings (August 2026): Revenue guidance of $420M implies continued 36% growth; any beat will test whether the 19-quarter streak reaches 20.
• Athena adoption metrics: Management expected to provide first material KPIs on Athena's revenue contribution in H2 2026.
• Potential GAAP EPS positive announcement in Q3 or Q4 2026, which may trigger index inclusion (Russell, potentially S&P 600).
• OpenAI relationship deepening: The expanded advertising operations partnership could evolve into a meaningful revenue line if OpenAI's own advertising business scales.
• Any regulatory action or ruling related to Culper allegations — this is a binary risk event in either direction.

Is AI a Threat to This Business?

Partially, yes. As large language models commoditize certain analytical and creative functions that Zeta's platform performs (audience segmentation recommendations, copy generation, campaign optimization suggestions), the AI-as-feature moat erodes. However, the AI threat is strongly offset by AI as amplifier. The core of Zeta's value — its proprietary consumer data — becomes more valuable, not less, in an AI world, because better models trained on better data produce better outcomes. Commodity AI without proprietary data is a tool without material. Zeta has the data; it just needs the right tools on top.

Athena — Zeta's Agentic AI Platform

Launched in general availability to all enterprise customers in Q1 2026, Athena is Zeta's AI marketing agent. It converts a company's marketing data into predictive answers, identifies opportunities before human analysts, and executes actions directly within marketing workflows. Early adoption metrics are genuinely impressive: 7x more agentic interactions year-over-year, and one cited Fortune 100 telco expanding to 18x their prior Zeta spend after Athena deployment. The OpenAI strategic partnership (announced January 2026 at CES) integrates OpenAI's most advanced models into Athena's reasoning layer, ensuring Athena tracks the frontier of AI capability without requiring Zeta to develop foundation models internally — a sensible resource allocation decision.

SuperGraph as an AI Data Asset

The SuperGraph's 245 million U.S. consumer profiles processed through 1 trillion monthly signals is a genuinely scarce training and inference asset. The expansion of the OpenAI partnership to include assisting with OpenAI's own advertising operations is particularly strategic: it means Zeta's data is not just being used by Zeta customers but is becoming infrastructure for OpenAI's commercial growth. This is potentially a revenue line in its own right.

Snowflake Open Semantic Interchange (OSI)

Zeta's membership in the OSI initiative — a Snowflake-led effort to create universal data standards for AI-powered marketing — is a sophisticated long-term positioning move. If the OSI standard is adopted broadly, Zeta's data schema becomes the reference point for AI marketing data across the industry, creating a de facto standard that benefits existing and future data products. The risk is that this standard becomes a commodity rather than a differentiator, but first-mover positioning in setting standards often creates lasting advantages.

Insider Ownership

David Steinberg owns approximately 10% of economic equity (3.5M shares, ~$56M) and controls more than 50% of total votes via Class B shares. His most notable open-market transaction was a $1M purchase of ~53,676 shares in November 2024 — during the height of the Culper crisis when the stock had just dropped 37%. This is the strongest possible insider confidence signal: buying with personal capital when sentiment is maximally negative. There have been no significant open-market insider sells in the disclosed period, though SBC vesting and related transactions represent ongoing stock flow.

Institutional Ownership

As of Q1 2025, institutional investors hold approximately 75% of outstanding Class A shares. The shift from private-equity dominated cap table at IPO to institutional-led ownership reflects growing buy-side confidence. The investor base includes a mix of growth-oriented technology funds and emerging long-only fundamental investors who have built positions post-Culper reset.

Analyst Consensus

The spread between the low ($22) and high ($44) targets is exceptionally wide, reflecting genuine uncertainty about whether Zeta's data quality concerns are resolved and whether the 2028 targets are credible. B. Riley maintained Buy and raised to $30 post-Q1. KeyBanc upgraded to Overweight at $22 (now below current price), citing early innings of AI-driven enterprise marketing. The uniformity of Buy ratings with no Sells is either a sign of genuine conviction or analyst conflict of interest — the lack of any bear-side coverage warrants slight skepticism.

Short Interest

Short interest spiked dramatically following the Culper report and has been declining as the stock recovered. The residual short position represents both residual bears who remain unconvinced the data allegations were defused and traders hedging the consent farm regulatory exposure. Continued short covering is a potential technical tailwind.

Asymmetry Assessment

At $23, the bull/bear asymmetry is approximately 2.5–3:1 upside to downside ($10 median upside vs. $4 median downside to the respective scenario midpoints, measured from current price). This clears the minimum 2:1 threshold for a worthwhile risk/reward, but only marginally so. The asymmetry was substantially more favorable at $12–15, where the risk/reward approached 5:1. The stock has meaningfully recovered from its trough and the margin of safety has compressed accordingly. The base case still generates a compelling ~19% annualized IRR, which is above a reasonable market hurdle. But the unresolved consent farm litigation means the bear tail is wider than usual for a software company at this growth rate.