The goals

Grabr, the global peer-to-peer shopping and travel marketplace, offers users a friendlier way for people to shop for international purchases around the globe.  At a time when international travel has grown to 1.3 billion flights per year—a 40 percent increase over the past decade—the ease of buying goods across borders hasn’t changed. High fees still dissuade many shoppers from making transnational purchases.

The Grabr marketplace solves this problem by connecting people who want to buy unique, hard to find products, as well as save money by connecting with travelers who can make money by buying and bringing products on their trips, as well as meet locals. Travelers have access to more affordable and meaningful travel experience by monetizing extra space in their suitcase. Shoppers receive the products they want while saving on tariffs, and Grabr monetizes this commerce by taking a small percentage of each transaction.

“We provide shoppers access to things they can’t get locally and travelers get subsidized trips so they can travel more often,” said Christina Leigh Morgan, Marketing Director at Grabr.

As the first platform to systematize peer-to-peer international shopping, Grabr is exploring uncharted territory. Operating a multi-sided marketplace that delivers goods to the other side of the world safely, quickly, and affordably without a logistics network is a complex undertaking. But that hasn’t slowed Grabr’s growth: Since its launch in 2015, Grabr has amassed over 1 million people on the platform, in seven primary markets. Travelers have made over $1.5m in their personal rewards delivering products to 59 countries.

“When you’re building something new, there are lots of unknowns,” said Christina. “It’s very easy to make assumptions around what people want, the prices they’re willing to pay, and the destinations they want to visit.”

This is where Mixpanel came in: “We needed a way to eliminate the guesswork, and that’s essentially what Mixpanel does for us,” said Christina.

Grabr deployed Mixpanel across its desktop and mobile apps shortly after launch as a way to measure its users’ behaviors. Mixpanel gave them the data to identify the metrics that would be most important to their growth: gross merchandise volume (GMV), user growth, and a healthy ratio of travelers to shoppers, or as they call it, liquidity.

But understanding the data isn’t just about knowing the topline business metrics. Taking these user insights drives Grabr’s innovation, product development, resources allocated to certain parts of the business, and the data-backed confidence to know what regions Grabr should expand to next.

“To achieve our growth objectives, we knew we’d have to let data guide our decisions,” said Christina.

The solution

Grabr chose Mixpanel because it was the most flexible platform that provided user insights so that its product, marketing, growth, and customer support teams can access to real-time data to make smarter decisions.

“So many companies choose Mixpanel because it’s a fast and flexible platform. Mixpanel doesn’t slow down when you accumulate lots of data like other platforms do. And unlike Google Analytics, which only displays sample data, Mixpanel gives you the whole dataset so you can actually rely on it to make decisions.”

With Mixpanel Core Reports such as Segmentation, Funnels, and Retention, the team queries user behavior from billions of tracked events as users and their devices circumnavigate the globe.

“To achieve our growth objectives, we knew we’d have to let data guide our decisions,” said Christina.

“Our most important metric is liquidity, which is defined by the percentage of people who publish a ‘grab’ (e.g. request for a product) and then the number of those requests that actually gets delivered.”

“Being able to track number of transactions on Mixpanel and then segment deeper to understand the unique users who transact (and continue to transact) on Grabr means that we have efficiency for both the members of Grabr, but also for our marketing team, too,” explained Christina. “We’re getting people to do the same purchase or delivery again and again because they find value in the marketplace, which is exactly what we want to see to drive the business.”

With Mixpanel Funnels, the team can see precisely how travelers and shoppers move through their user journeys. And where there are drop-offs, the team can make changes.

“Funnels is useful because it allows us to look at the unique number of people who took an action such as visiting the homepage, and then completed another action, such as publishing a grab or accepting an offer,” explained Christina . “Then we can filter it down to something as specific as the behaviors of shoppers from a specific city.”

The result of knowing what actions users take and in what order is that the team can eliminate steps in the journey and better allocate their marketing budget. “Once we understand user journeys, we can flip over to Retention to look at how we’re retaining each new weekly cohort. It lets us see where our feature changes and acquisition spend pay off.”

With A/B testing and Messaging, the team can try new variants of the app’s interface and nudge users to transact repeatedly with user-targeted marketing down different user paths.

“With Mixpanel, we can isolate user trends and nudge users in order to take a specific action, whether that’s to post a grab or shop again. By sending relevant communications tied to a user’s actions, we are more efficient in our marketing and more effectively encourage our users to re-engage with the platform.”

“For instance, if a user has a pending offer and hasn’t accepted for two days, the team can automate reminder emails. And they can use Properties to segment the emails so different personas of users get different versions. It works so well we now run all of our email and push notifications out of Mixpanel.”

Saphh Li, product and growth expert at Grabr also shared how the marketplace capitalizes on user insights to find new areas of growth. Access to specific user data allows the team to market and serve each segment of users differently, too. “We might find that Argentinian shoppers need fewer nudges to post a ‘grab’ than Brazilian shoppers do,” said Sapph. “That type of targeting increases our conversions with each group.”

“We have so many different markets,” said Sapph Li, product and growth expert at Grabr. “It’s beyond important that we get precise market data. Countries may be relatively close in proximity but may function differently as markets.

“With Mixpanel, we’re able to create localized reports for different regions, and then build different features, automated support documentation based on these user insights. With these user insights we’re able to serve diverse users from different regions and help them move through the platform more easily.”

Mixpanel usage now extends well beyond the few members on the product team. “It used to be just a handful of people in Mixpanel. But now, almost every team member gets access and basic training,” said Christina . “We think it’s important for everyone to be able to answer their own questions with data.”

“With Mixpanel, we can isolate user trends and nudge users in order to take a specific action, whether that’s to post a grab or shop again."

Understanding customers to personalize their language and support

Early on, Mixpanel helped Grabr pinpoint where travelers and shoppers lived, and where their trips initiated and terminated. This helped the team figure out which markets to prioritize, which languages to offer, and how to explain the shopping process to new users.

“In the beginning, we had a hunch that all of our shoppers would be U.S. citizens visiting markets like Argentina, Brazil, or Russia,” explained Sapph. “After digging through user trends with  Insights, we were surprised to find that we had it all backward. Many of the people making offers—that is, agreeing to pick up items for shoppers—lived in Latin America, and were bringing products back from the U.S.”  

This understanding led the team’s focus on localizing languages, especially on the traveler side. They added more educational content in those languages, as well as explanations of the U.S. customs system.

“Customs is different for shoppers who are just visiting the U.S. as opposed to those whose travel originates in the U.S.,” said Sapph. “We had to compensate for the reverse commute.”

The revelation that shoppers wanted goods from the U.S. led the Grabr team to invest more resources into studying international markets. In Brazil, for instance, they found that “it’s common for people to wait for family or friends to travel and bring things back for them,” said Grabr’s marketing manager in Brazil, Michelle Chahin. “We are offering technology for something that comes naturally for Brazilians.”

User insights into Brazilians and others allowed the Grabr team to select and prioritize its six most active markets, and analyze the user cohorts within each in even greater detail. After two and a half years of using Mixpanel insights to rearrange onboarding sequences, move buttons, and track user journeys, they increased their GMV by 117x.

2x higher conversions for shoppers who were referred

The Grabr team was curious about their referral rate, so they created a weekly report for the percentage of GMV that was coming from referrals. As they suspected, it was significant, but it led to a long, drawn-out internal conversation about how they might increase referrals. Some on the team simply wanted to promote the referral system as-is, but Christina, a long-time growth expert and former co-founder of a referral marketing software company, wasn’t so sure.

The team logged into Mixpanel and used Funnels to look at the health of their referral system. They compared two user funnels: One for all shoppers who signed up and later transacted, and another for shoppers who were referred, signed up, and transacted. They found that referred users were indeed more valuable—they were 7x more likely to convert. But when the team looked at the entire funnel from signup to order completion, they saw that it wasn’t very efficient.

The actual signup rate from people who were referred was fairly low. Users at that stage weren’t outperforming users of other groups.

“These were highly qualified users who should have been converting at a higher rate. It didn’t make sense,” said Sapph. Had the team simply tried to increase the overall rate of referrals, they would have been pouring leads down the drain.

“Insights from Mixpanel really changed our thinking. We had to become more efficient, not simply use brute force,” said Christina. “Unlike newly acquired users, referrals are a limited pool. If people keep referring friends but none of those friends sign up and that friend doesn’t get a referral bonus, they’ll get discouraged and stop.”

The team used what they learned from the referral funnels to begin testing and improving the stages where conversions were low. “We immediately made a series of modifications to the referral landing page. We A/B tested placements, fonts, and colors against each other,” said Sapph.

These data-driven iterations, and measuring their impact over time, doubled the percentage of people who actually signed up from referrals.

“That’s huge,” said Sapph. “That’s a really impactful stage of the funnel.” Once the referral funnel became more efficient, the team could promote the referral program with confidence because they knew that more new users would end up transacting.

“We’re always looking for points of leverage like that,” said Christina. “Helping us discover those areas of opportunity with user insights is one of Mixpanel’s strengths.”

Developing a culture of experimentation to drive results for different user personas

As the team began to collect more and more data on user behaviors, they found that the most expedient way to conduct tests was through Mixpanel Messaging.

“Having Messages is really important for our experiments,” said Sapph. “If we want to test whether we can motivate a certain behavior by offering an incentive or a reminder, we can do it right from the same interface.”

Some of the test messages are automated. For example, the team has set automatic reminders for shoppers who don’t receive an offer on their grab within a certain number of days. Others are one-off.

“All the messages we send with Mixpanel are some type of nudge,” explained Christina. “Whether it’s a customer’s receipt with their breakdown or a test to see what happens if someone gets an alert.”

Having one centralized repository for all user data makes robust testing possible. “We can trust our segmentations because we have their country, language, previous history, how many orders they’ve received, and more, all in Mixpanel,” said Sapph. “We can create different messages for each audience so the messages are more relevant and more interesting.”

One lesson Grabr has learned is that every campaign it runs must be fundamentally two-sided. Travelers and shoppers exhibit very different behaviors and Mixpanel allows the team to always create a different version of each message for each.

The power of combining data sources

Grabr’s teams also benefit from seeing unified user data in Mixpanel.

For example, with Mixpanel Platform, Grabr uses its bidirectional integration to Zendesk to examine customer support data and product data together with Mixpanel’s powerful data visualization and reporting. That way customer support agents are more informed, empathetic, and resolve tickets at a faster rate.  

“These powerful integrations help us make better decisions about where we scale our customer service team or where we automate our support efforts,” said Christina. “Through testing, we learned where exactly we should place a product in our notifications flow to better educate new shoppers about their frequently asked questions, which has reduced the number of inquiries our support team gets.”

“By tracking and measuring the types of questions we receive in Mixpanel, we’re able to know where we need to invest support resources and what education content we need to build out. With more automated support content, we see our time-to-resolution decrease and customers receive an answer right away, moving forward in their purchasing journey.”

“We can filter metrics by things like ticket close times, by order history, by region, or by country,” explained Christina. “With Mixpanel, we always know exactly where we need to add support resources.”

“Many of our agents rely on user data when resolving tickets in their day-to-day work.

With Mixpanel, our Support agents can access a comprehensive history of user activity to quickly understand the problem they’re experiencing and how the agent might close tickets faster, without asking too many unnecessary questions. Overall, this creates a more empathetic customer experience and helps our product team know of the most popular feature requests.”

What's next?

With user insights, Grabr is able to streamline its platform so people can shop the world in a friendlier way. When the team has the insights it needs to track individual user journeys, identify drop-offs, and make the journey easier, travelers get cheaper travel, shoppers get cheaper goods, and people who would never have otherwise met get to connect.

“We have lots of stories of travelers who have gotten coffee with their shopper, who then offers them unique travel tips,” said Christina. “There’s a human element to all of this, and it’s part of what makes this job great.”

With Mixpanel, Grabr is able to understand the complex behaviors of multiple segments of users on a multi-sided platform. It’s able to measure, track, and improve its GMV, acquire more users, and boost its liquidity.

In the future, Grabr will use Mixpanel insights for even more effective data-driven innovation. It plans to ramp up testing, provide even more proactive customer support, and deliver on its promise of making the world more accessible by delivering goods safely and securely to the other side of the world.