Build an Account-Based Marketing Program

Most account-based marketing playbooks open the same way: build a data warehouse, model your accounts in dbt, hire a reverse-ETL vendor, then start running plays. Mid-market RevOps teams read the playbook, look at their ops budget, and quietly close the tab.

Stage-based account-based marketing does not actually require a warehouse. It requires a single agreed-upon account stage, a few signals routed into the CRM, and a couple of plays tied to stage changes. The warehouse layer in the famous playbooks is solving for analytical reporting, not for the ABM motion itself.

This guide is for marketing-ops and RevOps teams running ABM at a B2B SaaS company without a data engineering team. We will cover what account-based marketing data looks like, how to design an ABM stage model from existing tools, the setup, and the edge cases that trip people up.

Why account-based marketing breaks without unified account data

The reason most ABM programs feel chaotic is not strategy. It is data routing. Sales sees one set of signals in the CRM. Marketing sees a different set in the email platform. Ads run off a list someone uploaded last quarter. Each tool has a partial view, and no place holds the account's current state authoritatively.

If sales believes an account is in active eval and marketing is still showing them top-of-funnel ads, the program is misaligned no matter how good the playbook is. A useful definition of account-based marketing is just this: marketing and sales agree on which accounts matter and what stage each one is in, and act accordingly.

The warehouse-first approach makes the warehouse the source of truth and syncs stage downstream. That works well for teams with warehouse infrastructure, but most teams can achieve the same outcome more directly. The same outcome can come from picking the CRM as the agreed source of truth and syncing stage signals into it from every other tool.

What account-based marketing data your team needs (and where it already lives)

ABM data is not exotic. It is the same engagement and firmographic data your tools already track, framed at the account level instead of the contact level.

You do not need every row on day one. Pick three or four sources, wire them into the CRM, and validate the stage rules before adding more. None of these sources require a warehouse, though Oneprofile also works with one if you have it. They each expose an API. The job is moving the relevant fields into one account record in the CRM and keeping that record fresh.

How to design an ABM stage model from existing tools

Most published ABM playbooks converge on the same five awareness stages: Unaware, Aware, Interest, Desire, and Action. Borrow that directly. The stages themselves are not the hard part. The hard part is defining which signals advance an account from one stage to the next, and writing those rules somewhere auditable.

A reasonable starting model:

• Unaware. No engagement signals from any source in the last 90 days.

• Aware. First meaningful signal: ad click, content download, email click, website session, or any inbound contact response.

• Interest. Sustained engagement: multiple sessions, repeat email engagement, a content download from the same domain, or a free-trial signup.

• Desire. High-intent behavior: pricing-page visits, multiple stakeholders engaged, demo request, or evaluating against listed competitors.

• Action. A clear conversion: opportunity created, contract sent, or paid plan activated.

Two design rules make this easier to maintain:

1. The stage field lives on the Account (or Company) object in the CRM. Every play reads from this one field. If three different tools each compute their own stage, you have rebuilt the original problem.

2. Stage rules are written as plain English in a shared doc, not just as click-trail workflows in a CRM UI. Workflows drift. A doc you review every quarter does not.

For deeper context on ABM as a discipline, the CDP Institute publishes activation research worth skimming.

Step-by-step: set up account-based marketing sync directly

This is the bare-minimum setup. A day or two of work, no engineering tickets.

1. Pick your matching key. Most teams match on company domain at the account level and email at the contact level. Verify your CRM, billing tool, and product database all populate one of these reliably. If any source is missing it on more than 10% of records, fix that first.

2. Add a stage field to the CRM. Create a single picklist field called account_stage on the Account or Company object, with the five stage values. This is the only field downstream tools will read.

3. Wire signals into the CRM. For each source, set up a sync that writes specific fields onto the matching CRM account.

• Billing: subscription status, plan, MRR, last payment date.

• Product database: last login, weekly active users, feature adoption flags.

• Support tool: open ticket count, last CSAT score.

• Email tool: last engagement date, total opens in last 30 days.

• Website analytics: last pricing-page visit, repeat-session count.

Map each one to a dedicated CRM property. Oneprofile handles this layer: authenticate each source, pick the records, map fields to CRM properties, set a schedule. New CRM properties get created automatically. Bidirectional sync means anything pushed to the CRM can flow back out to ads and email later.

4. Write the stage computation. Most CRMs can do this with native workflow rules. The logic is "if any condition for stage X is true and the current stage is below X, set account_stage = X." If your CRM workflow engine is too limited, a small scheduled script reading the same properties works fine.

5. Tie one play to each stage. Resist building twenty plays before validating any of them.

• Aware → trigger an email nurture sequence aligned to the entry signal.

• Interest → assign an SDR research task; add to retargeting audience.

• Desire → notify the AE; switch ads from awareness to product creative.

• Action → suppress from marketing audiences; route to sales-led motion.

6. Sync stage back to ads and email. Once the CRM holds the authoritative stage, push it to ad platforms as audience membership and to your email tool as a list or property. This is the same orchestration enterprise ABM platforms charge six figures for; it is also a basic property sync.

7. Add a suppression list. Closed-won customers, disqualified accounts, employees, and known competitors should be excluded from awareness ads. Treat suppression as a real audience, not an afterthought.

Edge cases when running account-based marketing across CRM, billing, and product tools

A few patterns reliably break ABM programs assembled this way.

Multiple contacts, one account. A buyer journey touches three to five people at the same company. Engagement scoring at the contact level over-counts. Roll signals up to the account before computing stage.

Free-tier users at named accounts. A self-serve signup from the IT manager at a target account is a strong Interest signal even if the email looks personal. Match on email domain when possible.

Currency and unit mismatches. Stripe stores amounts in cents; HubSpot expects dollars. Apply a transformation during field mapping so MRR shows up correctly.

Stage regression. Should an account move backwards from Interest to Aware if it goes quiet? Most teams say no during the active sales cycle and yes after the deal closes lost. Pick a rule and write it down.

Stage thrash. If your rules are too sensitive, accounts oscillate every sync run. Add a minimum dwell time (24 hours, say) before an account becomes eligible to advance.

The first sync is loud. Backfilling thousands of accounts at once produces a wave of CRM notifications and possibly a flurry of automated SDR tasks. Run the initial sync in a quiet window and mute downstream automations for the first few hours.

What changes after you turn this on

The biggest shift is conversational, not technical. Sales and marketing stop arguing about which accounts are real. The stage field becomes a shared vocabulary. Reps stop wasting outreach on cold accounts; marketing stops running ads at accounts that closed last quarter.

You will not get every signal right on the first try. Stage rules need a few quarters of tuning. That tuning is fast when the data is in one place and the rules are written down.

This is also the part of ABM that does not need a six-figure platform. Intent vendors, ad orchestration, and dynamic personalization are real, worth buying when the basic motion is working. They are worthless if the underlying account stage is wrong. Get the stage right first; everything else is a layer on top.