How ZALORA increased experimentation velocity and doubled its conversion rates with Twilio Segment

ZALORA, part of Global Fashion Group, is the leading direct to consumer (DTC) online fashion retailer throughout Asia. With over 3,000 brands, ZALORA gives customers access to a wide range of products from apparel, to shoes and accessories, to home and living products. Great customer experience lies at the heart of ZALORA's business, providing its shoppers with omni-channel experiences across mobile, tablet, and desktop. ZALORA continues to innovate the online shopping experience by creating new features such as visual search function, ‘Complete the Look,’ and ‘Get the Look’ which allow users to find items that fit their unique style preferences based on AI. 

To stay competitive in the crowded fashion e-commerce market, ZALORA wanted to continue to build on its customer-first business model by modernizing its technology stack. This would enable its team to create consistent user experiences across platforms, something it was previously struggling with. It went through an extensive evaluation process, searching for technology products that could support its 34-item “use case” checklist. 

After completing interviews and in-depth vendor audits, ZALORA emerged with a plan to create a modern data infrastructure. At the heart of that strategy sat a Customer Data Platform (CDP), powered by Twilio Segment. 

Today, ZALORA is able to: 

• Unlock a modern data infrastructure with established user identification and ID resolution 

• Access comprehensive customer data for experimentation and A/B testing 

• Double its conversion rates by focusing on Tier 2 and Tier 3 audiences 

• Increase the velocity of experimentation and A/B testing

• Save time, money and effort of development team resources 

Moving from fragmented data collection to a trusted, standardized approach 

ZALORA knows that its customers expect a seamless experience when shopping, whether they are casually browsing the app or logged-in to their unique user profile on their desktop. Without a CDP in place, this siloed data was extremely difficult to manage, which resulted in a poor experience for shoppers. 

Prior to Segment, user ID management at ZALORA was poorly managed, fragmented and frustrating for Product and Engineering teams. The challenge was that its team lacked the ability to merge user IDs before and after logging in. “As far as tools like Optimizely were concerned, it would be a different person, pre and post login, there was no way of unifying those users,” explained May Chin, Principal Product Manager at Zalora. 

Without the ability to create complete, real-time customer profiles across the user journey and various platforms, ZALORA missed out on a crucial opportunity to provide the most seamless experience for its shoppers. In addition, its internal team members lacked reliable data to run experiments and tests. This was highlighted in the setup of its tech stack. Events and translation data were individually passed to downstream destinations, like Google Analytics and passed again to BigQuery and finally to Tableau for visualization. This was not an ideal data infrastructure. 

Now, ZALORA uses Segment Connections 400+ out of the box integrations to easily collect, unify and connect all of its customer data, to create trusted, dynamic customer profiles across systems. Segment’s anonymous IDs are sent to all of ZALORAs downstream destinations, including Mixpanel and Optimizely, which enable its team to power its tech stack, with the same, reliable data. Going back to the example about customers' login experience, May says, “with Segment in the picture, we are able to fully recognize the user, across the entire journey and serve up a consistent customer experience.”

Segment enabled the ZALORA team to better leverage customer data for experimentation and A/B testing. Its teams can continuously gather standardized, trusted, real time data and optimize its customer experiences. 

With Segment, ZALORA has achieved:

• 250 active audiences built

• 50 traits per profile

• 25 SQL traits

• <10 minutes to build a new audience

• 1440 dev hours/year freed

Improving customer engagement strategies leading to increased conversion rates 

Segment, in combination with Mixpanel, was instrumental in achieving ZALORA’s overarching success metric: doubling its conversion rates. With a CDP in place ZALORA had access to real-time and historical customer tracking data. Using this data, it could deconstruct the entire e-commerce customer journey and begin to understand specific audiences. Using Segment, ZALORA computed four main audiences, built from tracked events, traits, and computed traits. To build these, it focused on a couple of leading indicators such as AOV and click through rates. 

Through audience segmentation, ZALORA has identified Tier 2 and Tier 3 users based on previous purchases and revenue potential, which allowed its teams to focus its efforts on its most valuable customers. “Segment has been really instrumental because it is where we build all of our audiences and pipe that data downstream,” said May. Using audiences and focusing on specific customer tiers allowed ZALORA to understand which levers it could pull and how it could further monetize customers in each of the tiers. 

Making data-driven decisions using reliable experimentation strategies 

Prior to Segment, ZALORA did not have data governance tools in place, which led to inconsistencies in its data-naming and data dictionaries. With Segment Protocols ZALORA created a shared data-dictionary with a live tracking plan. Now, it has access to trusted data that can be used by a wide variety of teams for experimentation. “Back when we didn’t have a CDP, people actively avoided touching analytics because it was so messy and unstructured. Now that we have Segment’s governance tools in place, issues like that are avoided,” says May. ZALORA previously had only 20-40 weekly active users utilizing its analytics stack, with Segment’s clean accessible data, it now has over 100 active data analytics users. This was a key metric its team has accomplished with the help of a CDP. 

With Segment, ZALORA’s experimentation velocity has increased significantly, primarily because of improved event tracking. “With all of the insights we now have at our fingertips we have been able to experiment a lot faster,” Says May, “the Segment and Optimizely integration has been really helpful by reducing implementation efforts; it's pretty much turnkey now and our teams can plug and play.” Zalora can focus on the A/B tests without having to worry about event tracking. It has seen its experimentation abilities scale, with a 10x scale in experimentation volume. 

Perhaps even more impactful, was the data used to make decisions around experiments and features that were not built and launched. May explains this by saying, “With all of the analytics available, we can actually invalidate experiment requests, and save a lot of time by not launching new tests.” Segment has allowed teams across ZALORA, including Engineering, Marketing, CRM and Commercial, to access complete, trusted, real time data with consistent audiences and customer segments. With this data and analytics available to everyone ZALORA has invalidated 195 feature and/or experiment builds, because the data showed they weren’t relevant. As a result, ZALORA saves 1440 hours of dev team support per year by not building these. 

What’s next?

ZALORA continues to take a structured approach to maintaining its modern data infrastructure, with Segment CDP at its core. It started with the high level goal of increasing conversion rates and plans to identify new areas of opportunity moving forward. Now that ZALORA achieved 100% customer tracking coverage, its plans to dive deeper into personalization and “extracting the full value from Segment,” as May describes it.