ABM programs fail for a lot of reasons. Bad ICP targeting. Misaligned sales and marketing. Generic content. But the failure mode that marketing ops people know intimately is the data plumbing problem: the CRM is not set up to think at the account level, engagement data is siloed in the MAP, intent signals are living in a platform nobody checks, and the sales team is working from a spreadsheet.
Good ABM infrastructure does not require a $200,000 tech stack or a team of three developers. It requires the right data model, the right automations, and the right integrations, built in the right order. This guide is a blueprint for marketing ops teams building ABM infrastructure from scratch or retrofitting an existing martech stack to support an account-based motion.
Think of ABM infrastructure as four layers, each building on the one below.
Layer 1: Data Foundation. Account records, contact records, firmographic enrichment, and the data model that connects them. This layer is the ground floor. Nothing else works without it.
Layer 2: Signal Collection. The mechanisms for capturing and storing behavioral signals (website activity, email engagement, ad clicks, product usage) and intent signals (third-party research behavior) at the account level.
Layer 3: Scoring and Routing. The logic that converts raw signals into prioritized scores, assigns accounts to tiers, and routes high-priority accounts to the right actions and owners.
Layer 4: Execution and Personalization. The outreach sequences, advertising campaigns, website personalization, and content delivery that are triggered by the scoring and routing logic.
Most teams try to build Layer 4 before they have Layer 1 properly configured. That is why so many ABM programs collapse into chaos after the first 90 days.
The CRM account object needs to carry a specific set of fields to support ABM operations. Default CRM configurations are built for contact-centric selling. ABM requires an account-centric view.
Required account fields for ABM:
Required fields to carry on the contact record:
Getting these fields built and populated before you launch any ABM campaigns saves enormous pain later. You cannot report on account-level engagement or run tier-based automations if the fields do not exist.
Raw account records from CRM imports or web form fills often contain minimal data: a company name and a domain. ABM scoring requires much more.
Enrichment sources and what they provide:
A firmographic enrichment platform appends employee count, revenue range, industry classification, headquarters location, company description, and sometimes tech stack information. The major options each have different coverage and pricing. Most marketing ops teams run scheduled enrichment jobs that update account records weekly for Tier 1 and Tier 2 accounts.
Technographic enrichment adds value for companies where tech stack is a predictor of fit. If your product integrates with or competes with specific platforms, knowing what a prospect uses is actionable. Technographic data providers scan public signals (job postings, API documentation, public install data) to infer technology usage at the company level.
Growth signal enrichment captures hiring trends, funding announcements, and news events that indicate account momentum. This is most useful for startups and growth-stage companies where headcount trajectory is a stronger ICP signal than absolute company size.
Build enrichment into your CRM workflow as an automated step rather than a one-time manual import. Accounts should be re-enriched when they are created, when key fields change, and on a quarterly refresh schedule.
The first behavioral signal layer is identifying which companies are visiting your website, even without form fills. Most B2B website traffic is anonymous: the visitor did not fill out a form and you have no contact record for them.
Reverse-IP resolution matches the visitor’s IP address to a company. The result is an account-level record: “[Company X] had three sessions this week, including two visits to the pricing page.” You do not know the specific individual, but you know the account is interested.
Configure reverse-IP data to write back to your CRM account records automatically. For each visit, capture: the company name, visit date, pages viewed, and session count. Store this as activity records on the account, not just in the reverse-IP tool.
Your marketing automation platform (HubSpot, Marketo, Pardot) collects individual-level behavioral data: email opens, clicks, form fills, page views for known contacts. For ABM, you need to roll these up to the account level.
Build account-level engagement score rollup automation. When any contact at an account takes an action, the account’s behavioral score updates. Define the scoring rules:
Apply score decay: subtract 2 points per day from the behavioral score if there is no new activity. This prevents accounts with old engagement from permanently occupying the top of the prioritization list.
Most marketing automation platforms support account-level scoring natively now. If yours does not, build the rollup in your CRM using a workflow that fires whenever a contact engagement activity is logged.
Third-party intent data requires an integration between the intent platform and your CRM. The configuration varies by platform, but the principle is consistent: intent signals for accounts in your CRM should write back to the account record in near-real time.
What to write back to the account record:
Build a CRM workflow that fires when a new intent signal arrives on an account: update the Intent Signal Strength field, update the Last Intent Signal Date, and create a task for the account owner if the signal strength is High. For medium signals, add the account to a daily digest report rather than creating an individual task.
With Layer 1 and Layer 2 in place, you have the data needed to build composite account scoring.
The composite score calculation:
Composite Score = (ICP Fit Score × Fit Weight) + (Behavioral Engagement Score × Behavioral Weight) + (Intent Signal Score × Intent Weight)
A reasonable starting weight: ICP Fit 40%, Behavioral Engagement 35%, Intent Signal 25%.
Build this as a calculated field in your CRM. The field updates whenever any of the three component scores change. This means the composite score is always current, not a weekly batch process.
Tier assignment automation. Build a CRM workflow that fires when the composite score changes and updates the ABM Tier field based on the score thresholds you defined. Include a condition that prevents tier downgrades from triggering during a period when there is an active opportunity (you do not want an account to be removed from Tier 1 just because their engagement score dropped temporarily while the deal is in negotiation).
Routing automations are the mechanism that connects data signals to human action.
High-priority alert automation: Fires when an account’s composite score crosses the Tier 1 threshold or when a High-intensity intent signal arrives for an account in any tier. Action: creates a task for the account owner with the specific signal details, sends a Slack notification to the SDR or AE, and adds the account to the high-priority queue report.
Tier promotion automation: Fires when an account moves from Tier 3 to Tier 2, or from Tier 2 to Tier 1. Action: updates the ABM Tier field, updates the Tier Assignment Date, creates a task for the account owner to review the account and take action, and enrolls the account in the appropriate ABM campaign track for the new tier.
Suppression automation: Fires when an account is enrolled in an active ABM sequence. Action: adds the account to the suppression list for generic demand gen campaigns. This prevents an ABM Tier 1 account from receiving a generic nurture email the same week their SDR is running a personalized sequence.
Stale account automation: Fires when an account’s Tier Assignment Date is more than 90 days old and the composite score has not changed. Action: flags the account for human review. Accounts that have been in a tier for 90 days without significant score movement may need to be re-evaluated or moved to a lower priority.
Every marketing touchpoint must be tagged consistently so that attribution and reporting work correctly. Build a tagging architecture before you launch any campaigns.
Required tags on every campaign or sequence:
Build a campaign tagging template and require it for every new campaign record created in the CRM. This is an ops discipline problem as much as a technical one: the tool infrastructure can enforce required fields, but you need buy-in from the campaign managers that they will fill them in.
For a multi-tier ABM motion, you need a program architecture in your MAP that reflects the tier structure.
Program structure:
Build the suppression infrastructure first. The biggest operational failure mode in mixed ABM and demand gen environments is active ABM accounts receiving generic demand gen communications. Suppress by ABM Tier field: any account with Tier = 1 or Tier = 2 is excluded from Tier 3 programs automatically.
If you are building from scratch and need to know what to prioritize, here is the minimum viable stack:
CRM (HubSpot or Salesforce): The system of record. Account objects must be configured with the fields described in Layer 1 before anything else.
Marketing automation (HubSpot, Marketo, or Pardot): Signal collection, lead nurture, email sequences, and program management.
Reverse-IP / website visitor identification: The cheapest and most immediate source of intent signal. Most ABM platforms include this.
LinkedIn Matched Audiences: Account-level advertising targeting using CRM account lists. This is the advertising layer for Tier 2 accounts.
Sales engagement platform (Outreach, Salesloft, or Apollo): Automated email sequences with personalization tokens, reply detection, and activity logging back to CRM.
Enrichment (Clearbit, ZoomInfo, or equivalent): Firmographic and technographic data appended to account records.
Total tooling cost for this stack varies significantly by vendor choice and account volume. The decisions that matter more than the specific tools are: which system is the source of truth for account data, how data flows between the systems, and what happens in each system when a key event occurs (new intent signal, score change, tier promotion).
ABM infrastructure requires ongoing maintenance. The most common failure mode after a successful launch is that the infrastructure degrades over time because nobody owns the maintenance cadence.
Weekly (ops owner): - Check for data sync failures between systems - Review suppression list accuracy (are Tier 1 and Tier 2 accounts appearing in Tier 3 programs?) - Check alert automation fire rate (if zero alerts fired this week, the automation may be broken)
Monthly (ops owner + marketing lead): - Review composite score distribution: are the right accounts in the right tiers? - Check enrichment coverage: what percentage of active accounts have complete firmographic data? - Review tagging compliance: what percentage of new campaigns were tagged correctly? - Update suppression lists if segment or tier criteria changed
Quarterly (ops owner + sales and marketing leadership): - Validate scoring weights against recent closed-won/closed-lost data - Update ICP segment definitions if targeting has shifted - Audit tool integrations: are all data flows still functioning as expected? - Evaluate whether any tools in the stack have become redundant or whether gaps have emerged