NBCU Peacock

Big picture

The client: NBCUniversal Peacock
My role: Lead Researcher @ Pilot.ly | Product/UX research on interactive ad format and copy
The project:

• Objective: Peacock was evaluating Infillion’s “Ad Choice” product, an interactive format that lets viewers choose between a shorter, interactive ad break or a longer, standard one. Before rolling it out, Peacock wanted to understand: Did it improve or harm the viewing experience? Was it impactful for the brands whose interactive ads were being served? And would viewers actually use it? Finally, what copy should be used to drive adoption of the product while still being in line with Peacock’s brand?

• Research: I led an experiment in a simulated Peacock environment built with Peacock’s engineering team, which mimicked the real product and allowed us to capture behavioral data (clicks, plays, scrubs, opt-ins). Viewers were randomized into four conditions: a control group with no choice, Infillion’s time-centric copy, and two Peacock-branded copy variants. We tracked opt-in rates to the interactive ad, then followed up with a post-exposure survey to measure viewer experience, perceived control, and brand lift for KitchenAid, the brand partner.

• Findings: The study showed that clear, time-focused copy almost doubled opt-in compared with more “on-brand” Peacock language. Viewers rewarded prompts that explicitly stated how many ads and how much time they were choosing. Offering Ad Choice did not harm overall viewing experience, and in fact, even those who opted out appreciated simply having the option. For brands, the interactive ads significantly improved upper-funnel metrics like awareness and recall.

• Impact: Based on these findings, Peacock decided to adopt the Ad Choice format using the time-centric copy, despite its less “relaxed” tone that was less in line with its brand.

The problem

Peacock was considering implementing Infillion’s “Ad Choice” product, an interactive ad experience where viewers could choose between a shorter, interactive ad break or a longer, standard ad break. Before rolling it out, they wanted to understand if it would create both a positive brand impact and a positive viewer experience. The main questions we sought to answer were:

• Does the Ad Choice product improve brand metrics (recall, favorability, consideration)?

• Does being offered the Ad Choice improve or harm viewer experience?

• If offered, do viewers actually want to use it? Who opts in to interactive ads vs sticking with regular breaks?

• Among those who opt in, is the experience positive? Or does requiring interaction create friction in a lean-back, streaming context?

• Which copy drives the most adoption? Infillion’s standard prompt is clear and time-centric, but doesn’t match Peacock’s more relaxed brand voice. Peacock wanted to test whether more “on-brand” copy would perform better or worse.

Methodology

To approximate a real product environment and capture both behavioral and attitudinal data, we used a mixed approach, including a simulated Peacock interface, A/B testing of the different Ad Choice options and copy options, and a survey to assess viewer experience afterwards.

Simulated Peacock environment

We worked with Peacock’s engineering team to build a “lite” Peacock experience. Viewers were shown an episode of TV (either Parks & Recreation of Vanderpump Rules) on a simulated version of Peacock's interface, where we were able to capture interactions like clicks, play/pause, scrubbing attempts, etc, in order to measure how "leaned in" or "leaned back" viewers were.

A/B test: four ad experiences

During the first ad break, viewers were randomized into one of four conditions:

1. Control – standard ad experience with no choice

2. Ad Choice 1: Infillion Copy – Ad Choice card using Infillion’s standard copy (explicit number of ads + time commitment)

3. Ad Choice 2: Peacock Copy A – Ad Choice card rewritten in Peacock’s more relaxed brand voice.

4. Ad Choice 3: Peacock Copy B – A second Peacock-styled variant.

In each of the Ad Choice cells, viewers could opt into either a shorter, interactive ad break, or a longer, standard ad break. We then measured:

• Opt-in rate (the share of viewers choosing the interactive ad).

• How those opt-in rates varied by copy, TV show, and timing during the viewing session (early vs later).

Post-exposure survey

After viewing, respondents completed a survey to capture:

• Viewer experience

  • Overall viewing experience

  • Perception of the ad experience

  • Attitudes toward the ability to choose

• Brand lift for KitchenAid

  • Aided brand recall

  • Brand favorability

  • Likelihood to consider/purchase and recommend

• Future intent

  • Willingness to use Ad Choice again

  • Interest in seeing more interactive formats in the future

We then compared these responses between the Control (no choice), the Ad Choice viewers who opted out (chose a regular break), and the Ad Choice viewers who opted in (experienced the interactive ad).

Findings & Recommendations

Simply being offered the choice is a positive experience for viewers, even if they opt out of the interactive ad.

In order for Peacock to assess whether they should implement the product at all, they were specifically interested in whether offering a choice might harm the experience for viewers who choose not to engage with the interactive ad.

However, we found that overall ad experience ratings were high and similar for both control and Ad Choice groups. Offering the choice did not hurt the perceived ad experience—in fact, around 90% of viewers felt favorable about simply having the ability to choose, regardless of whether they opted in or opted out.

In qualitative responses, viewers mentioned feeling empowered (“It made me feel more in control.”), appreciative (“It was nice to have a choice instead of one being forced on me.”), and having a sense of novelty (“I haven’t seen this before; it was interesting.”).

Interactive ads significantly boost brand recall, but not all brand metrics.

On the brand side, we found that aided recall for KitchenAid (the interactive ad partner) was nearly double among those who opted into the interactive ad vs those who saw only standard ads. Opt-in viewers were also more likely to rate the KitchenAid ad as “excellent” than opt-out or control viewers.

However, brand favorability, purchase intent, and likelihood to recommend were largely unchanged across conditions—suggesting that though the format makes the brand memorable, it doesn't meaningfully move the needle on lower-funnel metrics.

Clear, time-centric copy drives significantly higher opt-in.

Across all cells, the overall Ad Choice opt-in rate was 26%, but performance differed sharply by copy: Infillion’s standard prompt had the highest opt-in rate at 41%, significantly outperforming both Peacock variants.

Qualitative feedback that we gathered in open ends showed that viewers appreciated knowing exactly how many ads they’d see and how much time they were committing. Meanwhile, Peacock’s copy was on-brand but much vaguer, and created confusion (“I wasn’t sure what the options meant”).

When we asked control viewers (who didn't get the choice) a hypothetical question using Infillion's copy, their stated opt-in rate (~39%) was very close to the actual Infillion condition, reinforcing that the clear, time-based framing helped drive its adoption.

Viewers should be given the choice before they fully “lean back”

Finally, we observed strong evidence that Ad Choice works best early in the viewing session. Opt-in was higher for Parks & Rec, where the ad break was served 7 minutes into the 22-minute episode, than for Vanderpump Rules, where the ad break was 13 minutes into the 45-minute episode. Additionally, behavioral data from the simulated environment showed that opt-in viewers tended to be more actively engaged (more clicks, scrubs, and interactions)—but engagement slowed during longer content, so even if a viewer was actively engaged earlier in the session, they may have “lean back” by the time the ad break was served.

Recommendations

Based on all of these findings, we recommended that Peacock adopt the Ad Choice product, and use Infillion’s clear, time-focused copy (even despite that copy being less aligned to Peacock’s brand voice). We also recommended that they present the choice as early as possible, during the first ad break.

Impact

NBCUniversal Peacock decided to adopt the Infillion Ad Choice product using the time-centric copy we recommended. Peacock was satisfied, based on our findings, that the Ad Choice would not degrade the viewing experience, and would even help viewers feel a sense of control. The positive impact on upper-funnel brand metrics were also shared with Peacock’s brand partners in order to drive ad revenue.

Reflections

It's small, but I found myself surprised at just how well the Infillion copy did compared to Peacock’s. I think both me and the client were expecting a minor difference that could be written off in favor of sticking to Peacock’s brand voice, but the difference ended up being so large—more than 2x as good!—that the benefits were unignorable.

It really sharpened my ability to think about everyone with stakes in a product decision. In this case, there wasn’t just one “user” to satisfy — there was Peacock trying to protect its brand voice, Infillion wanting to prove the value of its Ad Choice product, the advertiser (KitchenAid) looking for meaningful brand lift, and, of course, the viewer. In both designing the study and presenting the findings, I had to be thinking about the trade-offs between these. In designing the study that was considering the positive brand impact vs. the lean-back experience for the viewer; in presenting results it was the Infillion vs. Peacock copy example. It made me think not just about what was working, but also who it would work for, and where interests were either aligned (the viewer + brand both like the Ad Choice) or in tension (Infillion vs. Peacock copy).

Another big takeaway for me was the power of cross-team collaboration in getting us as close to the real product as possible. The simulated Peacock environment that we built with the engineering team let us observe real behaviors—clicks, pauses, opt-ins—instead of relying only on surveys, and significantly strengthened our findings from the questionnaire alone. At the same time, this also highlighted some limitations of working as an external vendor. I know the product was adopted, but had no direct visibility into longer-term metrics after the product was launched. If I were on the product team, I’d definitely want to follow up with this work. Was adoption impacted after the novelty wore off? What about enjoyment of the experience (and would that impact the positive effect for brands?) How could we iterate on the Ad Choice product?