Growth should be measured by results customers actually achieved.

A digital banking app operating across several Latin American markets · ORCA Ads — User acquisition

Aim optimization at funded accounts.

Challenge: Installs and registrations were not the shortage. What stalled growth was funded accounts, and funding often lagged registration by a long time. Campaigns tuned to registrations bought a large volume that never reached funding, and incentive amounts easily attracted invalid installs.

Approach: Prediction was chained along the funnel, from install to registration, application, and funding. Even when the terminal event was sparse, bids could still use signals earlier in the funnel. Pre-bid invalid-traffic filtering and app-level exclusions ran on every auction.

Result: Inventory mix shifted toward users more likely to complete funding. Invalid requests were filtered before spend was committed. Cost started to be read on funded customers; registration volume was no longer treated as success.

"Registration volume was always enough. What stalled growth was funded accounts. Once campaigns were allowed to predict further down the funnel, the inventory mix changed quickly." — VP of Growth

A regional general-merchandise marketplace with a large base of active sellers · ORCA Commerce Media

Turn seller promotion budget into a measurable advertising business.

Challenge: Advertising was still seller-funded promotions, with little basis for renewal. Sellers bought placements but could not see SKU-level return. Internal teams also worried that adding ads in search would hurt conversion.

Approach: Sponsored products launched first on search and browse. Ranking weighed predicted purchase probability and bid together, with a relevance floor set by the marketplace, and self-serve campaigns generated from the seller catalog feed. SKU-level ROAS reporting was opened directly to sellers.

Result: Advertising became media revenue that could be renewed on performance. Sellers could see return at the product level. The relevance floor kept poorly matched products out of the top, and the search experience was protected.

"The worry was that monetizing search would sacrifice conversion. Poorly matched products cannot outbid well-matched products here — the campaigns themselves later proved that." — Director of Monetization

A food-delivery platform expanding across multiple cities · ORCA Ads — User acquisition and remarketing

Buy the second order.

Challenge: A large share of first-time customers never placed a second order, so promo-led acquisition was structurally hard to make profitable. Spend also landed outside delivery range, or in hours when capacity was already short.

Approach: Targeting narrowed to active service areas, budget switched by time of day, and the optimization event moved from first order to a second order within a defined window. Lifecycle campaigns remarketed customers who had stalled after one order.

Result: The share of newly acquired users who reached a second order became the outcome campaigns watched. Lapsed users were invited back at lower cost. Whether a new city could work was judged on the customer's own order data, not after burning through a long stretch of budget.

"It used to take a long stretch of budget before we knew whether a city was worth continuing. After optimization aimed at the second order and campaigns stayed inside fulfillable areas, the read got much clearer." — Head of Performance Marketing

See the advertising product line for retailers and marketplaces.

Read research and practical guidance on media-buying decisions, user value, and performance optimization.