How Indeed Overrode Instinct on a Major Brand Campaign with Outset

Sometimes the most valuable thing research can do is tell you what not to do. For Indeed's brand team, that lesson played out in Germany where a beloved character, strong stakeholder enthusiasm, and two rounds of rigorous AI-moderated research converged on a decision that was hard to make, but clearly the right one.

The Challenge: A Beloved Character With a Structural Flaw

Indeed's German market had a history with Ingrid. From 2021 to 2023, the character anchored Indeed's best-performing campaign ever in any market; so deeply embedded in the brand that German job seekers still associated her name with Indeed years after the campaign ended. When the opportunity arose to bring Ingrid back for a new campaign, the agency was excited. Internal stakeholders were excited. The instinct to move forward was strong.

Angela Kesselman, Brand Research and Insights Lead at Indeed, saw the opportunity and the risk. The new positioning needed a campaign that didn't just entertain job seekers, but clearly conveyed how Indeed helps them find work. The question was whether the Ingrid concept could do both.

The Solution: Two Rounds of AI-Moderated Testing

Indeed used Outset to run two sequential rounds of creative testing with German job seekers, each designed to answer a specific question.

Phase 1 was a discovery study: 195 interviews across three creative concepts, with 65 participants per concept, conducted via 20-minute AI-moderated video interviews. The goal was to understand which concept to pursue and where the risks were. Outset's multilingual capability made the research operationally straightforward. Setting the study language involves a single dropdown selection, and all participant responses are automatically translated back into the researcher's working language, eliminating the time and cost of external translation.

The findings from Phase 1 were clear. The Ingrid concept had real emotional appeal: participants found it warm, uplifting, and easy to connect with. But it failed on a more important dimension. As Kesselman put it, the concept had "the Ingrid sugar high.” It appealed on the surface, but lacked the underlying substance to meaningfully progress the brand. Participants couldn't articulate how Indeed was actually helping them find a job, so the concept entertained without earning credibility.

The agency reworked the concept to address that gap. Phase 2 was a pressure test: 38 additional interviews with German job seekers reviewing the revised animatic, designed to answer one question — had the core issue been fixed?

It hadn't. The feedback, as the research output described it, had deja-vu running high. The revised concept wasn't a clear rejection, but it wasn't a clear endorsement either. A sizable minority still questioned whether the spot showed how Indeed was genuinely helping them. And a new problem had emerged: a twerking scene in the revised concept read as inappropriate and off-brand.

The Results: Research Overrode Instinct

Armed with two rounds of consistent findings, Kesselman and her team used the data to make a call that went against the grain of internal enthusiasm. Indeed stepped away from the Ingrid-centric concept and pivoted to a different direction; one the agency had developed later in the process, initially seen as riskier, but which the research suggested had meaningfully higher upside.

"It's a hard thing to walk away from what your gut is telling you," Kesselman acknowledged. The Outset data made it possible.

As the slide deck from the project framed it directly: without the research, stakeholder enthusiasm wins and Ingrid launches as a structurally flawed campaign dressed up as the safe choice. With the research, Indeed pivots to a concept with a clearer brand role and a higher ceiling of impact.

The new campaign launched in Germany shortly after the research concluded. Additional creative testing conducted as the work progressed through development showed strong scores and the iterative nature of the testing, made practical by Outset's speed and cost, was what made that possible. Rather than a single evaluative pass at the end of a long development process, Indeed was able to run a check, respond to findings, and run another check, building confidence through the process rather than hoping for it at the end.