The 4-Part ABM Framework Behind $2B in Pipeline

Introduction

This session, hosted by Ganesh (Factors) with Steve Armenti: former Google marketer, founder of twelfth agency, and creator of the Howdy one-to-one ABM program, broke down how he thinks about ABM as two disciplines that have to be balanced, not chosen between: the technical, scalable side, and the human, one-to-one side. Steve walked through a real one-to-one campaign built for an actual prospect (down to the AI research and the physical experience it led to), the tiering system he uses to decide which accounts get which treatment, why chasing more signals stopped being a competitive advantage, the one low-effort exercise anyone can run this week, and where he personally draws the line on AI — including a story about AI research that stopped a campaign from going out to a contact who had passed away.

‍

About Steve Armenti

Steve Armenti spent 17–18 years in marketing, including time at Google, before founding twelfth agency, where he runs ABM programs for clients ranging from seed-stage startups to enterprise accounts. He also built Howdy, a one-to-one ABM motion that uses AI-driven research to plan deeply personalized experiences for high-value prospects. He teaches ABM-focused courses with CXL and is a regular podcast and webinar guest on account-based marketing, GTM strategy, and demand generation.

The Two Sides of ABM — Art and Science

• Steve's framework: marketing breaks down into art (brand, storytelling, creative feel) and science (orchestration, data, lists, intent) — and ABM inherits both.
• "We are living and evolving through the most technical marketing has ever been," he said — but that's exactly why the human side needs equal emphasis right now, partly as a reaction to the overuse of AI.
• The human side shows up through events, the sales process, relationship-building, and personalized moments — not just brand storytelling.

Where One-to-One, One-to-Few, and One-to-Many Actually Live

• One-to-many = "targeted demand gen." Untargeted demand gen might mean spending $1M on Google display ads to 100,000+ people; targeted narrows that universe — LinkedIn is a good example of a channel built for this.
• One-to-few = a smaller universe layered with extra effort — retargeting, re-engagement, using the signals and engagement data generated from one-to-many to get more precise and personalized. Steve calls this a good starting point for teams building their signal muscle, since it generates your richest first-party intent data.
• One-to-one = real human hours: deep research, events, sales-team collaboration, personalized gifts or experiences. In Steve's view, "really good true one-to-one doesn't scale because it requires so many humans" — outside of mega enterprise sellers doing billion-dollar deals, most teams can't keep adding headcount and maintain ROI.
• His warning on over-indexing toward scale: "if you overindex on too much scale, you start to sound like everybody else."
• The ROI math is genuinely different, not just smaller: a one-to-many/few campaign targeting 1,000 companies might spend $50K and land 2 customers at $100K ACV — a straightforward 4x ROI. A one-to-one campaign might target only 20 contacts and land the same 2 customers — same output, radically smaller denominator. Steve's take: "sometimes one-to-one is almost like a bit of a shortcut," especially when you have specific accounts you need to land and can't wait for a typical funnel to build.

A Real One-to-One Campaign, Walked Through Live

Steve shared an actual research document built for a prospect through Howdy — a CIO at a large trucking company:

• AI agents were trained to research unstructured web data (podcast transcripts, speaking engagements, executive bios, Reddit, personal blogs) plus social behavior across X, Instagram, and LinkedIn — trying to answer one question: what does this person care about outside of work?
• The research surfaced a 2024 interview where the CIO talked about racing cars in Japan earlier in his life, and that he's now a motorcycle enthusiast living in Portland.
• The team planned a real moment around it: renting supercars at Portland International Raceway for a few hours.
• The outreach wasn't a demo pitch — it was framed as relationship-building from a senior AE, referencing the Motorsports magazine interview directly and offering the experience as a gesture.
• Results: this kind of outreach gets replies "three or four times out of 10 contacts" — often with genuine surprise and appreciation for the thoughtfulness, after which it's on the sales team to run the process well.
• Factors is running its own version of this: 20–25 target accounts selected for a one-to-one motion, with the top 4–5 getting the deepest, most personal treatment.
• Steve's caution on tone: since marketers are a skeptical audience themselves, genuine and thoughtful campaigns land — but "purely commercial" gestures (his example: ads literally offering a Nintendo Switch for taking a meeting) "feel a little cheap," even if they generate a laugh.

How Steve Decides Which Accounts Get Which Treatment

Steve's tiering framework, from cold to hottest:

• Tier 3 — cold accounts: little engagement, minimal reliable CRM contact data, maybe a bit of third-party intent data.
• Tier 2 — accounts with either real engagement or internal importance. On the latter: if sales says an account matters, marketing should treat it as important too, even without supporting data. Steve's advice: "just be the bigger person. Go win them that account and then they'll love you forever."
• Tier 1 — documented engagement, real data, and internal importance together. These are the accounts that earn genuine one-to-one investment.
• The nuance most teams miss: one-to-one, one-to-few, and one-to-many can all apply within a single tier. Tier 1 accounts should still get scaled demand gen — targeted ads, emails, webinar invites — so that by the time a personalized one-to-one motion drops, the account already has some brand recall and the outreach lands harder.

Signals: Why More Data Stopped Being the Advantage

• At Google, Steve's team ran a robust, multi-provider data program with a dedicated data science team, and saw a real, measurable lift in conversion rates on accounts where campaigns and sales motions were built around signals.
• But his current view: "in a world where signals are abundant and AI can synthesize signals rather quickly, it's no longer a competitive advantage to use signals." The advantage now is finding the intersection of the signals that matter most to your market and ICP, combined with the value you actually create.
• His warning: relying on a single signal source (his example — a G2 insight about what a company's customers like) means your competitor can send the exact same message off the exact same data. "Both of you are just showing up as somewhat lazy."
• The real value comes from layering 3–4 signal sources plus your own first-party data (website account activity, CRM and marketing automation engagement) and looking at where they overlap — and how effective that overlap actually is at generating meetings, pipeline, or revenue.
• Doing this manually, even with a data science team, is a real burden — heavy QA, false positives, errors. This is where he sees tools built for signal aggregation (like Factors) becoming genuinely valuable rather than a nice-to-have.

The Lowest-Effort Place to Start

Steve's advice for teams who've never done any of this: start small, not with your whole GTM motion.

• Pick a cohort of just 10–20 contacts. Manually research each one — web research, LinkedIn profile, existing CRM and marketing automation behavior, past outreach, any sales responses. Build an actual understanding of who each person is, not just their title.
• His broader point: marketers have historically been too far removed from sales data. Spending occasional time in the CRM, looking at real lead and contact profiles the way an SDR would, builds a mental model of your ICP as actual people — and makes you a better marketer.
• Once you've done that manually, you'll know exactly what data you have, what's missing, and where to apply it as you scale.
• The follow-through: take that same 10–20 contact cohort and plan something genuinely unique for them — a mix of gifting and experience — then work with sales to track replies and meetings through the funnel.
• The economics behind this: Howdy's research shows the average SDR/BDR spends about 5.5 hours per contact on true one-to-one work (research, sourcing gifts or experiences, planning). For an AE earning six figures, that's a real sunk cost. In a typical 20-contact pilot, Steve is confident in landing 1–2 customers — which, at a $50–100K ACV, works out to roughly 3–5x ROI within about two months.
• Several prospects have taken this approach and built it themselves after learning about it from Steve directly — which he actively encourages.

List Building: Macro vs. Micro

• Macro lists are your total addressable or serviceable market — pulled from third-party data sources, usually owned by a strat ops or rev ops function.
• Micro lists are what matter most for ABM specifically: sub-lists within your TAM that match your tiering and prioritization criteria. These need to stay flexible, since accounts move between tiers as engagement changes.
• Data is the real currency here — take stock of what's available across your CRM, marketing automation platform, and enrichment providers. Tools like Clay were called out as useful for combining multiple data sources into a workable micro-audience for campaigns or multi-channel plays.
• Steve's advice: start smaller, think in cohorts of companies or contacts with similar patterns, based on real data rather than assumptions.

The ICP Mismatch Story

• Steve shared a live example — an active client who was certain their ICP was CIOs. The data said otherwise, and he noted this happens "more than you would think."
• The gap: asking "who is your ICP" gets a different answer than asking "who does sales actually get meetings with," "who is the financial decision-maker," "who is the security decision-maker," or "how does this person interact with the rest of the buying committee."
• A common pattern in software with a demo signup flow: senior stakeholders (CIO, CISO, CHRO) rarely sign up for demos themselves — that's usually a director or senior manager. The C-level contact typically shows up at meeting two or three, mainly to bless a decision that's already been researched and driven by someone more junior.
• Because of this, Steve's team deliberately bifurcates ABM campaigns by persona: C-level contacts get the same core messaging and creative feel, but more thought leadership and value-oriented content, while directors and managers get more tactical, direct-response content built to actually drive demo signups — because the data shows they're the ones doing the research and talking to sales. Brand awareness plays targeting the C-level still matter, so that by the time they show up in meeting two or three, they're already aware and subconsciously supportive.

Targeting a Buying Committee, One-to-One

Two approaches Steve uses depending on the situation:

• Small, high-value account list (e.g., ~5 accounts): target 5–10 contacts per company, spread across levels — directors, VPs, C-levels. The goal here isn't a single ideal contact, it's getting a meeting with the company at all, since deals at this scale (his examples: Boeing, Target, McKinsey-sized companies) are worth playing the long game for.
• Wider, more greenfield motion: spread across more companies with 1–2 contacts each, generally targeting lower levels since they're more likely to respond. For genuinely hard-to-reach C-levels at Fortune 1000 companies, Steve's team has gotten more creative — physical delivery to a known address, or in some cases, showing up in person. "We've done some wild things there."

Where AI Fits — and a Story About Where It Almost Went Wrong

• Steve's personal rule on AI: "if anything came from me that I would have to defend or stand behind to another person, like that's going to be incorrect or false or invented — for me, that's where I'd personally draw the line." He applies the same standard to brand and marketing output broadly.
• He's skeptical of what he calls "first draft marketing" — prompting AI once and running with the output. Even Howdy's AI workflows, after months of QA and technical rigor, still occasionally produce something wrong — which is why a human validates every source, every signal, and every message before it goes out.
• A cautionary story: while researching a list of contacts, one of Howdy's agents flagged that a contact had actually passed away — the agent found his obituary. He'd worked at a large manufacturing company for 25 years, had a family, and was still sitting active in a database. Without that flag, Steve noted, this person would have gone into an outreach sequence despite no longer being alive. "That's just kind of messed up... I'm sure there's lots more of these types of issues with data." The moment became a reminder that AI, used well, can catch real problems — but the same gap "in less equipped hands" could easily go the other way.

Do this today:

1. Pick one segment to audit manually. Go into your CRM and manually audit a couple of lists — a segment or a geo, whatever's easiest to grab.
2. Do the manual analysis first. Work through it by hand until you can start to see an overlap of signals and data, and make sense of it yourself.
3. Optionally, bring in AI to synthesize. If you want, export what you found and drop it into AI to help synthesize and find patterns.
4. Turn it into one bespoke play. Use what you found — the demographics, firmographics, signals, and patterns — to plan an ABM play or campaign that's unique to those specific individuals you researched.
5. Design for how seniority engages differently. Build the experience knowing C-level contacts engage differently than managers and directors.
6. Get it on paper and see what happens. See what works well, see what flows well, see where you get stuck, and where you might need more resources.

Why this matters: This gives you a good foundation for building your ABM program on your own ingenuity — instead of just taking what the market says, or what AI or research says.

‍