CDLX — Q2 2026 Financial Report Analysis

Overview

Cardlytics runs the "offers" tabs inside banking apps. A brand such as a restaurant chain pays Cardlytics to show a cash-back offer to bank customers whose card history suggests they might buy. Cardlytics passes part of that money back to the shopper as a reward and pays the bank a share for access to its app and data. In the second quarter of 2026 (April to June) that whole chain got smaller: GAAP revenue fell 36.5% to $36.9 million, mainly because a U.S. bank partner left the platform. Cost cuts kept a small positive adjusted EBITDA of $1.7 million, but the net loss widened to $14.9 million from $9.3 million and the company spent $10.7 million of cash. Two things outside day-to-day operations made the loss worse: a $6.2 million charge tied to the Bridg business it sold in March, and the absence of last year's $5.4 million currency gain.

At a glance

• Billings down 33.8% to $65.5 million. Billings is the total amount advertisers were invoiced. The 10-Q puts the drop down to "the change in bank partner relationships": fewer bank customers to show offers to means fewer campaigns to sell.
• Operating cash flow of –$8.6 million against adjusted EBITDA of +$1.7 million. The profit measure management highlights did not turn into cash. As the business shrinks, Cardlytics is still paying out rewards and bank shares it owes from earlier, bigger quarters.
• Cash fell to $28.0 million from $48.7 million at the end of 2025. The company owes $172.5 million of convertible notes due April 2029. It also has to keep at least $20.0 million of unrestricted cash in its bank accounts to meet a loan covenant, a floor that was lowered from $25.0 million in April.

How the money flows, and where it shrank

The layers of Cardlytics' income statement need a short explanation, because the company's revenue is not what advertisers paid:

1. Billings ($65.5M): the gross amount advertisers paid.
2. minus Consumer Incentives ($28.6M): the cash-back rewards passed on to shoppers. Cardlytics reports revenue after taking these out, because it doesn't sell the product the shopper buys.
3. = GAAP Revenue ($36.9M), effectively "billings less consumer incentives".
4. minus Partner Share and other third-party costs ($15.6M), mostly the cut paid to banks.
5. = Adjusted Contribution ($21.3M), the non-GAAP figure management calls "the dollars we keep after all of our partners are paid".

Revenue fell faster than billings because rewards fell more slowly than billings did. Rewards rose from 41.3% to 43.7% of billings. The 10-Q says this was "primarily due to strategic decisions to drive incremental performance for our advertisers as well as optimization of our network". In plain terms, Cardlytics funded richer offers to keep campaigns working for advertisers, and gave up revenue to do it.

Partner share fell even faster (–41.6%), helped by "changes in Partner mix". As a result, Adjusted Contribution fell less than revenue, and the share of revenue Cardlytics keeps rose from 54.0% to 57.7%. That is the main bright spot in the quarter.

The user base shrank too. MQUs (bank customers who made a card purchase in a month and whose data Cardlytics received) fell by 39.0 million, "primarily driven by an FI partner in the U.S. exiting the Cardlytics purchase intelligence platform." Each remaining user also earned less (ACPU down from $0.14 to $0.11), so the drop was not only the lost bank. By pricing model, revenue from served-based campaigns, where advertisers pay a percentage of all purchases by shoppers who were shown an offer, fell 44.8% to $11.8 million. Engagement-based revenue, paid per purchase after a shopper activates an offer, fell 33.5% to $23.9 million.

Key figures

Per-share figures reflect the 1-for-10 reverse stock split of June 5, 2026; prior-year figures are restated for it. The 10-Q reports per-share loss in two lines, –$1.50 from continuing operations and –$1.08 from discontinued operations. The –$2.58 total is their sum and matches the $14.9 million net loss divided by 5.76 million weighted shares.

Operating margin is operating loss divided by revenue, so here it shows how much of each revenue dollar is lost before interest and other items. It improved even though revenue collapsed, because costs were cut harder. Sales and marketing, R&D and G&A together fell 35.6% to $20.0 million. The filing points mostly to lower staff costs after the 2025 layoffs: staff expense was down $1.1 million in sales and marketing, $2.4 million in R&D and $1.4 million in G&A. Delivery costs also fell $1.4 million from cheaper data storage.

Takeaway: Cardlytics has cut costs about as fast as it lost revenue, which keeps adjusted EBITDA slightly positive. Cash is a different story. The quarter burned $10.7 million, the cash pile is $8 million above its covenant floor, and the September settlement of a Bridg-era legal claim took roughly $6.4 million more after quarter-end. Without a new bank partner or a turn in billings, the cost cuts only slow the decline.

What the headline numbers hide

• The adjusted EBITDA isn't turning into cash. Adjusted EBITDA of $1.7 million compares with operating cash flow of –$8.6 million and free cash flow of –$10.7 million, which includes $2.1 million of capitalized software spending that adjusted EBITDA never counts. For the half-year, operating cash use of $14.2 million was driven by a $10.3 million drop in the consumer-incentive liability and a $6.3 million drop in the partner-share liability, partly offset by $13.2 million of lower receivables. The 10-Q says rewards are paid "between one and four months following redemption". When billings fall, Cardlytics is still paying the bills from bigger earlier quarters while taking in less new money, so a shrinking business drains working capital.
• What adjusted EBITDA leaves out is large. To get from the –$14.9 million net loss to +$1.7 million, the company adds back $6.2 million of discontinued-operations loss, $4.1 million of depreciation and amortization, $2.4 million of stock compensation, $2.3 million of net interest, a $1.1 million loss on the PAR shares it received for Bridg, and $0.7 million of severance. Interest is a real cash cost: the 4.25% coupon on $172.5 million of notes alone is about $7.3 million a year.
• Stock comp fell partly for a one-time reason. Stock-based compensation dropped $4.3 million year on year (from $6.6 million to $2.4 million). The filing attributes this "primarily" to "higher forfeitures due to a reduction in headcount": departing employees forfeited unvested awards, and those forfeitures reverse expense already booked. That flatters the GAAP operating loss and will not repeat at the same size.
• Last year's currency gain distorts the comparison. Q2 2025 included a $5.4 million foreign-currency gain on the British pound, against $0.2 million this quarter. Strip that out and the pre-tax loss from continuing operations improved to about –$8.8 million from –$11.5 million. The headline "loss from continuing operations widened" is mostly an FX comparison effect.
• The Bridg charge is a legal indemnity. Discontinued operations lost $6.2 million on zero revenue, because Bridg was sold to PAR Technology on March 24. Almost all of it is $5.5 million of G&A, chiefly a $5.2 million increase in the accrual for indemnifying a former Bridg officer over his share of a settlement in a lawsuit (DailyGobble v. Jain), plus legal fees. The 10-Q expected $4.0 million to be covered by D&O insurance but had not booked that recovery. A September 4, 2026 8-K then disclosed a settlement: Cardlytics paid about $6.44 million and took over the officer's rights in his lawsuits against the insurer. The $4.0 million recovery now depends on Cardlytics winning or settling those suits.
• Partner concentration got worse. In the first half of 2026 the top three bank partners accounted for over 85% of the partner share paid (over 80% a year earlier), and the top two for over 60%. The earnings release names Chase and Wells Fargo as key dependencies. Most bank agreements can be ended on "90 days or less prior notice". The quarter shows what losing one bank does to the numbers.
• Balance sheet. Stockholders' equity is negative: a deficit of $16.0 million, widened from $6.5 million in December. Net repayments of $25.1 million on the credit line (most of it from the $23.0 million raised by selling the PAR shares) brought the line down to $15.0 million from $40.1 million. Cash fell by $20.7 million at the same time. Weighted shares rose 9.2% year on year, so existing holders are being diluted as well.

Guidance and outlook

The earnings release (8-K, August 5) gave Q3 2026 guidance:

The CEO said Q2 came in "within our guidance across all key metrics", and the CFO spoke of a "game plan for sequential growth and self-sustainability". The midpoints above don't show that growth yet: revenue ($36.5M), billings ($64M) and adjusted EBITDA ($1.5M) are all at or slightly below Q2. The year-on-year declines should narrow in Q3 simply because the comparison base is lower once the lost bank is in both years. That narrowing alone won't mean the business is growing. In the 10-Q, management still expects "to incur additional operating losses in the near term" and believes it has enough liquidity for at least the next 12 months.

Our read on Q3 (10-Q due around early November):

1. Sequential billings. Anything above Q2's $65.5 million would be the first real sign that new advertisers or deeper bank partnerships are replacing what the exiting partner took away. The fourth quarter is seasonally strongest, so Q3 is the cleaner test.
2. Cash against the $20 million floor. Q2 ended with $28.0 million of cash. The roughly $6.4 million Bridg-related settlement came after quarter-end, and Q2 free cash flow was –$10.7 million. Even with $20.7 million of unused credit-line capacity, the Q3 cash balance and any further covenant amendment are the numbers to check first.
3. Reward intensity. Consumer Incentives rose to 43.7% of billings this quarter. If that keeps rising, revenue will keep falling faster than billings.