How We Used AI to Increase HubSpot Email Conversions by 82%: A Case Study

By kbodnar@hubspot.com (Kipp Bodnar)

Download Now: The Annual State of Artificial Intelligence in 2024 [Free Report]

We turned our standard nurture email flow into an AI-driven conversion powerhouse. Here’s what we did, what worked (and what didn’t), and what we learned along the way.

When our marketing team began discussing how to strategically incorporate AI into our workflows, we knew we wanted big results. But here’s the thing about big results: They don’t come from trying everything at once.

With limited resources and unlimited possibilities, we needed to hone in on which AI applications would deliver the biggest impact.

Email marketing seemed like a natural starting point for us. We’d been running optimization tests on our nurture flows for years, but after a while, the gains became incremental by a few percentage points here and there.

We needed something that was a total game-changer. Something that had both meaningful influence on top-of-funnel metrics and practical usability across our marketing team. But what — and how?

In a recent Marketing Against the Grain episode, HubSpot VP of Marketing Emmy Jonassen and I share how we experimented with AI to transform our email performance. We’ll also explain how we achieved an 82% increase in conversion rates — plus, all the lessons we learned along the way.

Identifying the Challenge

First, let me explain what we were doing before AI. Like most marketing teams, we approached email personalization through segmentation — grouping leads based on similar characteristics, then tailoring content to those groups.

For example, if someone downloaded marketing-related content, we’d send them more marketing resources rather than sales content.

It wasn’t a bad approach. But it was essentially educated guessing at the group level. We were saying, “People like you typically want this,” rather than understanding what each individual person was trying to accomplish. We wanted to do better than that.

The Hypothesis: Moving From Groups to Individuals

The more we looked at AI’s capabilities, particularly its ability to analyze multiple data points and identify patterns, the more we saw a path to true one-to-one personalization at scale.

So, we asked ourselves: What if AI could help us understand not just what group or cohort someone belongs to but also the specific job they’re trying to get done?

For example, rather than sending marketing content to all “marketing people,” we wanted to be able to pinpoint when a specific marketing manager at a specific company is ready to build their influencer strategy for this specific upcoming product launch. From there, we could send them exactly what they need for that task.

It was a tall order … but we were willing to give it a try.

The Setup: Building our AI Solution

To test our hypothesis, we first designed a process that would enable AI to do what humans can’t: Analyze thousands of individual user intents at scale and craft tailored recommendations. Here’s how the process worked.

When someone fills out a form to download HubSpot content, we collect a few key pieces of information: their business URL, company size, and what …read more

Source:: HubSpot Blog

      

Aaron
Author: Aaron

Related Articles