Sage cuts cost per conversion by over 30% with Twilio Segment

Challenge

Fragmented customer data across marketing, web, email, and paid media made it difficult for enterprise software provider Sage to respond quickly to onboarding progress or upgrades, often leading to wasted ad spend and irrelevant messaging.

Solution

Sage rolled out Twilio Segment as the backbone of its customer data strategy to bring together customer activity across product and marketing touchpoints and support event-triggered onboarding and engagement journeys.

Introduction

Sage helps small and mid-sized businesses remove administrative burden and focus on growth by providing accounting, human resources, payroll, and financial software. Before these tools can support their operations, customers need to set them up and actively adopt them. When they don’t, they may churn or underuse features that could help their businesses manage costs and stay compliant.

This is why increasing product activation and everyday usage is a priority for Sage’s marketing and direct channel teams. “We want to make sure people are activating and adopting the products they’ve purchased so they can get the full value from them,” said Richard Fahey, Enterprise Architect at the company.

It was difficult to respond to what customers were actually doing in the product or on the website in near real time because customer data lived in separate tools used by marketing, web, email, and paid media teams. Consent data also resided in various systems, which made it difficult to apply preferences the same way across all channels. Identity resolution was likewise inconsistent.

This disconnect showed in the customer experience. For example, a prospect who converted could still see paid ads for days despite already visiting a Sage product page or signing up for a trial. Similarly, a customer who upgraded might still be in a nurture flow meant for first-time buyers.

Addressing Manual Audience Creation and Engineering Bottlenecks

Building marketing lists at Sage followed a familiar pattern. The marketing team would request a list of customers and prospects, and data engineering would write SQL queries to pull it together. The list would then be passed back for review, and requested changes would often be sent back for another round of revisions. Ultimately, sales would call through the final list.

As Fahey put it, “Marketing wants a list to produce another list, but somebody will evaluate that list to produce another list so salespeople can pick up the phone and call people.”

In addition, every audience refresh needed an engineering ticket. Manual extraction, transformation, and loading could take 10 hours per audience, and some requests were in the queue for two or three weeks because of limited engineering capacity. Campaign schedules had to work around those delays, with audiences refreshed monthly and suppression lists often lagging recent conversions or onboarding progress.

Sage had previously implemented another customer data platform before rolling out Twilio Segment to bring together customer data and improve segmentation. However, adoption by marketing and engineering stalled and skepticism persisted from the earlier deployment, as it had not delivered the consistency the business expected.

Working within the enterprise architecture function, Fahey and a specialist team focused on scenarios where Segment could quickly show value, particularly for onboarding journeys and event-triggered engagement.

Hands-on Workshops for Marketing and Engineering

Sage began running on-site hackathons with marketing, engineering, and data teams working together for two or three days on well-defined use cases such as automated churn signals. Instead of submitting requests to a service desk, marketers could work directly with engineers and test ideas.

“If you have the right people with access to the tools and systems, you can get months’ worth of work done in three days,” said Fahey.

One experiment involved consuming a dataset in hourly chunks to automate a churn model. It did not work as planned, but the learning helped clarify how warehouse data should feed downstream tools. Gradually, marketing and data teams learned to create audiences without waiting for engineering support.

Marketing now spends less time generating static lists and more time building customer journeys that respond to real behavior. Even when something breaks, the process still helps marketing and data specialists find gaps in identity resolution, consent, and event architecture.

Combining Web, Product, and Paid Media Signals

With Segment in place, Sage connects behavioral and transactional signals from web activity, product usage, CRM systems, and backend tools. Data now feeds directly into activation tools in the Twilio stack, including SendGrid.

Identity resolution through Segment Unify brings together anonymous and known histories into a single customer profile, while Profile Sync makes those profiles available to downstream tools such as paid media platforms and marketing automation systems. As a result, paid media and email platforms work from the same underlying customer record. Because Segment and SendGrid work natively together, Sage can activate audiences in email without complex integrations or custom connectors, reducing setup time and overhead.

Warehouse data from Snowflake can be cleaned and pushed back into activation tools using reverse ETL (extract, transform, and load). Consent preferences from sage.com are updated within 24 hours, instead of weeks, reducing the risk of contacting customers who have changed their preferences.

Audience refresh rates moved from monthly to hourly, with the option to refresh as often as every 15 minutes if needed. This cuts the delay between customer action and campaign response. “We calculated that if you were to refresh an audience each time and to load that manually into an activation system, that’s 10 hours of work we’ve saved per audience,” added Fahey.

Faster Onboarding and Reduced Wasted Ad Spend

Campaign onboarding is now more than three times faster. Marketers can build suppression lists without engineering tickets while paid media audiences update automatically when a prospect becomes a customer.

More relevant targeting cut wasted ad spend and improved segmentation in platforms such as Google, LinkedIn, and Facebook. This contributed to a 37% reduction in customer acquisition costs and a 27% increase in conversion rates.

Sage is experimenting with personalization in Segment to support customer onboarding and upgrades. “Ideally, we’d be able to tailor each customer journey based on their behaviors, whether it’s through email, product interactions, paid media, social, or voice,” said Fahey.

“We can’t do it manually, and we need the automation to support our operations as we grow.”

Sage also plans to expand event-triggered journeys that respond within specific time windows, rather than waiting for scheduled campaigns.

“Our goal is to deliver messages that feel personal and valuable because they’re based on real-time data, not just pre-set flows,” said Fahey. This includes win-back campaigns for stalled onboarding and product recommendations linked to patterns in usage.

Fahey shared that data privacy and consent controls were among the early visible wins. “Our lawyers always love the data privacy box function of Segment. They can see that data privacy concerns are satisfied, that a choice has been made and, if needed, reconfirmed historically,” he added.