How Ramp, Snowflake, and Hightouch run ABM (Dave's Newsletter)

How Hightouch, Ramp, and Snowflake Actually Run ABM

As hot as AI is right now... every time we talk about ABM we get more engagement than most other topics combined.

I don't know why, but it's been that way for four years of running Exit Five now.

I wonder if it's because ABM usually means "we're not just doing random acts of marketing" so it's an attractive strategy. Less chaotic. Or if it's because ABM usually removes a lot of the conversations about "credit" in marketing since you have to be aligned to sales.

Last week I sat down with a group of marketers doing ABM at Ramp (Drew Pinta), Snowflake (Casey Patterson), and Hightouch (Brian Kotlyar) and they shared how their teams actually run ABM. Not the theory, the real stuff. We pulled out four key takeaways for you below.

1. Cap the list before you build the program. 20 accounts per AE, max.

Casey's team at Snowflake fell into a trap that probably sounds familiar. ABM was working. Sales loved the intel. So they wanted more accounts. Her team, half of them former salespeople, kept saying yes.

The result: individual ABM managers covering 200 accounts each.

"That's not ABM anymore," Casey said. "That's targeted demand gen. And it doesn't scale the way we're doing it."

Her fix: hard account limits per person. For major accounts, each manager covers around 20. For prospect-focused work, 2-3 ABM accounts per rep, which comes out to 120-150 per manager total. The exact number will vary by account size and team. The point is to have a cap and stick to it.

At Hightouch, Brian's rule is simpler: 20 accounts per AE, maximum.

"We just don't believe that a human being, even with AI tools, can do a good job working more than 20 accounts."

2. The CRO signs off on who makes the list. Not marketing.

This is where most programs get set up to fail. The list gets built by marketing, handed to sales, and debated for two weeks.

At Hightouch and Ramp, it works differently. Account selection happens at the CRO level. There's data science modeling, enrichment, account research, and then the CRO signs off on who makes the cut.

At Ramp, Drew's team builds a first-pass model scoring fit and intent. That list goes to sales leadership, who whittle it down based on things no model can catch. His example: "This account has a three-year contract with a competitor. Take them off the list and put them back in three years."

Casey put it simply: "There's no amount of intent data that outweighs the fact that a sales rep knows that account intimately."

The list also has to move. Not a spreadsheet you set in January and revisit in Q4. When something changes, the list changes with it.

3. ABM accounts convert to stage one at 32% higher rates. Here's how to prove it.

If you need to sell this to a skeptical CFO or CRO, Brian shared the data from Hightouch's own program.

Accounts in their treatment group (the ones receiving coordinated ABM outreach) convert from unengaged to a stage one opportunity at a 32% higher rate than the general population. Of those that reach stage one, 38% more convert to stage two.

He called it a "double dip effect." Even among accounts that were already good targets, the ones receiving ABM treatment moved through the funnel faster and at higher rates.

The setup that makes this measurable: Hightouch runs a control group alongside the treatment group within the same target account list. Some dream accounts get the full program. Some don't. The difference in progression tells you if it's working.

On attribution: "Fighting over attribution is the biggest waste of time in a B2B company," Brian said. "Just don't do it." Instead, track how accounts move through pre-opportunity stages. Are more accounts going from unaware to lightly engaged? From engaged to an open opportunity? That's the story to tell.

4. You don't need intent data to start. You need a list.

If you're new to ABM or rebuilding a program that's grown too big to manage, Brian's advice was the most practical thing said in the whole session.

You don't need an intent data platform. You don't need a data science model on day one. Every AE you have can probably name 20 accounts they'd love to be in front of. Start there. Get in a room with your head of sales, agree on why each account is on the list, and write it down.

Snowflake's ABM team

That's ABM. The intent signals, the gifting programs, the personalized landing pages come later. They're tactics layered on top of a good list. If the list is wrong, none of it matters anyway.

As Brian put it: "If you want to do ABM for effectively zero dollars and start immediately, just pick good accounts."

The full recording is available here. Casey, Drew, and Brian went deep on a lot more in the Q&A: how to build ad audiences for small account lists, how gifting works at different funnel stages, and what the scrappy version looks like when you don't have a Super Bowl suite budget.

— Dave

P.S. Are you running ABM right now? What's working — and what questions would you ask Brian, Casey, and Drew if you had them for another hour? Hit reply and I'll ask them.

‍