How to Build High Performing ABM Audiences from Scratch
When people talk about “ABM audiences,” it often sounds like a neat, finished product you can just buy, upload, and turn into pipeline. In reality, most ABM audiences are a patchwork of assumptions.
A list from last year. A handful of “target accounts.” A few intent signals. A layer of job titles. A prayer. Then the campaign launches, performance looks “fine,” and nobody can explain why a handful of accounts are surging while most are dead silent.
That’s the real problem ABM teams run into: you can’t optimize what you can’t clearly define.
High performing ABM audiences don’t start with platforms or tactics. They start with clarity—about who’s actually in-market, who can influence a deal, and how your targeting logic holds up once it hits the messy reality of B2B data. This is how to build ABM audiences from scratch in a way that performs, holds up under scrutiny, and gets better over time.
Start with a hard truth: “Target accounts” aren’t an audience
A target account list is a business decision. An ABM audience is an operational asset.
The account list says: we want these logos. The audience says: these are the specific people we can reach, measure, and learn from inside those logos.
Most ABM underperforms because teams confuse the two. They select accounts and assume the audience “just exists.” But inside any account are multiple teams, competing priorities, internal politics, and wildly different levels of relevance to your product.
If you don’t decide who matters—and why—your audience ends up being whoever your platform can find. That’s not ABM. That’s randomized B2B reach with a nicer name.
Build from the inside out, not the outside in The common impulse is to start broad:
“Let’s target IT directors in healthcare companies over 1,000 employees.”
That’s outside-in. It’s simple, scalable, and usually mediocre. High performing ABM is inside-out. It starts with a few grounded truths about real deals:
● What triggers the purchase conversation?
● Who initiates it?
● Who blocks it?
● Who signs?
● What has to be true internally for budget to move?
Your audience should reflect that internal reality. If you can’t say, in plain language, why a person is in your audience, you don’t have an audience. You have a guess.
Step 1: Define the “deal unit,” not just the persona
Personas are helpful, but they’re often too fuzzy to build targeting logic. Instead, define the deal unit: the smallest group of roles that consistently shows up in a real buying process.
For example, a security product might repeatedly involve:
● A technical evaluator who tests and compares
● A budget owner who needs risk justification
● An operator who has to live with deployment
● A procurement stakeholder who slows everything down
Notice what’s different here. You’re not writing a persona deck. You’re mapping how decisions actually get made. This matters because high performing ABM audiences don’t just “include decision makers.” They include the people who create momentum inside the account. Momentum is what turns awareness into meetings.
Step 2: Write audience rules like you’re designing a system
Most targeting setups sound like a vibe:
“Senior titles, relevant departments, intent topics.”
A high performing ABM audience sounds like a system:
● If an account is in ICP, include these buying roles.
● If an account is outside ICP but showing strong signals, include only evaluator roles.
● If an account is already in pipeline, include expansion roles and cross-functional stakeholders.
● If an account is in a regulated vertical, shift emphasis toward risk and compliance functions.
You’re creating logic that matches how you want to allocate attention. ABM is resource allocation dressed up as marketing. Treat it that way.
Step 3: Separate “reachable” from “important”
This is where most teams quietly compromise. There are always roles you want—CIO, CISO, VP-level stakeholders—that are harder to reach consistently. So the audience expands to whoever is easiest to identify at scale.
Suddenly your “ABM audience” is full of adjacent roles because they’re abundant in data.
That trade-off isn’t automatically wrong. But you need to acknowledge it. High performance comes from intentionally balancing importance and reachability. If you only chase importance, your audience becomes too small to learn from. If you only chase reach, you end up with engagement that doesn’t move deals. The solution is to design tiers.
Step 4: Build three tiers: Core, Influence, and Expansion
If you’re building from scratch, don’t try to cram everything into one audience.
Create tiers that reflect how deals happen. Core roles are the people closest to evaluation and ownership. If they don’t engage, nothing else matters. Influence roles shape internal consensus. They don’t always drive the process, but they can speed it up or shut it down.
Expansion roles matter when you’re already in conversation or when your product touches multiple functions.
This tiering does something critical: it makes performance interpretable. If Core is responding but Influence isn’t, your messaging is likely too product-centric. If Influence is responding but Core isn’t, you may have a relevance mismatch or the wrong trigger.
If Expansion is outperforming Core, you may be building awareness while missing the actual buying motion. Without tiers, everything blurs together and you get “mixed results” forever.
Step 5: Choose signals that reflect stage, not just interest
Teams love intent because it feels like certainty. But intent is noisy. It can mean curiosity, research, student projects, competitor monitoring, or genuine buying momentum.
The mistake is treating all signals the same way. Instead, tie signals to stage.
● Early stage signals should widen reach carefully, mostly to evaluators.
● Mid-stage signals should tighten toward roles that move internal decisions.
● Late-stage signals should reinforce with cross-functional confidence builders.
If your signals don’t change your audience logic, they’re not signals. They’re decoration.
Scenario 1: The “everything looks engaged” trap. A mid-market SaaS team builds an ABM audience using:
● ICP accounts
● Department targeting: “Operations” and “IT”
● Seniority filter: manager and above
The campaign performs well on clicks and site engagement. Sales reports that meetings aren’t improving. On inspection, the engaged users are mostly operations managers and analysts—people who care, but don’t own the problem the product solves.
The audience was reachable, not decisive.
They rebuild the audience into tiers and tighten Core to roles directly responsible for the workflow the product touches. They keep the reachable layer, but reclassify it as Influence.
Now engagement becomes a diagnostic, not a vanity metric. And meetings rise because the audience is finally aligned to the deal unit. The takeaway: engagement isn’t the goal. Correct engagement is.
Step 6: Make exclusions as deliberate as inclusions
Most ABM builds are inclusion-only. That’s how you end up targeting people who will never buy, even inside your best accounts. Exclusions feel uncomfortable because they shrink reach. But they increase signal quality.
Common examples:
● Excluding roles that consume content but don’t influence purchase
● Excluding subsidiaries or business units that don’t match your ideal deployment environment
● Excluding regions where you can’t realistically sell or support
● Excluding departments that overlap in name but not in function (a classic enterprise problem)
If you don’t define what “not relevant” looks like, your platforms will happily spend your budget finding it for you.
Step 7: Decide what “from scratch” means operationally
When teams say “build an ABM audience,” they often mean one of three things:
- Build a list of target accounts
- Build a set of ad platform audiences
- Build a measurement framework for ABM performance
High performance requires all three.
If you only do (1), you have a strategy without execution.
If you only do (2), you have execution without a strategy.
If you skip (3), you have activity without learning.
An ABM audience isn’t complete until you can answer:
● Who is included and why?
● How many people per account are reachable?
● Which roles are over- or under-represented?
● How does performance vary by tier?
● What happens to pipeline accounts versus net-new accounts?
If you can’t answer those, you can’t improve the audience. You can only replace it.
Scenario 2: When “accuracy” kills learning
An enterprise team builds an extremely strict ABM audience.
They only include VP+ titles in a narrow department set, inside a highly curated account list. It’s beautifully precise. Performance is unpredictable. Some accounts surge. Most show nothing.
The real issue is statistical and operational: the audience is too small to learn from, and the buying unit is broader than their filter admits. They loosen the Core tier slightly to include director-level evaluators and add an Influence tier that captures adjacent stakeholders who shape internal requirements.
Results stabilize. The team learns faster because they’re observing a real buying system, not a tiny sliver of it. The takeaway: precision is powerful, but too much precision can slow feedback loops.
Common misunderstandings
● “ABM audiences are just account lists.” Account lists are a starting point. Audiences are how you operationalize attention inside those accounts.
● “More intent data automatically improves targeting.” Intent helps when it changes decisions. If it doesn’t shift audience logic or messaging, it’s just noise.
● “If we include decision makers, we’re good.” Deals move through evaluators and blockers long before they reach a signature.
● “High match rates mean high quality.” Match rates can improve while relevance gets worse. The goal is outcomes, not coverage.
A practical playbook to build high performing ABM audiences
Start with your last 10 real opportunities. Not the personas you wish you had—actual deals that moved. Write down the roles that consistently appeared, including the “hidden” ones that slowed or accelerated progress. Group those roles into a deal unit and translate them into tiered audiences: Core, Influence, and Expansion. Build targeting rules that change based on account status: net-new, showing signals, in pipeline, or in active sales motion.
Define exclusions early. If a role never influences the deal, don’t pay to learn that lesson again.
Validate reachability per account. If you can only reach two people in an account, don’t pretend you’re running ABM there. Launch with measurement that separates tiers. Track not just engagement, but who engaged, in which accounts, and what happened next.
Then iterate on a cadence. ABM audiences aren’t built once—they’re tuned like an instrument, with every campaign teaching you what to tighten and what to widen.
Build the audience like it’s a product, and performance stops being a mystery.
