10 December 2024

The AI Strategy That Doubled His Email Conversion Rate (Step By Step)

2Frameworks

6Insights

Frameworks in this episode

Strategy7 steps

AI Use-Case Prioritization 2x2

Rank AI ideas by customer impact and team-wide productivity potential

Read →

Marketing7 steps

Job-to-Be-Done AI Email Personalization Pipeline

Infer each lead's job and match it to the most useful available content

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster31:30

Personalized Copy Wasn't the Main Source of the Conversion Lift

The team's first iteration assumed that making the email text more personalized would create the step-function improvement. Testing showed that the decisive factor was recommendation relevance: accurately identifying the recipient's job to be done and surfacing content that could help accomplish it.

The initial hypothesis emphasized personalized output
Testing separated copy personalization from recommendation quality
Job inference proved more important than clever wording
Relevant content created the strongest conversion effect

“what we found was really where the step function conversion rate improvements came is not necessarily in how personalized this text is”
Emmy Jonathan · 32:00

“that's where the magic is”
Emmy Jonathan · 32:00

#personalization#jobs to be done#email copy#experimentation

Hot Take· 1

Hot Take23:30

A Deep Content Library Becomes an AI Distribution Advantage

HubSpot's recommendation system benefits from having a large catalog of courses and educational material that the model can match to individual needs. The broader point is that companies which consistently create useful customer education give AI more valuable assets to retrieve and distribute.

AI recommendations need useful assets to recommend
A larger relevant catalog creates more potential matches
Existing educational investment compounds through AI retrieval
Customer value creation improves the usefulness of AI distribution

“the more you are focused on educating creating value for your customer the easier working with AI to deliver that value is going to be”
Kip Bodnar · 24:00

#content strategy#ai advantage#customer education#content library

Explainer· 1

Explainer13:00

The Three Levels of Intent Inside HubSpot's Demand Engine

HubSpot distinguishes people who want to buy software, people who want to try software, and people who want to learn about a related topic. Educational intent represents the largest audience but requires nurturing before those leads are ready to evaluate or purchase software.

Buying intent is expressed by raising a hand to speak with sales
Trying intent is served through freemium acquisition
Educational intent is served through content and nurturing
The learning audience may be more than ten times larger than the evaluating audience

“there's at least 10x more people in your Market out there learning than they are evaluating”
Kip Bodnar · 15:00

#demand generation#buyer intent#lead nurturing#freemium

Story· 1

Story24:00

The AI Email Test That Lifted Conversion by 82%

HubSpot's AI nurturing experiment produced a large step change after years of diminishing returns from conventional optimization. The team double- and triple-checked the result, while improvements across opens and clicks suggested that recipients also found the emails more valuable.

Conversion rate increased by 82%
Open rate improved by nearly 30%
Click-through rate improved by more than 50%
Multiple engagement gains supported the conversion result
The team verified the unusually large lift before accepting it

“our conversion rate increased by 82%”
Emmy Jonathan · 24:00

“open rate I think was close to 30% maybe just a just a hair Ender clickthrough rate over 50% conversion rate over 80%”
Kip Bodnar · 25:00

#conversion rate#email marketing#experiment results#engagement

Takeaway· 2

Takeaway10:30

Why AI Products Must Ship Before They Can Become Great

The speakers argue that an AI experience cannot be perfected entirely in development because its weaknesses emerge through real interactions. Teams should release a sufficiently trained version, set expectations, collect feedback, and improve it iteratively rather than waiting for theoretical perfection.

Real-user behavior supplies essential training feedback
Waiting for perfection delays the data needed to improve
Early expectations should acknowledge that the first release will not be great
Iteration turns an adequate launch into an optimized experience

“until you actually release it into the wild you're not going to be able to get it to Perfection”
Emmy Jonathan · 10:30

“get it out quickly and then iterate and train and perfect”
Emmy Jonathan · 11:00

#ai products#iteration#shipping#user feedback

Takeaway33:30

Pair AI Builders With Workflow Experts

The experiment combined an AI specialist with an email-automation expert who understood first-conversion nurturing and HubSpot's personas. This pairing joined technical model capability with the operational and customer knowledge needed to make the system useful in production.

AI expertise covered model and system implementation
Email expertise covered automation and nurturing mechanics
Persona knowledge helped shape relevant behavior
Cross-functional pairing was treated as essential to execution

“you have a marketing uh someone who is just a subject matter expert in the automation work the Persona the email and then you have…”
Kip Bodnar · 34:00

#team design#ai implementation#marketing operations#domain expertise