Today’s expert shares the data-driven, personalized monthly drip campaign designed to educate, not irritate. And it doubled follow-up actions while cutting unsubscribe rates by 28%.
Thawing Cold Contacts
HubSpot’s marketing team had a list of about 300,000 low-intent small business contacts who received the same generic marketing emails. Some filled out a form. Some poked around on free software. All of ’em got the same spiel.
“We used to try to offer meetings and demos,” says Kelsey Craddock, manager of downmarket automation at HubSpot. “There hasn’t been a really successful way for us to ensure [this cohort] is receiving consistent, helpful communication from us.”
Kelsey and her team put their heads together to rethink the whole process.
“We completely scrapped the old program. What we didn’t take with us was the idea that everybody would be interested in the same thing, [or] that these people want to talk to sales if they’re not doing anything sales-driven. Essentially, we stripped the entire program of anything sales-forward and just focused on delivering value.”
But how do you figure out what’s valuable to hundreds of thousands of people with different roles, industries, and needs? And how do you serve it without making your email team cry?
The team realized AI could help, but not where you’re thinking. Writing 300,000 AI-generated emails would tax their credit cost (and probably the readers’ goodwill). Instead, it did the homework: Sorting contacts and delving into what they care about.
The result was that independent follow-up actions (think form fills, signups, chats, etc.) more than doubled from the old benchmark.
The script behind the drip.
I asked Kelsey what the minimum a marketer would need to replicate this. I simply meant what tools, but her answer cuts to the heart of the tactic.
“What you absolutely need is knowledge of your database: A list of contacts and an understanding of where they are in the buyer journey,” then she takes a beat. “It would be a lot more helpful to have something that connects to your CRM.”
Here’s how they turned those three things into magic.
1. Segment by the job, not the job title.
The first step is not trying to guess who they are and what they like. Instead, it’s determining “a contact’s most likely job to be done.”
To do that, the team gathers 18 signals from within the CRM: basics like job title and department, plus personalized data like pages viewed and past interactions with HubSpot.
AI uses those signals to sort contacts, at scale, into one of 29 marketing or sales roles. That role is written to a contact property called “Job To Be Done” (JTBD) in HubSpot’s CRM, so every list, send, and report can use it.
“It’s divided further by whether they’re a free user actually engaging with our tools, or just a lead engaging with content on our website.”
Kelsey says the secret to this tactic is helping them accomplish that job to be done.
2. Help them accomplish that job.
“This program is all about value. How can we help? Here are resources.”
To do that, the team needed to understand the real challenges each segment encountered.
Our sales team already uses a tool called Gong to create call transcripts for training and quality-assurance purposes. Using an AI tool called Glean, Kelsey’s team analyzed those transcripts for recurring pain points.
What did users struggle with? What content already resonated? How were they using (or not using) their free tools?
To be clear, research was applied across audience segments, not individual contacts. The idea was to shape what content the team created, not to make content for specific recipients.
Those insights enabled human marketers to create six months of education-first emails, each aimed at a segment’s biggest struggles.
“It’s all about meeting them where they’re at and helping them decide what the next step is, so it doesn’t feel so overwhelming,” Kelsey explains. “If it’s marketing, we’re giving them things like the State of Marketing report, so they can learn from other marketers. If it’s sales, maybe we’re coaching them on how to send their first sequence.”
3. Good guardrails make good neighbors.
Kelsey notes that another key to this tactic is to know when not to send an email.
When creating the email automation workflows, her team included suppression lists for open deals, new customers, or recently qualified leads.
In other words, if someone else is talking to them, the monthly drip stays out of the conversation. This ensures that nobody feels bombarded by marketing emails.
“If they’re in a thoughtful conversation with sales, we suppress [the emails] and let the human-to-human interaction do the education.”
4 Tips for Effective Drips
Automation is powerful, but it also makes it easy to go overboard. I asked Kelsey what separates the good, the bad, and the ugly.
Here are her tips for creating an effective nurture campaign.
- Protect deliverability at all costs.
Kelsey emphasizes that this technique is absolutely not about “emailing thousands of contacts off the bat and hoping for the best.”
“You don’t know what you have until it’s gone when it comes to deliverability,” she adds. If too many people flag you as spam, email service providers will believe them.
The team doesn’t blast everyone at once. HubSpot's CRM calculates an engagement score for every contact, then ranks recipients so that emails are sent from highest to lowest. If a red flag pops up, they throttle the send before low scorers ever see an email.
- Don’t ask too much too soon.
“Make sure the resourcing for a program like this isn’t self-serving,” she advises. Instead, it should be “grounded in delivering value and ensuring you have the right content to do that.”
Kelsey’s team tried including content about using HubSpot tools and was surprised to find it performed poorly.
“HubSpot Academy content hasn’t worked very well. It’s too early to drive them there.”
For this audience, general education content was the key to slowly warming up cold leads.
- Put AI upstream, not in the email.
“Instead of putting AI-driven, personalized content directly in the email, we use AI to [determine] what people would be interested in and do really thoughtful grouping.”
Using AI to generate massive quantities of emails would get expensive quickly. Instead, Kelsey’s team focused their credits on finding useful insights.
And by connecting those insights within their CRM, they can use and share that data across campaigns.
- Think about what would move you.
You probably guard your inbox jealously. Your email recipients do the same. So take some time to think about what you would actually like to receive.
“Put yourself in the shoes of someone in your database who hasn’t received any relevant messaging recently,” Kelsey advises. “And look at what content you have that could give value without asking for anything. What would their journey look like? What would they need over time?”
