AI Pet Portrait: GTM for a Niche AI Product

Context

AI Pet Portrait launched in October 2023. The idea came from an internal data signal: in-app pet retouching templates on BeautyPlus had consistently performed well, signaling that demand existed even before a dedicated product did. There were no AI pet portrait products on the market at the time. The design team acted on that signal, training a LoRA model to generate high-quality stylised pet portraits. When the team decided to move forward, the primary GTM goal was clear: use this product to rank on the App Store and Google Play charts, and convert that visibility into organic user acquisition.

My role was to execute the in-app GTM strategy: converting bought-in traffic into engagement, managing in-app promotional channels, and tracking funnel performance.

The Challenge

• The audience ceiling was real and known from the start. BeautyPlus's user base skewed heavily toward young women interested in photo editing. Without needing formal research to confirm it, we could reasonably assume that pet owners interested in an AI pet portrait product would be a small minority within this user base.
• Promotions had to reach the full user pool, which meant the majority of impressions were going to people the product was not built for. Non-pet-owners simply had no reason to click. Among those who were pet owners, the bar to engage was still high: uploading 10 to 15 photos was a lot to ask of users. The addressable audience was small to begin with, and the fraction of that audience willing to complete the upload was smaller still.
• Cold-start signal was hard to read. My expectations for click-through rate were set low from the start. And even when the data came in, a low number was difficult to interpret: it could reflect genuine lack of interest, or simply the fact that most users were not pet owners. That ambiguity made it harder to decide on next steps with confidence.

Pre-launch

Accepting the niche and planning for it

Given that targeting was not possible, the question was how to calibrate the push. Too much promotional weight for too long would take up space that other features needed. Too little, and we would not accumulate enough data to draw any meaningful conclusions about whether the product was working.

My decision was to focus the cold-start effort on pop-up windows, which offered the highest reach and most direct exposure. Working with colleagues handling ad optimisation and ASO, we executed a concentrated push within a short window: two weekends at maximum priority.

For in-app promotion, the logic was deliberate: two weekends of concentrated exposure is enough to see whether the product could achieve our App Store and Google Play ranking goals. If the downstream data looked strong, we would continue pushing. If not, the team would discuss and decide whether the issue was the product concept, the promotional creative, or something in the funnel, and how to reallocate resources from there. The goal was to get a clear read as fast as possible without committing disproportionate resources to a product that had not yet proven itself.

Phase 1: October launch

Aggressive cold-start, then read the data

There were three main entry points into the product. Pop-up windows drove the highest volume and served as the primary awareness channel. Homepage banners had lower click volume but attracted users with higher purchase intent. The MiniApp Tab generated the fewest entries but the most committed users. The pattern was consistent with what we had seen across other MiniApp launches: channel volume and channel quality move in opposite directions.

Of the approximately 846,000 users who saw the Pet Portrait entry across all channels, only 6.5% clicked through to try the product, significantly lower than AI Portrait (19% Western, 13.5% Japan/Korea) and much lower than AI Filter's 30% or more. This gap was expected and consistent with the niche hypothesis.

The steepest drop-off was at the photo upload step. Only 39% of users who reached the photo selection page completed the upload. Two hypotheses: users did not have 10 to 15 suitable pet photos available, or the photo quality guidelines felt too strict for casual pet photography. Pets do not hold still; most phone photos have motion blur, partial frames, or inconsistent lighting.

One finding from style performance worth noting: all four styles showed remarkably even purchase distribution, with Pawtrait at 29.9%, Master Chef at 27.2%, Wizard at 21.6%, and Birthday Party at 21.4%. No single style dominated, which suggested users had no strong preference for any particular aesthetic, or liked them all equally.

Phase 2: Christmas campaign

Scaling the niche, and what breaks when you do

By December, Pet Portrait had become one of the more robust performers in the MiniApp portfolio. The launch did not push the app to the top of the store rankings, but at its peak the product ranked as high as #3 in Japan's App Store. The cost per acquisition from related ad spend remained low throughout, providing a stable growth lever without significant budget pressure.

The October data had demonstrated that users who made it through the funnel converted at a strong rate at the payment step. New styles launched in waves through December, and the Christmas period proved to be a natural fit for the product. December 25th became one of the highest single-day revenue days across the entire MiniApp ecosystem.

During the Christmas campaign, all three MiniApp products, AI Filter, AI Portrait, and AI Pet Portrait, were given high promotional priority, each having launched new Christmas-themed style packs. Of the three, AI Pet Portrait ended up contributing approximately 40% of total MiniApp GMV in December, a result that exceeded expectations. Revenue grew over the period, but as promotional volume tripled during the holiday season, conversion rate was diluted in parallel, falling from 0.3% in October to 0.1% in December, a predictable outcome of reaching further beyond the core audience.

Key Takeaways

All metrics and performance data referenced in this case study reflect the 2023–2024 period. Sensitive figures have been anonymised or aggregated, and no proprietary or commercially confidential information is disclosed.