Shippit masters marketing attribution, boosts ROAS 4X with RudderStack

Highlights:

• Shippit created a Customer 360 in their Snowflake data warehouse using RudderStack Profiles and built an attribution solution on top, solving a problem that had plagued the business for over four years
• Data, Marketing, and every other business team are now aligned around a central source of truth maintained by the Data team in the data warehouse
• Data trust is at an all-time high, and business teams are now confidently using data to make both strategic and tactical decisions

Key Stats:

• Marketing achieved a 4X increase in ROAS
• Marketing is saving over $30k a month in advertising cost savings
• Data built a Customer 360 in the data warehouse with a team of only 3

Shippit is Australia's leading multi-carrier shipping software. The company enables retailers of all sizes to deliver excellent post-purchase experiences with multi-carrier allocation technology that brings together retailers, carriers, and customers. The platform features a streamlined fulfillment workflow and an allocation engine that picks the fastest or cheapest courier for each delivery. It helps retailers simplify their shipping operations, no matter how complex.

Shippit works with thousands of leading retailers across Australia and New Zealand, including Target and Sephora, sending millions of deliveries every month. In 2020, the company expanded to serve Singapore and Malaysia. Their platform regularly drives dramatic results, reducing delivery times, cutting freight costs, and unlocking sales growth through better inventory access and shipping options. This reputation for results fueled rapid growth throughout the past decade, but the growth didn’t come without challenges. As Shippit’s team built a growth engine in flight, they accumulated tech debt that hampered their ability to understand the customer journey and continue efficient go-to-market spending.

Shippit struggled to solve marketing attribution for years before eventually accepting the status quo. That is until Nitt Chuenprateep, Head of Data and Analytics, implemented RudderStack Profiles to build a Customer 360 in their Snowflake data warehouse. Shippit now has an attribution solution running on top of the customer 360. The business is finally aligned around a central source of truth with a high level of trust in their data and the data team. Most importantly, every go-to-market team can now confidently make strategic decisions on where to allocate resources and tactical decisions on optimizing their efforts. The Customer 360 is driving measurable business results.

Challenge: A fragmented attribution nightmare

Shippit’s growth comes through four primary channels: inbound, outbound, partnerships, and events. Their core digital properties include three websites covering two different brands, which all connect to their Ruby application, where signups take place. While this baseline reveals some complexity, it’s the collateral damage from rapid growth that made attribution so difficult to solve.

With four channels and teams moving fast – often using their own tools ad hoc – the system to capture customer data and conversions became fragmented, and it got messy. Over time teams deprecated tools, adopted new ones, and changed workflows. UTM usage, to no surprise, varied across teams and tools. As the company grew, attribution became increasingly important and increasingly out of reach. Nitt Chuenprateep, Head of Data and Analytics explained:

“Inbound might start from the website or a landing page built in Pardot or Unbounce. There were at least three different ad platforms, or you might come in through one of our events tracked in Splash, which interacts with Salesforce. Then there are partner referrals and outbound leads from ZoomInfo or Lusha — so you’re trying to create an attribution model that works across all these different systems over four channels.”

Another layer to the challenge came from the evolution of Shippit’s business model. As the company grew, it moved its focus towards larger enterprises and adopted a named accounts approach. This is why each of the four channels is critical. While paid marketing attribution is important, and a tall challenge in itself, events, partnerships, and outbound now account for a major portion of their business. The disparate systems leveraged by the teams across these channels presented a challenge with greater scope and consequence.

On the paid advertising side, the focus on named accounts made it critically important for Shippit to understand not just which campaigns convert customers but which campaigns convert the right customers. Shippit was spending tens of thousands of dollars a month on ad campaigns that seemed successful on the surface because they converted customers, but they converted the wrong types of accounts – SMB accounts that might only bring in $32 every 30 days, for example.

To effectively allocate resources across channels and optimize spending within each, Shippit needed an attribution solution that could account for the complexity of its data stack and the specifics of its business model. Twice, they brought in the product engineering team to build a solution. Both times, these efforts stalled. Eventually, the business just accepted the status quo.

But the pain persisted, and there were consequences. With what Nitt called “nonexistent attribution,” the Marketing team had little way to understand what was working. Unable to tie conversions to revenue numbers, they couldn’t solve the campaign optimization puzzle because they didn’t have all the pieces. When running experiments, it was difficult for them to determine what actually happened and why. Tensions between teams ran high because there was no “unified view of users to expose ‘this user did this, and it led to that,’” said Nitt. Every team operated and made decisions based upon a different, limited source of truth.

The attribution problem, however, was just one concrete manifestation of a higher-order problem that impacted the entire business: the lack of a solid data foundation. To solve attribution, Shippit needed to model data from across their stack into reliable, complete customer profiles and establish a single source of truth to align teams from every channel. They finally made headway when the Data team took charge and decided to solve the problem at the root.

The Data team set out to make a world where “you are all able to look at everything together,” said Nitt. “And that’s what RudderStack Profiles delivered. Finally, there’s a way for us to bring all of the data together, to marry revenue data from the application with activity data, and to actually govern it from one central source. It’s something we’ve tried to have for years.”

Solution: A solid data foundation unlocks an attribution solve for the entire business

We call unresolved identities a silent data project killer because identity resolution is often the root issue that blocks data teams from completing projects that deliver real competitive advantage. Shippit experienced this firsthand.

The problem behind attribution that frustrated the business to the point of acceptance wasn’t attribution itself – it was a lack of identity resolution across a fragmented data stack. The first step to solving attribution was to solve identity resolution at the root and create a Customer 360 in Snowflake.

RudderStack Profiles made this possible for Shippit – even with a lean team. Using Profiles, Nitt solved the problem that had plagued the company for years, building a customer 360 and establishing a source of truth with a team of just one analyst and a data-savvy growth marketer. All in three months.

Because Shippit uses RudderStack Event Stream, Profiles provided a baseline identity graph out of the box. Nitt and his team then used Profiles to add all of the relevant non-Event Stream data from their warehouse to the Identity Graph and build features on top to produce an actionable customer 360. Instead of writing and maintaining complex code for this, Profiles' declarative approach enabled them to do it by defining the ID graph and feature spec through a simple YAML config. Profiles delivered a major win that accelerated time to value for the small team working on the project without comprising control or flexibility.

It was Profiles that motivated Shippit to revisit the attribution problem. With the customer 360 in place, Nitt knew accurate attribution was within reach. His team could now deliver previously blocked, high-ROI, use cases to their business stakeholders, and attribution was an obvious place to start.

To solve attribution, the Data team built a model on top of the Customer 360 in Snowflake that calculates attribution with data from every source across the entire customer journey. Today, Shippit understands which touchpoints lead to revenue because they’re tying data from touchpoints throughout the user journey into unified user profiles that include revenue data from their application.

The Results: Confidence to optimize every channel

Shippit’s customer 360-fueled attribution model immediately impacted the bottom line. With a newfound understanding of which campaigns convert the right types of users and which ones don’t, the Marketing team quadrupled ROAS, cutting ineffective campaigns and saving $30k per month without affecting throughput. “Now we can tell which users count as conversions according to the ad platform but aren’t actually bringing in any revenue. With RudderStack, Marketing knew exactly which campaigns to turn off,” said Chuenprateep.

Shippit’s Marketing team has more confidence than ever, and not just for paid advertising. RudderStack drives effective engagement across the customer journey. “I rely on attributes in Intercom and custom fields in Pardot/Salesforce that come directly from our Customer 360 via Rudderstack. With daily syncs, I’m confident I’m reaching the right people with the right message,” Lifecycle Marketing Manager, Charlotte Travers, explains. “Ultimately our confidence comes from a trust in our Data team and the systems they’ve set up.”

Shippit is even using Profiles to govern segments for their paid campaigns. The Data team has created cohorts around core business definitions in the warehouse. In collaboration with Marketing, they can build audiences on top of these cohorts to fuel different campaigns from the same data source, ensuring consistency across channels.

Before implementing RudderStack Profiles, data trust was low, and data, or the lack thereof, was a frequent point of contention between teams. Today, the Data Team has established a comprehensive source of truth in their data warehouse that keeps every team on the same page, and data trust is at an all-time high. Because the Customer 360 is driving measurable results, the business teams have a newfound appreciation of the data teams work and the power of data to drive competitive advantage.

Looking ahead:

The impact of Shippit’s Customer 360-based attribution model goes well beyond its inbound channel owned by Marketing. Nitt and his team are beginning to use the new data to understand what’s working for every channel:

• Events: Sponsoring and attending events requires major investment from Shippit, and it’s difficult, especially in new markets, to know which events are worth attending. Now, they can make data-driven event investment decisions based on revenue results.
• Outbound: Nitt’s team can now work with Sales to assess which outbound cadences and messaging lead to high-value conversions on a granular level to begin understanding the value of outbound touches and optimizing outreach accordingly.
• Partnerships: Not every partnership is fruitful. Some, like the partnership with Shopify, are major revenue creators for Shippit, but others drive little value. Using the new attribution model, Shippit can more quickly assess which partnerships are worth investing in.

Attribution was a pain point at Shippit for years due to a formidable underlying problem: Unresolved identities blocked multiple high-ROI projects, including attribution. Previous attempts to address the issue stalled, because it was seen as too complex and time-intensive. But when Head of Data and Analytics, Nitt Chuenprateep, implemented RudderStack Profiles, they were able to build a comprehensive customer 360, providing the foundation to finally solve attribution. Now, every team is confidently using data to drive both strategic and tactical decisions.

Shippit data stack

Destinations: Salesforce, Intercom, Amplitude, Webhook

Sources: JavaScriptSDK, Ruby SDK, NodeJS, PHP, Webhook

Warehouses: Snowflake