Cardlytics, Inc. (NASDAQ: CDLX) — A Melting Ice Cube Wearing a Debt Overcoat

Published: June 7, 2026 (data through Q1 FY2026, 10-Q filed 2026-05-07; FY2025 10-K filed 2026-03-04) Sector: Communication Services — Advertising / Marketing Technology (financial-channel “card-linked offers”)

Standing disclaimer for the analytical body (Sections 1–15): this article takes no investment recommendation and sets no price target. It discusses valuation only as embedded expectations and scenarios. The sole exception is the clearly-labeled “Author’s Take” block immediately below, which is a subjective view.

⚡ Author’s Take

This block is the author’s own subjective opinion. It is general information, not investment advice. The analysis in Sections 1–15 below carries no recommendation and no price target.

Verdict: AVOID as an investment — speculative levered option only. Tag: “A melting ice cube wearing a debt overcoat.”

Directional view (post-1:10 reverse split; ~5.8M shares; stock ~$9 post-split / ~$0.91 pre-split): I would not own the equity here, and I would not short it either — the borrow is hard, the float tiny, and the wipeout is partly priced. This is not a value stock despite a 0.23x price/sales optic; it is a deeply out-of-the-money call option on a ~$197M enterprise value clearing ~$144M of net debt. My honest “fair value” for the equity is a distribution, not a point: roughly $0 in the bear case (genuine wipeout), ~$0–$9 in a stabilization base case, and ~$16–$28 only if revenue re-accelerates to $160M+ at recovering margins. At ~$9 the market is already paying the high end of the base case and embedding the recovery multiple — so the asymmetry is against you: you are underwriting the bull just to break even.

What is the market getting right? That a 0.8x EV/sales business with positive operating cash flow and no 2026–27 maturity is not an imminent bankruptcy. What is it getting wrong? The discount looks like value but is a value trap: revenue is shrinking ~13%/yr and accelerating (Q1’26 −39% YoY), the single most important customer (Bank of America) left and built its own engine, the largest remaining customer (Chase) bought a direct competitor (Figg), management destroyed ~$714M of acquired capital and is fire-selling the centerpiece (Bridg) at ~5¢ on the dollar, and stockholders’ equity is already negative. The framing is a falling knife with a financing cliff — not a quality-compounder-at-a-price and not a clean special situation. Conviction: medium-high on “avoid as a core holding”; low on the binary outcome (turnarounds in rented-distribution ad-tech occasionally work). Flips bullish if: the post-Bridg/post-BofA revenue line visibly stabilizes for two consecutive quarters AND a new anchor FI signs on improving (not dilutive) economics, de-risking the 2029 refi. Flips to zero if: another top-three FI partner (Wells Fargo, Amex, U.S. Bank) declines to renew, or revenue stays in double-digit decline into 2027 — at which point the 2028 term loan and 2029 converts force equitization.

1. Executive Summary

Cardlytics operates a financial-media network embedded inside banks’ digital channels: it ingests anonymized cardholder purchase data from financial-institution (“FI”) partners (historically Chase, Bank of America, Wells Fargo, American Express, U.S. Bank), renders card-linked cash-back offers inside online/mobile banking, and sells closed-loop, purchase-verified advertising to marketers. In theory this is an attractive, hard-to-replicate asset: privileged access to real purchase data at national scale. In practice, the company has spent five years demonstrating the opposite — that the scarce assets (the data and the customer relationship) belong to the banks, not to Cardlytics, and that the independent middleman captures a thin, shrinking, economically negative slice.

The numbers are stark. GAAP revenue has fallen from $309.2M (FY23) → $278.3M (FY24) → $233.3M (FY25), with Q1 FY2026 down −39% year-over-year to $34.3M. The company has never produced GAAP operating profit; cumulative goodwill/intangible impairments of roughly $714M since 2022 have written off nearly the entire ~$856M spent on the 2021 Dosh and Bridg acquisitions; the accumulated deficit is $1.404 billion; and stockholders’ equity is now negative (−$6.5M at YE25, −$4.0M at Q1’26). “Adjusted EBITDA” of $10.1M in FY25 is a stock-comp mirage — the $28.1M SBC add-back alone exceeds it. Free cash flow has been negative every year.

Two events crystallize the broken thesis. First, Bank of America — a top-three partner — gave notice of non-renewal in April 2025, ended the relationship in February 2026, and migrated “BankAmeriDeals” to its own in-house engine. Second, JPMorgan Chase, Cardlytics’ single largest partner, acquired Figg (the only other bank-direct card-linked-offer integrator) and now powers its own “Chase Media Solutions.” The company’s anchor channel is being insourced by the very partners that supposedly constitute its moat. Management is simultaneously divesting Bridg to a PAR Technology affiliate for ~$27.5M in stock — a ~5-cents-on-the-dollar exit from a $578.9M purchase.

Financially the company is a levered stub: ~$204M of debt (a $172.5M 4.25% convertible due 2029 with an $18 pre-split conversion price, plus a ~$40M term loan due 2028) against ~$60M of cash and ~$53M of equity market value. There are no 2026–27 maturities, so bankruptcy is not imminent; but the equity is a deeply out-of-the-money option on the enterprise clearing its debt, and the 2028–29 maturity wall must be refinanced by a sub-scale, shrinking ad-tech business. A 1-for-10 reverse split (effective June 5, 2026) cured a Nasdaq sub-$1 bid-price deficiency cosmetically; it changes nothing economically and is itself a distress marker.

Bottom line: No durable competitive advantage, a structurally disadvantaged position in an otherwise-growing market, serial value-destructive capital allocation, declining revenue, negative equity, and a refinancing cliff. The cheap-looking sales multiple is a value trap. This article takes no position (see Author’s Take for the subjective view).

2. Business Overview

2.1 What Cardlytics does

Cardlytics runs two platforms, and is in the process of becoming one:

• The Cardlytics platform (the core, ~85%+ of remaining revenue): a “purchase-intelligence” advertising network operated inside the digital banking properties of FI partners. The bank supplies (a) anonymized, tokenized cardholder transaction data and (b) the advertising real estate — the offers carousel in online and mobile banking. Cardlytics sells campaigns to marketers (“advertisers”), who fund cash-back rewards (“Consumer Incentives”) that appear to cardholders as activated offers (e.g., “spend $50 at Retailer X, get $10 back”). When the cardholder transacts, Cardlytics measures the purchase against the ad exposure — closed-loop, purchase-verified attribution that walled-garden digital advertising (Google, Meta) cannot natively replicate for offline spend.
• Bridg (held-for-sale; being divested): a SKU-level customer-data platform acquired in 2021 to extend into retail-media / CPG measurement. A January 2026 asset-purchase agreement sells substantially all Bridg assets to a PAR Technology affiliate. Post-close, Cardlytics is a pure-play FI-channel card-linked-offers business. (Dosh, a consumer cash-back app acquired in 2021, was folded into the core platform.)

2.2 How the money flows — the “thin middle slice”

The revenue architecture is the single most important thing to understand, because it explains why scale has not produced profit. The FY2025 funnel (FACT, FY25 10-K):

Of every dollar of billings, cardholders take ~39% and banks take ~27%; Cardlytics keeps gross profit of ~27% of billings before any operating cost. Critically, Partner Share flows through cost of revenue, not as contra-revenue (Cardlytics books revenue as principal), so GAAP “revenue” overstates the economics actually retained. The party with pricing power in this chain is the bank: the FY2025 Partner Share + Consumer Incentives together consumed ~66% of billings, and that share has trended against Cardlytics over time — it needed a June-2023 amendment with Chase merely to retain more economics, and unit monetization still fell.

2.3 Operating metrics — the inversion

Cardlytics historically disclosed Monthly Active Users (MAUs) and Average Revenue Per User (ARPU). As of January 1, 2025 it stopped reporting MAU and ARPU, stating it “does not receive equivalent user data from newer bank partners” — a transparency downgrade that itself signals newer partners share less data on worse terms. The replacement metrics tell the story bluntly: Monthly Qualified Users (MQUs) rose ~18% to 224 million in FY25, while Average Cardlytics Revenue Per User (ACPU) fell ~25% to $0.50 (from $0.67). More users, materially worse monetization — the opposite of what a network-effect or scale business should show.

2.4 Revenue model quality

Revenue is largely non-recurring at the campaign level: marketers buy via insertion orders “with no commitment beyond the campaign.” There is no contractual subscription backlog; the “recurring” element is behavioral (large marketers re-up) rather than contractual. FI-partner contracts are multi-year but terminable on short notice (as little as 90 days in some cases), and certain FIs hold the ad-server source code in escrow, releasable on defined events so the bank can self-operate — a structural concession that all but invites insourcing (and is exactly what Bank of America did).

Verdict (Business Overview): A genuinely clever measurement product sitting in a structurally weak commercial position — Cardlytics rents both its data and its distribution from counterparties who can, and increasingly do, take the function in-house. Revenue quality is low (campaign-based, no contractual recurring base, terminable channel).

3. Industry Dynamics

3.1 Market structure and size

The card-linked-offers (CLO) / financial-media category is real and growing. Third-party estimates put the CLO market at roughly $9.2–9.4B in 2025 growing toward ~$27–29B by 2034 (~13% CAGR), nested inside a broader retail-media complex heading toward ~$100B+ by 2028 (INTERPRETATION, third-party market research; treat magnitudes as directional). The structural tailwind is real: the deprecation of third-party cookies and the privacy crackdown on identity-based targeting make purchase-verified, first-party-data advertising more valuable, and offline purchase data (where banks see everything) is the scarce input.

3.2 Who owns the profit pool

Here the competitive lens is decisive: the profit pool sits with the owners of the scarce assets — the banks and the card networks — not with the independent enabler. The competitive set has converged on Cardlytics from three directions:

• Card networks integrating forward into media: Mastercard Commerce Media (launched October 2025), Visa Offers Platform, and American Express Offers are simultaneously infrastructure and competitors. They sit on transaction data at a scale Cardlytics cannot match and need no bank intermediary.
• Banks insourcing: Chase Media Solutions (powered by Figg, which Chase acquired), Bank of America’s in-housed BankAmeriDeals, plus PayPal and Klarna building their own rewards/offers engines. Every bank that insources is both a lost customer and a new competitor.
• API-first challengers: Kard, PayLead, Array, Prizeout, Affinity Solutions, Empyr/Augeo, Rewards Network — a fragmented field competing on integration ease and economics.

3.3 Capital-cycle read

This is not a recovering industry with capital exiting (the classic supply-side “buy” setup). It is a capital-inflow boom: ~13% demand growth is attracting deep-pocketed entrants who already own the scarce inputs. The structural warning applies directly — market growth is the enemy of an economies-of-scale advantage, because growth lets new entrants reach minimum efficient scale. A rising tide here lifts the networks and the banks; it does not protect a sub-scale, capital-starved independent whose “scale” is rented from the very entrants flooding in.

Verdict (Industry Dynamics): A structurally attractive industry to be a bank or a card network in, and a structurally bad one to be Cardlytics in. The category grows, but the independent FI-channel middleman is being disintermediated from both ends.

4. Competitive Position

4.1 The claimed moat

Cardlytics’ self-description is essentially a scale + exclusive-data-access moat: “the only company that enables marketing through FI channels at scale,” with a national footprint of bank partners and privileged purchase data creating a two-sided network (more banks → more data/reach → more marketers → more value to banks). On paper this maps to the strongest advantage type: economies of scale reinforced by customer captivity.

4.2 Pressure-testing — the moat is failing in real time

Run each advantage type against the evidence:

• Supply / cost advantage — absent. Infrastructure is commodity cloud (AWS); there is no cost edge versus a Mastercard or a well-funded startup. The escrowed ad-server source code means even the technology is contractually transferable to partners.
• Demand / captivity (customers) — weak and deteriorating. Marketers buy per-campaign with no commitment and multi-home across channels. FI partners face low switching costs and short termination windows; the source-code escrow lowers them further. The proof: Bank of America, a top-three partner, gave non-renewal notice (April 2025), exited (February 2026), and stood up its own engine. Customer captivity that evaporates on 90 days’ notice is not captivity.
• Economies of scale + captivity (the only potentially durable one) — fails the captivity leg. The “scale” — the FI footprint — is rented, not owned. The captive party (the bank) can leave and take the function in-house, which is precisely the observed behavior. A scale advantage you do not control is not a moat; it is a supplier relationship.

4.3 The killer facts

Two facts falsify the network-effect thesis outright:

1. The largest partner owns a competitor. JPMorgan Chase — Cardlytics’ biggest FI relationship — acquired Figg, the only other bank-direct CLO integrator, and now runs Chase Media Solutions. Cardlytics’ anchor distribution is controlled by an entity now building the competing product.
2. Unit economics invert with scale. A real network effect shows rising monetization as the network grows. Cardlytics shows MQUs +18% while ACPU −25% — monetization falls as the user base grows. The empirical signature is the opposite of a network effect.

Both empirical moat tests reject a moat here: market-share stability (lost a top-three partner; revenue shrinking double-digits) and ROIC (chronic GAAP losses, negative FCF, ~$714M of impairments writing off the entire growth-by-acquisition strategy).

4.4 The “if-it-disappeared” test

A simple standard: if a claimed moat can’t be tied to a financial outcome that would deteriorate without it, it isn’t a moat. Cardlytics’ purported data/scale advantage should manifest as pricing power and rising ARPU. It manifests as falling ACPU, shrinking revenue, and a partner exodus — i.e., there is no financial outcome being protected. This is the same archetype seen at Digital Turbine (APPS): an ad-tech middleman on a “rented choke point,” monetizing access it does not own. Cardlytics is the worse version — it is losing its anchor channels, not merely failing to grow them.

Verdict (Competitive Position): No durable competitive advantage. A clever product in a structurally subordinate commercial position, being disintermediated by the very partners and networks that define its market.

5. Growth History and Forward Opportunities

5.1 The historical record

The growth that existed was acquired, not organic — the 2021 Dosh/Bridg deals inflated reported revenue, then unwound. Organic core-platform billings have been falling since 2023. The FY25 decline decomposes to −$98.8M from existing marketers (spending less / churning) partially offset by +$39.9M from new marketers — i.e., the installed base is shrinking faster than new logos can replace it. Q1’26’s −39% reflects the first clean quarter without Bank of America’s contribution and the Bridg reclassification.

5.2 Forward opportunities (the bull’s list) — and the discount on each

• Privacy tailwind / cookie deprecation raising the value of purchase-verified, first-party data. Discount: the banks and networks capture this, not the rented middleman.
• New FI partners to replace lost reach. Discount: the “newest FI partner” is diluting ACPU — Cardlytics appears to be buying distribution at worse economics, which grows users but not profit.
• Retail-media / CPG expansion. Discount: the vehicle for this (Bridg) is being sold for scrap; the strategy is being abandoned, not pursued.
• New ad-decisioning engine / self-service tooling to improve fill and yield. Discount: real but unproven; it must reverse a structural decline, and it competes with banks’ own engines.
• Operating-cost discipline producing positive operating cash flow. Discount: this is margin from shrinkage (S&M −25%+, R&D −25%+), not growth; you cannot cut your way to a moat.

Verdict (Growth): Low-quality, negative growth. The historical “growth” was acquisition-driven and has reversed; forward opportunities are real for the category but are being captured by better-positioned players, and Cardlytics’ own actions (Bridg divestiture, ACPU-dilutive new partners) cut against the bull list.

6. Financial Quality

6.1 Multi-year financials (USD thousands; FY25 10-K, FY23 10-K, Q1’26 10-Q, SEC EDGAR XBRL)

6.2 Earnings quality — three red flags

1. “Adjusted EBITDA” is a stock-comp mirage. FY25 Adjusted EBITDA of $10.1M is smaller than the $28.1M of stock-based compensation added back to compute it. SBC is a real, recurring, ~12%-of-revenue cost that dilutes shareholders every year; a profit metric that exists only by ignoring it does not describe an economically profitable business. Adjusted net loss was still −$17.3M in FY25.
2. Operating cash flow is flattered by capitalized software. FY25’s first-ever positive OCF (+$9.3M) coexists with negative free cash flow (−$6.5M) because $15–19M/yr of software development is capitalized rather than expensed. The business is not self-funding.
3. Serial impairments are not “one-time.” Goodwill/intangible writedowns recurred every year from FY22 to FY25 (~$453M, $71M, $132M, $59M). When impairment is annual, it is a verdict on the acquisition strategy, not a non-recurring item — and the run-rate “adjusted” profitability that excludes them flatters a business that destroyed capital on a massive scale.

6.3 Margins and operating leverage — the wrong kind

Gross margin has crept up (~38% → ~45%), but not from scale economics — it rose because revenue fell slower than partner/incentive costs in some periods and because cost cuts hit the P&L. Operating margin is negative every year. There is no evidence that economics improve with scale; the entire history shows the reverse — more users, lower ACPU; more revenue (2021–23, acquired), larger losses.

6.4 Balance sheet and solvency (the central risk)

• Stockholders’ equity is negative (−$6.5M YE25; −$4.0M Q1’26). Accumulated deficit of $1.404B has consumed essentially all paid-in capital.
• Debt: 2024 Convertible Notes — $172.5M, 4.25%, due 2029, conversion price $18.00 pre-split (≈$180 post-split; deeply out-of-the-money vs. ~$9 post-split), issued April 2024 to retire the 2020 Notes at a discount ($13M extinguishment gain). Plus a ~$40M term loan/credit facility due 2028 (Banc of California; $25M minimum-cash covenant). 2020 Notes residual matured September 2025.
• Liquidity: ~$60M cash/securities at Q1’26. With near-breakeven OCF and modest FCF burn, this is several years of runway if revenue stabilizes — but the $25M minimum-cash covenant limits usable liquidity, and there is no buffer for a sharp revenue leg-down.
• Maturity wall: ~$40M in 2028 and $172.5M in 2029. Nothing in 2026–27 — so no imminent default — but ~$212M comes due into a business generating ~$10M of (SBC-flattered) Adjusted EBITDA. Cash redemption is impossible; the converts must be refinanced or equitized. Refinancing a sub-scale, shrinking ad-tech name on acceptable terms is the central solvency question.
• No going-concern / substantial-doubt language appears in the FY25 10-K or Q1’26 10-Q (FACT) — bankruptcy is not imminent, but negative equity plus the 2028–29 wall make the equity a levered option.

6.5 ROIC / ROE

Not meaningful — persistent losses and negative equity make returns-on-capital metrics undefined or misleadingly signed. The economically honest statement is that return on invested capital has been deeply negative: ~$856M deployed into acquisitions returned ~$714M of impairments and a shrinking core.

Verdict (Financial Quality): Poor. Economics do not improve with scale; “profitability” exists only on an SBC-flattered non-GAAP basis; equity is negative; the business is not self-funding; and a refinancing cliff looms. Earnings quality is low (SBC add-backs, capitalized-software-propped OCF, serial impairments, a mid-stream revenue-presentation change).

7. Capital Allocation

7.1 The record — serial value destruction

This is among the clearer cases of capital destruction in the small-cap universe. At the 2021 ad-tech/SPAC peak, prior leadership paid roughly $856M combined for Dosh and Bridg, then impaired ~$714M of it within four years — near-total destruction of deployed capital. The Bridg divestiture removes any doubt about whether this was a temporary mark: a business bought for $578.9M is being sold for ~$27.5M in PAR Technology stock — ~5 cents on the dollar. That is scrap value, not a value-recovering exit.

7.2 M&A scorecard

7.3 Financing decisions — the one competent axis

Liability management has been more competent than asset allocation. The April 2024 refinancing of the $230M 1.00% 2020 Notes via the $172.5M 4.25% 2029 converts retired $183.9M of principal at a discount (+$13M extinguishment gain) and pushed the maturity wall from 2025 to 2028–29. Operating costs were cut hard (S&M, R&D, G&A each −25%+ over two years), producing the first positive operating cash flow in FY25. These are competent damage-control moves at the trough — necessary, but not value creation.

Against that, the March 2024 ATM sold 3.91M shares at $12.80 ($48.3M) into a falling stock, and SBC of ~$28–45M/yr continues to dilute against a now-negative book. The capital-cycle read is textbook value destruction: raised cheap equity/debt and overpaid for acquisitions at the top of the cycle; impaired and distressed at the bottom — negative equity, a sub-$1 (pre-split) stock, a Nasdaq deficiency, and a fire sale of the trophy asset.

7.4 Incentive alignment

Short-term cash bonus and 2025 PSU metrics are Adjusted Contribution and Adjusted EBITDA (budget-aligned). There is no ROIC or return-on-capital metric — so the ~$714M of M&A value destruction never directly hit management’s scorecard. That is the central alignment flaw: the comp plan rewards operating-axis budget achievement while being blind to the capital decisions that destroyed the equity. To its credit, the plan did pay down with poor results — FY25 corporate achievement was 30%, CEO Gupta was paid $166,100 against a $550,000 target, and the departed CFO received $0. But the committee then granted 1,000,000 fresh RSUs to Gupta, repricing incentives at the bottom and adding dilution — a pattern that rewards staying through value destruction rather than penalizing it.

7.5 Insider behavior

Over 36 months (123 Form 4s): grants/settlements dominate (A=6, M=26), with only 4 sales — all sell-to-cover for RSU tax withholding (per Form 4 footnotes), not discretionary selling. The lone genuine conviction signal is new CFO David Evans’s open-market purchase of 200,000 shares (~$0.64, ~$128k) in May 2026 — modest, fresh-into-the-role, and the only “code P” buy of size. Aggregate director/officer ownership is thin at 5.9% pre-split. Net insider posture is mildly positive, but the ownership stake is too small to be a strong alignment signal.

Verdict (Capital Allocation): Failing grade. ~$856M deployed, ~$714M impaired, the centerpiece fire-sold at ~5¢/$, negative equity, a reverse split to dodge delisting. Competent recent debt-maturity management and a small CFO buy are the only offsets, and they are damage control, not value creation.

8. Changes and Headwinds — Last Two Years

A roughly chronological, 8-K-driven timeline of the events that reshaped the thesis:

The throughline: the partner base eroded (BofA gone; Chase conflicted via Figg), the strategy reversed (Bridg sold), leadership churned (new CEO and CFO), and the capital structure went distressed (negative equity, reverse split, Nasdaq deficiency) — all within ~24 months.

Verdict (Changes/Headwinds): Decisively thesis-weakening. The single most important change — losing Bank of America and watching Chase buy a competitor — is structural, not cyclical.

9. Risk Analysis

9.1 Risk matrix

9.2 Catastrophic-loss / total-loss assessment

The realistic path to a total loss of equity is not a sudden bankruptcy but a slow one: revenue continues to fall, the enterprise value never recovers above ~$144M net debt, and at the 2028–29 maturities the company refinances on terms that equitize the converts or raises rescue equity at a price that wipes out current holders. Given negative book equity today, this tail is not remote — it is a live base-rate outcome for over-levered, shrinking businesses. The offsetting fact is the absence of near-term maturities and positive operating cash flow, which buys time for a turnaround to take hold.

10. Valuation Discussion (Embedded Expectations)

No price target. No recommendation. Valuation is framed as embedded expectations and scenarios.

10.1 Current multiples

EV ≈ $197M (~$53M equity + ~$204M debt − ~$60M cash); ~5.81M shares post-split; net debt ~$144M.

10.2 Peer comparison

Median peer P/S ~1.9x vs. CDLX 0.84x. The discount is a value trap, not value. CDLX is the only name shrinking double-digits, has no durable moat (lost BofA; Chase owns Figg), and ~73% of its EV is debt. The comparable Digital Turbine (APPS) setup points to an instructive conclusion: a discount to peers is justified, not an opportunity, “unless both the inflection and the deleveraging materialize.” CDLX is that setup but worse: APPS had +20% revenue, positive FCF, and was deleveraging; CDLX has −16%/−39% revenue, negative FCF, and is losing anchor partners.

10.3 Embedded expectations

At 0.84x EV/sales the market is not pricing a cheap-sales story — it is underwriting a recovery. To justify EV ~$197M, the business must roughly stabilize near a $130M+ steady-state revenue at ~1.5x EV/sales (197 ÷ 130 ≈ 1.52x) — a multiple above where stable-but-shrinking CRTO trades (0.46x). More importantly, the equity (~$53M) is a deeply out-of-the-money levered call on enterprise value clearing ~$144M of net debt. On an EBITDA basis almost every plausible scenario produces negative equity value; the only way the equity is worth its current price is a sales-multiple re-rating predicated on stabilized or growing revenue. In other words, the current price already embeds the bull case’s recovery multiple — the market is not giving you the turnaround for free.

10.4 Scenario analysis

Post-Bridg / post-BofA run-rate; ~2-year horizon; 5.81M post-split shares; net debt ~$144M. Assumptions explicit; all values approximate.

An EBITDA-multiple cross-check is harsher still: a base-case 5x EBITDA → EV ~$52M → equity ≈ −$92M. The equity is positive only under a sales-multiple re-rating on stabilized/growing revenue. The bear is a genuine equity-wipeout case: at ~$100M revenue and collapsed margins, EV sits below net debt and the 2028 term loan plus 2029 converts force distressed refinancing or equitization — and with negative tangible equity, the secured term loan ranks ahead of the converts, which rank ahead of common.

10.5 Valuation overhangs

• Reverse split (1:10, June 5, 2026): a Nasdaq-bid-compliance maneuver — cosmetic, value-neutral, and itself a distress signal. ~$9 post-split has little cushion if the decline persists.
• Debt wall: ~$212M maturing 2028–29 into a ~$10M-Adj-EBITDA, shrinking business. Cash redemption is impossible; refinancing a sub-scale, declining ad-tech name is the central bear risk.
• Own-history valuation percentiles were unavailable for CDLX — own-history cross-check could not be performed (flagged).

11. Variant Perception

Consensus view. A broken, over-levered micro-cap left for dead — sub-$1 (pre-split), reverse-split, shrinking, losing partners. The Street has largely stopped covering it; what remains is a distressed/special-situation audience.

The strongest bull case. Cardlytics sits on a genuinely scarce asset (purchase-verified, first-party bank data) in a category with a real privacy tailwind. The cost base has been cut to near-breakeven cash flow; there are no 2026–27 maturities; new leadership (ex-Google CEO, new CFO who bought stock) is refocusing on the core FI-channel business after shedding the Bridg distraction. If two or three new FI partners sign and the new ad-decisioning engine lifts yield, revenue could stabilize and re-accelerate toward $160M+, at which point a modest sales-multiple re-rating clears the debt and the deeply out-of-the-money equity multiplies. It is a small-cap, asymmetric, levered call.

The strongest bear case. The “scarce asset” belongs to the banks, who are insourcing it (BofA done; Chase owns Figg) and out-scaled by the card networks (Mastercard/Visa/Amex). Revenue is in accelerating decline (−39% in Q1’26), unit economics are inverting (ACPU −25%), capital allocation destroyed ~$714M, equity is already negative, and ~$212M comes due into a ~$10M-EBITDA business. The reverse split and Nasdaq deficiency are distress markers, not turning points. The most likely outcome is continued erosion into a 2028–29 refinancing that equitizes the converts and wipes out common.

The 3–5 assumptions that decide it:

1. Do remaining anchor FI partners (Wells Fargo, Amex, U.S. Bank) renew on non-deteriorating terms? (Bear: no / on worse terms; Bull: yes / improving.)
2. Does revenue stabilize, or keep falling double-digits? (The single most important variable.)
3. Can the 2029 converts be refinanced without equitizing common? (Depends entirely on #2.)
4. Does the new ad-decisioning engine lift ACPU, reversing the monetization inversion?
5. Is the privacy tailwind captured by Cardlytics or by the banks/networks?

Falsification tests. Bull is falsified if any additional top-three FI partner declines to renew, or if revenue remains in double-digit decline through 2027. Bear is falsified if revenue posts two consecutive quarters of stabilization/growth and a new anchor FI signs on improving economics — which would materially de-risk the 2029 refi.

12. Fact vs. Interpretation

13. Open Questions

1. Did Wells Fargo, American Express, and U.S. Bank renew — and on improving or deteriorating economics? Wells Fargo is now the load-bearing partner; any further non-renewal is thesis-ending.
2. Who is the unnamed “newest FI partner” diluting ACPU, and on what terms? Is Cardlytics buying distribution at a loss to mask lost billings?
3. What is the pro-forma, post-Bridg/post-BofA steady-state revenue and Adjusted Contribution run-rate? The clean continuing-operations base is the key input to the refi math.
4. Are there covenants on the term loan beyond the $25M minimum-cash threshold (e.g., Adjusted-EBITDA or liquidity tests) that could trip in a downturn?
5. What is the realistic 2029 refinancing path given negative equity and deep-OTM converts — extension, new secured debt, or equitization?
6. Did the FY25 figures fully reclassify Bridg to discontinued operations, and if so what is continuing-ops revenue precisely? (Presentation changed mid-stream; reconcile before any run-rate conclusion.)

14. What Must Be True

For the bull case to work (all roughly required):

• Remaining anchor FI partners renew on non-deteriorating terms, and at least one new anchor FI signs.
• Revenue stabilizes within ~2 quarters and re-accelerates toward $160M+ at recovering margins.
• The new ad-decisioning engine lifts ACPU, reversing the monetization inversion.
• The 2029 converts are refinanced without equitizing common.
• Falsification: any additional top-three partner non-renewal, OR revenue still in double-digit decline through 2027.

For the bear case to work (any one largely sufficient):

• Another major FI insources or fails to renew.
• Revenue continues double-digit decline, keeping EV below net debt.
• The 2028–29 maturities force distressed refinancing or equitization.
• Falsification: two consecutive quarters of revenue stabilization/growth plus a new anchor FI on improving economics.

The asymmetry is unfavorable to the equity: the bull requires several things to go right in sequence; the bear requires only one of several independent failures.

15. Source Appendix

See the Source Appendix below for the full, dated source list. Primary sources: Cardlytics FY2023/FY2024/FY2025 Forms 10-K; Q1 FY2026 Form 10-Q (filed 2026-05-07); 2026 DEF 14A proxy; Forms 8-K (CEO/CFO changes, BofA non-renewal, Bridg APA, convert issuance, reverse split, Nasdaq deficiency); Forms 3/4/5; SEC EDGAR XBRL company facts (CIK 1666071). Secondary: third-party CLO/retail-media market estimates and peer market data. All accessed June 7, 2026.

Appendix A — Diligence Questionnaire

Cardlytics, Inc. (NASDAQ: CDLX). Supplemental to the analysis above. Fact / Interpretation / Assumption labels applied where material. Data through Q1 FY2026.

General

What thoughtful questions have investors asked about this company? The distressed/special-situation audience that remains focuses on: (1) Is there a clean, post-Bridg, post-Bank-of-America revenue run-rate, and has it stopped falling? (2) Can the 2029 convertibles ($172.5M) be refinanced without wiping out the common? (3) Do the remaining anchor FI partners (Wells Fargo, American Express, U.S. Bank) renew — and on what economics, given Bank of America left? (4) Is the “exclusive bank-data” moat real, or do the banks and card networks own the asset? (5) Is positive operating cash flow durable, or an artifact of cost cuts and capitalized software? (Interpretation, from the evidence base.)

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Neither in the usual sense — there are no GAAP earnings (persistent losses). The business is in secular, not cyclical, decline: revenue −10% (FY24), −16% (FY25), −39% (Q1’26). The deterioration is driven by structural partner loss and disintermediation, not the ad cycle. (Fact/Interpretation.)

Driven by the external environment or internal actions? Both, adversely: external (Bank of America insourcing, Chase owning Figg, networks entering) and internal (the failed Dosh/Bridg strategy, ACPU-dilutive new partners). The recent cash-flow improvement is internal (cost cuts). (Interpretation.)

How stable are revenues? Low stability. Revenue is campaign-based with no commitment beyond the campaign; FI contracts are multi-year but terminable on short notice (as little as 90 days), with ad-server source code in escrow. FY25 saw −$98.8M of existing-marketer spend. (Fact.)

Outlook for products/services? The core card-linked-offers product has a real privacy tailwind but a weak commercial position; Bridg (retail media) is being sold. (Interpretation.)

How big will this market be? The CLO market is ~$9.2–9.4B (2025) growing ~13%/yr toward ~$27–29B by 2034; retail media broadly toward $100B+ — growing, global, but the profit pool accrues to banks and networks, not the independent middleman. (Interpretation, third-party estimates.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — card networks (Mastercard Commerce Media Oct 2025, Visa, Amex), banks insourcing (Chase, BofA, PayPal, Klarna), and API-first startups are all converging. (Fact/Interpretation.)

How profitable is the business (ROIC, ROE)? Not meaningful / deeply negative. Persistent GAAP losses, negative stockholders’ equity, ~$714M impaired against ~$856M deployed. ROIC has been substantially negative. (Fact.)

How profitable is the industry; how many competitors; barriers to entry? The category is profitable for asset owners (banks, networks). Barriers to entry for an independent are low-to-moderate and falling (commodity cloud, escrowed source code, banks willing to build). Many competitors. (Interpretation.) No genuine supply, demand, or scale-plus-captivity advantage survives scrutiny.

Can the business be easily understood? Yes — a financial-channel ad network with a transparent (if unfavorable) billings→revenue→gross-profit waterfall. (Fact.)

Can it be undermined by foreign low-cost labor? Not materially relevant; the threat is domestic insourcing and network competition, not labor arbitrage. (Interpretation.)

Do brands matter? Modestly — Cardlytics has scale recognition among marketers, but the consumer never sees the brand (it appears as the bank’s offers), and marketers buy on performance, not brand. (Interpretation.)

Nature of competition / customers’ switching costs? Two customer types: FI partners (low switching cost — proven by BofA’s exit; source code in escrow) and marketers (no commitment, multi-home). Switching costs are weak on both sides. (Fact/Interpretation.)

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The privileged data-access relationships are not capitalized — but they are not owned and are eroding, so this is a negative off-balance-sheet item, not hidden value. (Interpretation.)

Off-balance-sheet liabilities? None material flagged; Partner Share obligations are contractual cost-of-revenue, not off-balance-sheet. (Fact, per filings.)

How conservative is the accounting? Aggressive in presentation: SBC-flattered Adjusted EBITDA (add-back > the metric itself), ~$15–19M/yr capitalized software propping OCF, serial impairments treated as below-the-line, and a mid-stream revenue-presentation change. The audited GAAP picture (negative equity, losses) is honest; the non-GAAP framing is generous. (Interpretation.)

How CapEx-hungry is the business? Low physical capex, but meaningful capitalized software development (~$15–19M/yr), which is the real reason FCF is negative while OCF is positive. (Fact.)

Capital Allocation & Management

How much FCF does the business generate; how is it used; philosophy? Negative FCF every year (−$6.5M FY25, −$7.9M Q1’26). There is no FCF to allocate; cash has been consumed and supplemented by debt/ATM. Philosophy 2021-vintage: growth-by-acquisition (failed); current: cost discipline + maturity extension + asset shedding (damage control). (Fact/Interpretation.)

Significant acquisitions recently? No new acquisitions; the story is divestiture — Bridg to a PAR affiliate (~$27.5M) after a $578.9M purchase. (Fact.)

Buying back shares? No — the company issues shares (ATM 3.91M @ $12.80 = $48.3M in 2024; ~$28–45M/yr SBC). Heavy dilution, not buybacks. (Fact.)

Issuing large amounts of new shares to insiders? SBC ~12% of revenue; 1,000,000 fresh RSUs to the CEO at the trough. (Fact.)

Compensation policy / motivations of management. Bonus/PSU metrics are Adjusted Contribution and Adjusted EBITDA — no ROIC/return-on-capital metric, so M&A value destruction never hit the scorecard (central flaw). Pay did fall with poor FY25 results (30% achievement; CEO $166k vs $550k target; ex-CFO $0). New CFO bought 200k shares (~$128k) — a modest positive signal. Director/officer ownership ~5.9% pre-split. (Fact/Interpretation.)

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — ordinary U.S. common stock (Delaware C-corp), Form 1099. (Fact.)

Dividend policy? None; pays no dividend (and cannot — negative equity, cash preservation). (Fact.)

How profitable is the business? Unprofitable on GAAP; marginally cash-flow-positive at the operating line only after cost cuts and capitalized software. (Fact.)

Is net income diverging from cash from operations? Yes, sharply — FY25 net loss −$103.5M vs. OCF +$9.3M, the gap being non-cash impairment ($58.8M), SBC ($28.1M), and D&A. The non-cash add-backs are large and recurring. (Fact.)

Risks & Downside

What factors would cause the stock to decline further? Any additional FI-partner non-renewal (esp. Wells Fargo/Amex), continued double-digit revenue decline, a dilutive rescue equity raise, Nasdaq delisting, or signs the 2029 converts will equitize. (Interpretation.)

Risk of a catastrophic loss? Yes, material. Net debt (~$144M) exceeds equity market value (~$53M); equity is a levered option. The 2028–29 maturity wall against ~$10M Adjusted EBITDA is the catastrophe vector. (Interpretation.)

Chance of a total loss? Non-trivial. With negative book equity today and a refinancing cliff into a shrinking business, a 2028–29 equitization that wipes out common is a live base-case-adjacent outcome, not a remote tail. Mitigant: no 2026–27 maturities and positive OCF buy time. (Interpretation.)

Recent News & Events

Has the business environment changed recently? Drastically, for the worse: Bank of America exited and in-housed (Feb 2026); Chase (largest partner) owns competitor Figg; Mastercard launched Commerce Media (Oct 2025). (Fact.)

Significant acquisitions / divestitures? Bridg divestiture to a PAR affiliate (Jan 2026 APA). (Fact.)

Change in accounting policies? Mid-stream revenue-presentation change (FI Share reclassified) and cessation of MAU/ARPU reporting (Jan 1, 2025). (Fact.)

Recent changes — markets, facilities, management? New CEO (Amit Gupta, Aug 2024), new CFO (David Evans, 2025/26), 1-for-10 reverse split (June 5, 2026), Nasdaq deficiency/second compliance period. (Fact.)

Appendix B — Source Appendix

Cardlytics, Inc. (NASDAQ: CDLX), CIK 0001666071. All sources accessed June 7, 2026. Primary sources prioritized.

Primary — SEC filings (EDGAR, CIK 1666071)

Primary — corporate / market data

Secondary — reverse split / Nasdaq compliance

Secondary — industry / competitive

Notes on data reliability

• Public market-data feeds were garbled by the 1-for-10 reverse split (effective 2026-06-05; split-adjusted trading from 2026-06-08): some feeds showed the post-split share count (~5.8M) against a pre-split price, producing a nonsensical ~$3.5M market cap. All capital-structure figures were reconstructed from the Q1’26 10-Q and back-solved against the enterprise-value calculation. Pre-split price ~$0.91; pre-split shares ~58.08M → ~5.81M post-split; equity market cap ~$53M.
• Own-history valuation percentiles were unavailable for CDLX — own-history percentile cross-check could not be performed.
• Revenue-presentation change: the FY25 10-K reclassified FI Share into revenue, and Bridg was moved to held-for-sale/discontinued operations; figures are presented as reported, with the reclassification flagged as an open item before any run-rate conclusion.