Perplexity Shopping: $180K Ad Spend Lessons for 2026

Attributing value accurately in the complex digital advertising ecosystem requires a deep understanding of how different platforms generate and interpret user signals. Our recent campaign for “Perplexity Shopping,” a new e-commerce aggregator, provided a compelling case study in dissecting these data signals to refine attribution analytics and drive demonstrable return. We allocated a significant budget to this initiative, aiming for aggressive user acquisition in a crowded market.

Campaign Overview: Launching Perplexity Shopping

The objective for Perplexity Shopping was straightforward: drive initial user sign-ups and first purchases for their new platform, which aggregates deals from various online retailers. We launched a complete digital advertising campaign across several major channels, including paid search, social media, and programmatic display. The campaign duration was set for three months, from January to March 2026, with a total budget of $180,000. This budget was distributed to prioritize platforms known for strong lower-funnel performance while reserving funds for upper-funnel brand awareness.

Our initial strategy focused on a blend of direct response and awareness tactics. For direct response, we targeted users actively searching for deals or specific product categories. Awareness efforts aimed at broader audiences with an interest in online shopping or saving money. The key performance indicators (KPIs) included Cost Per Lead (CPL) for sign-ups, Return On Ad Spend (ROAS) for purchases, Click-Through Rate (CTR), and overall conversion volume. We understood that a new platform would face challenges in immediate recognition, so a phased approach to optimization was built into the plan.

Initial Strategy and Creative Approach

The creative strategy centered on showing Perplexity Shopping’s unique value proposition: simplifying the deal-finding process. We developed a series of ad creatives featuring crisp visuals of popular products alongside compelling price comparisons. For paid search, ad copy emphasized keywords like “best deals,” “discount shopping,” and “price comparison.” Social media campaigns used short video ads demonstrating the platform’s user interface and ease of use, with clear calls to action (CTAs) to “Sign Up Now” or “Find Deals.”

Targeting for paid search was primarily keyword-based, focusing on high-intent commercial queries. On social platforms, we layered interest-based targeting (e.g., “online shopping,” “e-commerce,” “couponing”) with lookalike audiences built from an initial seed list of beta testers. Programmatic display campaigns employed contextual targeting to place ads on websites related to retail, finance, and consumer technology. The goal was to reach users at various stages of their shopping journey, from initial research to purchase intent.

Our landing page experience was designed for minimal friction. Users clicking on an ad were directed to a page that immediately highlighted current top deals and offered a prominent sign-up form. We continuously A/B tested different headline variations, CTA button colors, and image placements to maximize conversion rates. This iterative testing process was critical, especially in the first few weeks, to gather enough data signals to make informed decisions.

Campaign Performance: Early Metrics and Challenges

The initial phase of the campaign yielded mixed results, as often happens with new product launches. After the first month (January 2026), our overall campaign metrics showed:

• Impressions: 12,500,000
• CTR: 1.1%
• Conversions (Sign-ups): 18,750
• Cost Per Conversion (Sign-up): $4.80
• ROAS (Purchases): 0.75x

While the volume of impressions and sign-ups was encouraging, the Cost Per Conversion for sign-ups was slightly above our target of $4.00. More concerning was the ROAS, indicating that for every dollar spent, we were only generating $0.75 in revenue. This suggested a significant disconnect between user acquisition and actual purchasing behavior. The data signals were clear: we were attracting users, but they weren’t converting into buyers at a profitable rate.

One particular challenge emerged from the attribution analytics. Different platforms reported vastly different conversion numbers, creating a tangled web of overlapping credit. For instance, Google Ads claimed credit for 6,000 sign-ups, while Meta Ads reported 7,500. Our internal analytics, using a first-touch attribution model, showed only 18,750 unique sign-ups across all channels. This discrepancy underscored the necessity of a more sophisticated attribution approach.

We identified that many users were clicking through multiple ads across different platforms before signing up or making a purchase. A user might see a display ad, then later search for “Perplexity Shopping reviews,” click a paid search ad, and finally convert. Both platforms would claim the conversion, leading to inflated individual channel performance reports and a skewed understanding of true impact. It’s a common problem, and one that requires more than just glancing at platform dashboards.

Data Signals and Attribution Refinements

To address the attribution challenge and improve ROAS, we moved from a simple first-touch model to a data-driven attribution model. This model, available within Google Analytics 4 (GA4), uses machine learning to assign fractional credit to different touchpoints in the conversion path, taking into account various factors like position and engagement. This shift provided a much clearer picture of which channels were truly influencing conversions.

We also implemented server-side tracking for key conversion events, particularly for purchases. This involved sending conversion data directly from our server to ad platforms, bypassing potential browser-side tracking limitations. According to a 2023 IAB Global Privacy Report, reliance on client-side tracking alone can lead to up to a 30% underreporting of conversions due to ad blockers and browser privacy features. By implementing server-side tracking, we aimed to capture a more complete and accurate dataset.

Upon analyzing the new attribution data, several insights emerged:

• Paid Search consistently played a strong role in the final stages of the conversion funnel, often being the last click before a sign-up or purchase.
• Social Media (Meta Ads) demonstrated significant influence in the earlier stages, driving initial awareness and consideration. Many users first encountered Perplexity Shopping through a social ad.
• Programmatic Display contributed to brand visibility but had a lower direct conversion rate. Its value was more pronounced in assisting conversions that were later completed through other channels.

This refined understanding of channel roles allowed us to reallocate our budget more effectively. We increased investment in paid search to capture high-intent users and optimized social campaigns to focus on engagement metrics beyond just clicks, such as video watch time and landing page scroll depth. The philosophy was simple: if social was good for awareness, let’s measure its awareness contribution, not just its direct conversion count.

Optimization Steps and Results

Based on the updated attribution analytics and deeper understanding of Perplexity Shopping’s data signals, we implemented several optimization steps during February and March 2026:

1. Budget Reallocation: Shifted 15% of the social media budget to paid search and an additional 10% from programmatic display to a retargeting pool.
2. Creative Refinement: Developed new social ad creatives that highlighted specific product categories and time-sensitive deals, encouraging immediate action. For paid search, we expanded keyword coverage to include long-tail queries.
3. Landing Page Optimization: Introduced dynamic content on landing pages, personalizing displayed deals based on the ad clicked by the user. This increased relevance and reduced bounce rates by 8%.
4. Bid Strategy Adjustment: Switched from maximize conversions to target ROAS bidding for purchase campaigns on platforms that supported it, setting a target of 1.2x.
5. Audience Segmentation: Created distinct retargeting audiences for users who signed up but didn’t purchase, offering them exclusive first-purchase discounts.

These adjustments led to a noticeable improvement in performance during the second and third months of the campaign. By the end of March 2026, the cumulative campaign metrics were:

The improvements were substantial. The Cost Per Conversion for sign-ups was halved, and more critically, the ROAS for purchases doubled, moving from unprofitable to a healthy 1.5x. This meant that for every dollar spent, we were now generating $1.50 in revenue, a significant turnaround. The power of accurate attribution and responsive optimization cannot be overstated here. It’s the difference between scaling a viable business and burning through a budget.

What Worked and What Didn’t

What worked:

• Data-Driven Attribution: Shifting to a more sophisticated attribution model was the single most impactful change. It provided actionable insights that platform-specific reporting simply couldn’t.
• Server-Side Tracking: This dramatically improved the accuracy of our conversion data, ensuring we weren’t making decisions based on incomplete information. It’s a foundational element for any serious performance marketing effort today.
• Aggressive A/B Testing of Landing Pages: Continuous optimization of the user experience post-click directly influenced conversion rates. Small improvements here compound quickly.
• Targeted Retargeting: Tailoring offers to users who had shown intent but hadn’t converted proved highly effective in driving purchases.

What didn’t work as expected:

• Broad Programmatic Display with Last-Click Attribution: Initially, this channel consumed a disproportionate amount of budget relative to its perceived direct conversion impact. Without proper attribution, its true value (assisting conversions) was obscured. It’s not that programmatic is bad, but its role needs to be properly measured.
• Generic Social Media Creatives: Early broad creatives, while generating impressions, didn’t drive strong enough engagement or conversion intent. Specificity, even in awareness campaigns, is key.
• Over-reliance on Platform-Reported Data: Trusting each platform’s self-reported conversions without cross-referencing and applying a unified attribution model led to initial misjudgments about channel performance. This is a common trap, and one that requires constant vigilance.

One editorial observation: many marketers get caught in the trap of optimizing for the metrics platforms want them to optimize for, rather than the metrics that drive actual business outcomes. The shift to data-driven attribution forced us to look beyond vanity metrics and focus on the true value chain.

Future Optimizations and Learnings

Moving forward, our plan for Perplexity Shopping includes further refinement of our audience segments, exploring new ad formats, and integrating predictive analytics to forecast user lifetime value (LTV). We will continue to invest in improving our data signals by enhancing first-party data collection and integrating it with our advertising platforms. Specifically, we’re looking at implementing a customer data platform (CDP) to unify user profiles and enable more personalized ad experiences across all touchpoints.

We’re also exploring the use of AI-powered creative optimization tools to automatically generate and test variations of ad copy and visuals. This will allow for even faster iteration and identification of high-performing assets. The goal is to move towards a more proactive, rather than reactive, optimization strategy, anticipating shifts in user behavior and market trends.

The campaign for Perplexity Shopping highlights a critical lesson: successful digital advertising in 2026 demands more than just running ads. It requires a sophisticated approach to attribution analytics, a willingness to challenge platform-reported data, and continuous optimization based on granular data signals. Businesses must invest in strong tracking and analysis infrastructure to truly understand the impact of their marketing spend and achieve sustainable growth.

Accurate attribution is the foundation of effective digital advertising. Without it, businesses risk misallocating budgets and missing significant growth opportunities. By focusing on complete data signals and advanced attribution models, marketers can achieve a clearer understanding of campaign performance and drive superior results.