Account-based marketing has been "the next big thing" in B2B for a decade. Most programs still fail. The strategy is rarely the problem. Teams buy the software before building the system: they get the orchestration platform, skip the sales alignment, run some account-targeted ads, and call the experiment dead two quarters later.

This is the playbook we run instead. It works with or without dedicated ABM tooling, and it's organised the way the work actually happens: pick, learn, play, align, measure.

Step 1: Pick accounts like a portfolio manager

Your account list is 80% of your outcome. Build it from evidence, not enthusiasm:

  • Mine your closed-won data. What do your best 20 customers share across industry, size, tech stack, trigger events, and buyer titles? That's your ICP, written by reality.
  • Score every candidate account on fit (does it match the ICP?) and signal (is there any evidence of timing, like funding, hiring, tech changes, or content engagement?).
  • Tier ruthlessly: Tier 1 is the 10 to 25 accounts you'd reorganise your quarter to win. Tier 2 is the next 100 to 300 with strong fit. Tier 3 is the broader ICP, handled programmatically.
The most common ABM mistake is a Tier 1 list with 200 accounts. If sales can't tell you who owns the pursuit of each named account, it's not Tier 1. It's a mailing list.

Step 2: Research until the message writes itself

For every Tier 1 account, build an intelligence brief before any outreach: company priorities (from earnings calls, job posts, leadership interviews), the likely buying committee, current vendors, and the trigger that makes now the right time. AI agents inside n8n workflows can compile 80% of this automatically; a human adds the judgement layer.

The test for "enough research": could you write a first line that the prospect would believe took effort? If not, keep digging. Personalisation that isn't specific is just mail merge with extra steps.

Step 3: Run plays, not campaigns

A campaign broadcasts; a play converges. Each play is a coordinated sequence across channels with a specific goal for a specific set of people:

The surround play (Tier 1 openers)

Weeks 1 and 2: LinkedIn ads to the buying committee establish familiarity. Week 2: personalised cold emails to 2 or 3 committee members, each angle matched to their role. Week 3: the AE follows up with a specific point of view on the account's situation. Air cover, ground game, close.

The trigger play (any tier)

An account raises funding, hires a new VP, or visits your pricing page → an automated workflow promotes the account, alerts the owner, and launches the matching sequence within 24 hours. Timing does the personalising for you.

The revival play (closed-lost)

Six months after a lost deal, check what changed. Champion moved? Competitor stumbled? New funding? Then re-open with "since we last spoke" relevance. Closed-lost accounts already know you; they're the warmest cold list you own.

Step 4: Make sales co-owners, not recipients

ABM dies in the handoff. Prevent that structurally:

  • Sales co-selects the account list. If a rep didn't vote for an account, they won't chase it.
  • Engagement alerts go to the account owner in Slack or the CRM ("two people from Northwind read the pricing page today") with a suggested next move attached.
  • Every Tier 1 account has a pursuit plan (who we're targeting, the angle, the next three touches) visible to both teams in the CRM.
  • One scoreboard, reviewed together, every two weeks. Marketing and sales looking at different dashboards is how "whose lead was that" arguments are born.

Step 5: Measure accounts, not clicks

ABM metrics that matter, in funnel order:

  1. Coverage: do we have contacts and a play running for every target account?
  2. Engagement: are the right people at target accounts interacting (visits, replies, ad engagement, event attendance)?
  3. Meetings: first meetings with named accounts, the cleanest mid-funnel signal.
  4. Opportunities & pipeline: open pipeline at target accounts versus a holdout or historical baseline.
  5. Win rate & ACV: the endgame. Targeted accounts should close more often and bigger than untargeted ones.

Report quarterly against the baseline, and resist the urge to claim every dollar. Credibility with finance is itself an ABM asset.

What's different in 2026

  • AI research at scale makes Tier 2 personalisation feel like Tier 1. Account briefs that took an analyst a day now take a workflow ten minutes.
  • Your buyers ask AI about you. When a target account researches the category in ChatGPT or Perplexity, GEO decides whether you appear. ABM and AI visibility are now the same fight on two fronts.
  • Signal data got cheap. Website de-anonymisation, job-change tracking, and intent feeds that cost six figures in 2020 are now mid-market line items. The advantage has shifted from having signals to acting on them fast.
Start-this-week version: pull your last 20 closed-won deals, write the ICP they prove, pick 15 Tier 1 accounts with sales in the room, and run the surround play on five of them. That's a real ABM pilot, no platform purchase required.
OM
The Omnitics Team We design and run ABM programs end to end. Bring your dream-account list and we'll sketch the plays free.