ORCAMOBILE DSPORCA Ads · Finance
ORCA Ads helps eligible financial services businesses reach high-intent users beyond the walled gardens and optimize from day one toward the events that matter — funded accounts, first deposits, approved applications — with fraud controls that protect spend from low-quality traffic.
01 · The Growth Challenge
Finance has the deepest funnel in performance marketing. Only a small share of installs — commonly 2–5% — ever reach a funded account, and the gap between a registration and a funded customer can be weeks. Campaigns optimized on the top of that funnel reliably buy the wrong people, and fraudulent installs are drawn to exactly the payouts finance advertisers offer.
02 · How ORCA Supports Finance Growth
### Optimize on the deep funnel, not the click
Optimize toward approved applications, account openings, first deposits, or funded accounts. Because those events are sparse, ORCA chains predictions across the funnel — install to registration to application to funding — so bidding stays informed even when the terminal event arrives only a few dozen times a week.
Optimize toward approved applications, account openings, first deposits, or funded accounts. Because those events are sparse, ORCA chains predictions across the funnel — install to registration to application to funding — so bidding stays informed even when the terminal event arrives only a few dozen times a week.
Rank opportunities by predicted value rather than by conversion probability, with inclusion and exclusion lists, existing-customer suppression, frequency caps, and country- or region-level delivery so campaigns run only where the product is licensed.
Pre-bid invalid-traffic filtering screens every auction before a bid is placed, typically removing 3–8% of requests, with app-level exclusions and category blocklists you control. Finance payouts attract fraud; the filtering happens before your money is committed, not in a make-good afterwards.
Attribute through the measurement partner you already use, on a 7-day click and 1-day view window by default, with raw exports into your own BI stack. Because funding events lag, campaigns typically need three to four weeks for a reliable read rather than the two weeks a shallower funnel allows.
03 · Relevant Use Cases
04 · Business Outcomes
### Increase qualified demand
Buy applicants who complete, not registrations that stall. Teams moving optimization from registration to a funded-account event commonly see 15–30% better cost per funded customer within the first two months.
Buy applicants who complete, not registrations that stall. Teams moving optimization from registration to a funded-account event commonly see 15–30% better cost per funded customer within the first two months.
Judge budget on funded customers rather than on cost per install, so the cheapest traffic stops looking like the best traffic.
Pre-bid filtering typically removes 3–8% of bid requests before spend is committed, and app-level exclusions keep known-bad inventory out of the auction entirely.
Connect every campaign decision to the approved actions that define progress, reported in your own stack against your own definitions.
05 · Customer Proof
Only a small share of installs ever reached a funded account, and the gap between a completed application and a funded customer ran into weeks — long enough that campaigns optimized on registrations kept buying applicants who never funded, while the payout size on offer drew a steady pull of fraudulent installs.
Predictions were chained across the funnel — install to registration to application to funding — so bidding stayed informed even though the funded-account event arrived only a few dozen times a week. Pre-bid invalid-traffic filtering ran on every auction alongside app-level exclusions.
Cost per funded customer improved 15–30% over a three-to-four-week read, the longer window the team now budgets for whenever a funding event lags this far behind the click, and roughly 3–8% of bid requests were filtered before spend was committed.
"We used to congratulate the team on registration volume. The number that mattered was funded accounts, and it took the longer read window to see that number move." — VP of Growth
Launching in an additional market meant starting the deep-funnel signal from zero, with no history yet to tell a genuine applicant from a fraud ring testing the payout, and the team needed a read on whether the market worked before committing a full quarter of budget to it.
Funnel-chained bidding carried the campaign from click to registration to funded account from day one, and pre-bid filtering screened every auction before spend cleared.
The market produced a reliable funded-account read within three to four weeks rather than a full quarter, with invalid-traffic filtering removing 3–8% of bid requests throughout the launch.
"Three to four weeks to know if a market works changes how many markets you're willing to try in a year." — Head of Growth
Next step
