Trade Me sees 20% campaign performance improvement using new tools

The big win

With increasingly heavy business demands on their data science team, Trade Me's marketing team needed a solution to enable them to rely less on data science resources and empower them to solve their own data science challenges.

Using CustomerAI Predictions, Trade Me's marketing team was able to easily leverage flexible, out-of-the-box predictive AI models to predict buyer and seller behaviour, then use those predictions to run more targeted, timely, and personalized campaigns without needing a dedicated resource from their data science team. Armed with AI Predictions of which users were most likely to bid on an item being auctioned on their marketplace, make a purchase on an item being auctioned, or even create a listing for a specific product category—like "Computers" or "Motors"—Trade Me's marketing team drove a 20% increase in campaign open rates, a 10% increase in campaign click-through rates (CTRs), a 1.3x increase in seller listings for computers, and achieved 2-3x better campaign performance than they did with ad audiences using a leading ad platform's AI technology.

Since implementing Segment, Trade Me has "unified our customer data from multiple devices into complete customer profiles and successfully migrated from our DMP to Segment CDP without negatively impacting our ad monetization and while improving our audience targeting capabilities, which we see as a key success," said Greg Varley, Programmatic & Data Manager at Trade Me.

As the largest online auction website in New Zealand, Trade Me is "the place you go to find your first home, launch a new business, purchase a car, or take the next step in your career," with over 650,000 Kiwi visiting the site every day. 

Although Trade Me has a data science team that supports marketing use cases and experiments, increasing demand for data science across the business has made it more difficult for the marketing team to get the resources it needs.

"Every time we have a use case," Vaibhav Miskeen, Senior Analyst of Customer Data & Martech, said, "it gets prioritized based on how it contributes to the overall objectives of the business. Although marketing drives revenue, other important initiatives were being prioritized, so we looked for a solution where we rely less on data science resources and more on platforms to do the hard work for us," said Vaibhav.

Twilio Segment has become a critical platform the team now relies on, providing Trade Me with unified customer profiles and a trusted data foundation while also putting the power of predictive AI directly in the hands of their marketers. Leveraging the out-of-the-box predictive models within Segment's CustomerAI Predictions allowed the team to quickly predict future buyer and seller behaviour then run targeted campaigns across channels without needing additional data science resources. As a result, the team recorded significant uplift in their email and ad campaigns while improving return on ad spend (ROAS) and operational efficiency. 

Boosting campaign performance and ROAS with predictive AI 
Using CustomerAI Predictions, Trade Me's marketing team were able to easily predict which users were most likely to engage and convert in different ways that drive sales on their online marketplace. 

With the flexibility of Prediction's out-of-the-box Custom Predictive Goals model, they could identify which users were most likely to bid on an item being auctioned on the marketplace, make a "Buy Now" purchase on an item being auctioned or make a purchase from their "Shopping Cart" in the next 30 days. They then used these predictions to send targeted and timely email campaigns to boost these engagement and conversion metrics. Measuring email metrics as well as website and app sessions per email sent, they observed an uplift in which fewer emails were sent, but more sessions were generated by targeted audiences. Additionally, the team recorded a 10% improvement in click-through rates (CTR) and over a 20% increase in open rates for campaigns built with CustomerAI Predictions.

On the seller's side of the business, the marketing team used the Custom Predictive Goals model to determine the likelihood of casual and pro sellers to create a listing and, more specifically, their likelihood to create a listing for specific product categories, like "Computers" or "Motors." Using these AI-powered Predictions to target promotional campaign offers to casual sellers who were most likely to create a listing resulted in a 1.3x increase in listings in Trade Me's "Computers" category.

When Vaibhav's team tested paid campaigns to casual and pro sellers that Segment determined had a high likelihood to make a "Motors" listing versus the audiences they built using a leading ad platform's AI technology, the audiences built with CustomerAI Predictions delivered 2-3x higher return on ad spend (ROAS).

Migrating from a DMP to a first-party data strategy with a real-time CDP
When Greg Varley joined Trade Me as Programmatic & Data Manager in 2015, the company was using a data management platform (DMP) to track customer behaviour on its website. However, due to limitations in their DMP, Trade Me had: 

• User data is siloed across platforms, making it challenging to piece together information unified profiles and gain a deeper understanding of their audiences.
• Data signals are only being captured as long as the cookie lifespan, meaning they couldn't get a true longitudinal view of the customer.
• To rely on big tech companies, like Apple and Google, to maintain their current policies in order to track their users at a device level. For instance, if Apple stopped businesses from being able to track identifiers for advertisers (IFDAs), companies like Trade Me would risk losing a large portion of their audience data, which is valuable to advertisers on their platforms.

In 2020, they began making the shift to first-party data and quickly identified the need to make the switch from their DMP to a customer data platform (CDP).

While they understood this shift presented an opportunity for them to know their customers better, offer more personalized experiences, and improve marketing ROI, they needed to assess how this migration would affect their business and identify a solution that would be the best fit for their needs. 

"Some providers had a good offering, but they didn't really suit Trade Me's use case. We chose Segment because we knew it could provide the robust privacy controls, streamlined data collection, and precise audience targeting we needed to support all our advertising use cases," said Greg.

According to Greg, Segment placed a large emphasis on data governance and provided tools to ensure data was being brought into the platform in a way that protected Trade Me's downstream use cases. Greg added, "Without foundational consistency in the data, the outcomes we're trying to deliver to our customers and users would not be as strong, efficient, or effective as they are."

Trade Me successfully migrated from their DMP to Segment CDP, future-proofing their data infrastructure, protecting their annual advertising revenue, and unlocking new growth opportunities with CustomerAI.

Enhancing security by limiting access to user data
Trade Me maintains its commitment to data security and user privacy by restricting access to the user interface with Personal Identifying Information (PII) during the migration to Segment. Only a select group of authorized individuals are granted access. Other Trade Me team members can access the platform but with redacted user data.

Trade Me also has governance that dictates who has access to and permission to use the user interface (UI) and create audiences within Segment.

By tightening its processes and having governance around who can access the Segment dashboard and use data, Trade Me has become more efficient and effective in building audiences for its advertising strategies.

Expanding the use of CustomerAI Predictions to improve ad targeting and scale marketing efficiency
The team will continue using Segment's out-of-the-box CustomerAI Predictions to scale their search engine marketing (SEM) efforts—aiming to better optimize their ad spend by focusing budget on users with a high propensity to buy while minimizing budget or suppressing ads to users with a low propensity to purchase. They plan to test CustomerAI Predictions-powered audiences across their Meta, and Google retargeted ads to drive more efficient ad spend and greater ROAS.

"The marketing team has been making very good use of CustomerAI Predictions, and we are keen on continuing to use it to drive campaigns across digital channels, including Facebook, Google Ads, and email," Vaibhav noted.

With Segment, Trade Me can now:

• Use out-of-the-box CustomerAI Predictions to improve campaign performance, grow ad revenue, and optimize ad spend
• Get a complete understanding of customers across all channels and devices
• Future-proof their business by moving away from a DMP that relies on third-party cookies to a first-party data-powered customer data platform
• Target audiences across iOS devices, which they were unable to do with their DMP 
• Increase operational efficiency by solving marketing challenges without the need for data science resources 

"We have a much deeper and more complete understanding of our audience now that we've made the move to Twilio Segment." - Greg Varley, Programmatic & Data Manager at Trade Me.
Results

• 20% increase in open rates for campaigns built with CustomerAI Predictions
• 10% increase in CTR for campaigns built with CustomerAI Predictions
• 1.3x increase in seller listings in the "Computers" category leveraging CustomerAI Predictions
• 2-3x better campaign performance with CustomerAI Predictions vs a leading ad platform's AI technology

Want to bring together clean, consented customer data for real-time insights to help you know each individual like they are your only customer? Learn more here.