ABM playbook for one-to-few programs: The account list, the plays, the metrics and a 90-day plan

An account-based marketing (ABM) playbook for one-to-few programs is the operating plan for running ABM against clusters of 5 to 15 accounts that share an industry or business issue. It sets the account list, the plays by stage, the split of work with sales, the metrics against a matched control group and a 90-day plan. A RevvGrowth benchmark guide found that 64% of business-to-business (B2B) marketing teams run an account-based strategy.

What a one-to-few ABM playbook is and the figures behind it

A one-to-few ABM playbook is the operating plan for running account-based marketing (ABM) against small clusters of accounts instead of single accounts or hundreds of them. Demandbase defines one-to-few as the research-based principles of one-to-one ABM applied to clusters of 5 to 15 accounts that share similar business issues or belong to the same industry. As an example, not a benchmark: a 30-account list split at 15 accounts per cluster gives 2 clusters, and the first plan runs 90 days.

• 64% of B2B marketing teams now run an account-based or target-account strategy, according to a RevvGrowth benchmark guide. The denominator is marketing teams, not companies or budget.

A program built around one contact per account misses most of the people who decide.

The six steps, in order

Demandbase sets out six steps. The right-hand column is how we apply each one to a cluster.

For the Measure step on paid media, Demandbase suggests measuring "cost per engaged account". The formula is plain: ad spend on a cluster divided by the number of accounts in that cluster that met your engagement threshold. Run it per cluster, not across the whole list, so a strong cluster does not hide a weak one. For the velocity side of the lagging indicators, see pipeline velocity.

What are the three tiers of ABM

In 2003, ITSMA formalized and promoted ABM, and Momentum ITSMA still tracks the discipline. Its 7th annual benchmark report was based on an international survey of more than 200 ABM leaders and practitioners, published with the ABM Leadership Alliance.

The three tiers differ in how many accounts get the same program and how much of that program is built by hand. In one-to-one, a dedicated, senior‑level marketer partners with the account team to build programs for "markets of one." One-to-few applies the same research to a small cluster, an approach that allows personalisation while achieving some scale. One-to-many is personalization at scale for hundreds of named accounts.

How teams blend the tiers

The blend is set by cost. Few, if any, companies can afford to do true One-to-One ABM to even a fraction of their individual accounts. The same source warns that skipping one-to-one for the handful of key accounts could mean missing a great deal of opportunity.

A working sequence we use to place accounts:

1. Name the few accounts where losing the relationship would change the revenue plan. They go to one-to-one, whatever the count.
2. Group the next layer by shared business issue, not by size. Each group becomes a one-to-few cluster with its own play.
3. Put the rest of the target list into one-to-many segments run through your advertising and automation stack.
4. Review placement each quarter, moving accounts up a tier when intent data or sales activity rises.

As an example, not a benchmark: a team with 30 accounts in one-to-few clusters and 90 in one-to-many manages 120 named accounts below the one-to-one tier. In our reading, the 30 deserve most of the custom content budget per account, while the 90 share segment assets. The demand gen budget allocator helps test that split before you commit spend.

How many accounts belong in a one-to-few program

Cluster size is set by the definition covered earlier. Total program size is a capacity question. We size a one-to-few program from three constraints and take the lowest answer: marketer hours, sales owners and pipeline coverage.

The three constraints

1. Marketer hours per cluster. Each cluster needs shared research, a tailored message, a content set and a run of coordinated plays. Estimate the monthly hours one cluster takes, then divide the marketer's available program hours by that figure. The result is the number of clusters the marketer can run.
2. Sales owners. Every account needs a named account executive who will act on engagement. Count the accounts each executive can work alongside their existing book, then multiply by the number of executives.
3. Pipeline coverage. Work backward from the pipeline the program owns. Divide the target by the average opportunity value and by the share of accounts you expect to open an opportunity. That tells you the minimum number of accounts the program needs. The pipeline coverage calculator and the reverse funnel calculator run this step.

If coverage asks for more accounts than hours or sales owners allow, the target needs to change, or the extra accounts move to one-to-many. Adding accounts to the clusters is the wrong fix.

Worked example

This is an example, not a benchmark. One ABM marketer runs 3 clusters of 12 accounts, supported by four account executives.

• 3 clusters of 12 accounts gives 36 accounts in the program.
• If each cluster takes 30 marketer hours a month, 3 clusters need 90 hours a month.
• 36 accounts across 4 account executives is 9 accounts per executive.
• 90 marketer hours spread over 36 accounts is 2.5 hours per account per month.

In our reading, 2.5 hours per account only works because research and content are built once for each cluster and then reused. If one account starts to need its own assets, it has outgrown one-to-few. Move it up a tier, or cut a cluster to free the hours.

Recheck after the first quarter

Replace the estimated hours with the hours you actually logged. Ask each account executive how many accounts they really worked. For paid media, Demandbase suggests measuring cost per engaged account to judge the ABM ad budget. That figure shows whether spend per cluster holds up as the program grows. Then resize to whichever constraint binds first.

How to select target accounts and group them into clusters

Each step has one owner and one written output, so the next step starts from a file rather than from memory.

1. Find your best customers. Owner: marketing operations, pulling from the CRM. Demandbase's account identification guide points to "Closed-won deals with the highest lifetime value (LTV)" and "Customers with fastest deal cycles and lowest churn." Output: a sheet of best customers with lifetime value, cycle length and churn status on every row.
2. Define firmographics and technographics. Owner: the demand generation lead with product marketing. Read the shared traits off the best-customer sheet: industry, size, revenue range, tech stack. Output: an ideal customer profile (ICP) and a non-ICP list. The same guide lists red flags as "incompatible tech stacks, poor revenue potential, regulatory barriers, or industries with low adoption rates."
3. Score fit and intent. Owner: marketing operations. Operatix advises that "Scoring should consider key factors like fit, intent, engagement, and potential revenue." For engagement, Demandbase suggests a 90-day lookback for accounts that engaged but did not convert. Output: a ranked list. See intent data and lead scoring for the mechanics.
4. Have sales validate the list. Owner: sales leadership and the account executives. Output: a signed-off list with a reason code for every account sales removes.

Scoring template

The stakeholder row follows Demandbase's note on "Multiple stakeholders from the same company researching" as a signal.

Worked example

These are example inputs, not a benchmark. If 90 accounts clear scoring and sales keeps 30, sales has removed 60. Those reason codes go back into the non-ICP list for the next cycle. The 30 remaining accounts are then split by the need they share, within the cluster size set earlier, and each cluster gets one play.

Which plays to run for each cluster by stage

Each cluster from the previous section shares one business issue. The plays below keep that issue constant and change the channel, the content and the trigger as an account moves from unaware to customer. Intent signals decide when an account changes stage, so define both kinds before you build the table.

First-party intent data comes from your own touchpoints: "visits to your pricing page, resource downloads, or webinar registrations." Third-party intent data "captures activity across external platforms and websites through data providers." See our intent data entry for how teams score each kind.

The stage table

Rules for triggers

Treat a single spike as a reason to look, not a reason to call. Some intent spikes reflect early-stage interest, job hunting, or competitive analysis rather than buying intent. We require a third-party signal to be confirmed by a first-party one before sales development outreach starts.

For customers, the same data runs in both directions. Teams use intent signals to uncover cross-sell and upsell opportunities, and they watch for the point when customers start researching competitors so the account team can step in early.

Buying-group coverage

A stage change should depend on breadth, not on one enthusiastic contact. Map the roles you expect in each buying group, then track how many you have reached.

As an example, not a benchmark: a cluster with 90 mapped stakeholders and 30 of them engaged has coverage of 30 / 90, or 33%.

To check whether the triggers work, compare whether accounts with intent signals convert from marketing-qualified leads to sales-qualified leads at a higher rate than accounts without them. The definitions are in our marketing qualified lead and sales qualified lead entries.

How sales and marketing split the work in one-to-few ABM

In our reading, a one-to-few program fails at the handoffs before it fails at the plays. ZoomInfo's ABM playbook puts it plainly: most breakdowns happen when outreach and follow-up aren't connected, and marketing often runs campaigns sales reps don't see. The fix is a written service level agreement that names an owner for every recurring task.

The same playbook describes the one-to-few tier as coordinated by small marketing, sales pods. Each pod covers one cluster. Its first job is to decide who runs point for each account so ownership is clear.

The responsibility template

Copy this into your planning doc and fill in the bracketed terms with sales leadership before launch.

Outreach timing sits with one owner because accounts in a cluster get touched by several channels at once. ZoomInfo describes orchestration tools that manage timing across outbound, ads, and nurtures so accounts don't receive conflicting or mistimed messaging. A shared calendar does the same job at small scale.

The weekly review

ZoomInfo calls the weekly ABM standup the reset point. Keep it to one agenda: which accounts moved, which stalled, which signal went unanswered, and which play changes this week. As an example, not a benchmark: with 30 accounts in a 90-day plan, the pod reviews the whole list every week and records one decision per account, so nothing waits a month for attention.

Why the metrics have to be shared

The IAB defines ABM as a strategic approach where sales and marketing are aligned and focus on agreed accounts.

In our reading, alignment holds only when both teams report the same numbers. For one-to-few, ZoomInfo lists the tier as measured by pipeline creation, velocity, and win rate improvements. Put those three on one dashboard that both teams own. Marketing should not report engaged accounts while sales reports bookings. Track pipeline velocity per cluster, and let a missed service level show up next to the account it cost.

Which metrics measure ABM, with formulas and a worked example

The table below gives one formula per metric. Each one is calculated twice, once for the cluster and once for a matched control group, because a number with nothing to compare against cannot answer the question leadership asks.

For the cycle-length input, Vector suggests that teams compare the days to close for the targeted accounts against the others. Revvgrowth reports that ABM accounts often generate deal sizes between 11% and 50% larger than non-ABM accounts, depending on region and segment. Treat that as a range to test against your own data. It is not a target.

Worked example: 12 accounts against a matched control

These inputs are an example, not a benchmark. Pick 12 accounts for the cluster. Then pick 12 more that match them on segment, size and stage, and leave those 12 out of the plays.

Run the plays for 90 days. Suppose the cluster creates $90,000 in qualified pipeline, which is $7,500 per account across its 12 accounts. Suppose the control creates $30,000, which is $2,500 per account across its 12 accounts.

The gap is $7,500 minus $2,500, or $5,000 per account. That difference is the figure to report. Repeat the same comparison for engagement rate, win rate, deal size and velocity once deals in both groups close.

Why the holdout handles attribution

Vector notes that the complexity of attribution makes ROI difficult to measure for account programs. Revvgrowth says benchmarks should help a team tell whether ABM is driving higher-value deals or simply redistributing revenue that would have closed anyway.

A holdout answers that second question directly. The control accounts get the same sales coverage and the same baseline marketing, so whatever they produce is what would have happened without the plays. In our reading, the difference between the two groups is a cleaner estimate of incrementality than any touch-based model, because it does not depend on deciding which email or ad earned the credit.

Two rules keep the comparison honest. Lock both lists before the program starts. Measure both groups over the same window, using the same definitions of pipeline stage.

The 90-day plan and how long ABM takes to show results

A one-to-few program needs a fixed calendar before it needs more plays. We split the first quarter into three phases of about a month each. Every phase has one owner, one output and one pass test. If the test fails, the phase gets extended rather than the next one starting late and thin.

Week by week

• Days 1 to 30. In the first two weeks, lock the list and clusters. In the second two, map buying groups and record a baseline engagement snapshot, so the readout has something to compare against.
• Days 31 to 60. Launch one play per cluster, not all of them at once. Run the weekly standup from the first week of launch. By the end of the phase, every engaged account should have a logged sales touch.
• Days 61 to 90. Pull engagement against the baseline, review open opportunities, and make the keep, fix or cut call for each play.

What to expect by day 90

In our reading, a 90-day readout should report engagement movement and early opportunity creation, not a final win-rate verdict. Engagement moves within weeks because it only needs a buyer to respond. In our reading, pipeline and win rate move on the sales cycle, which in most one-to-few programs runs longer than a quarter. As an example, not a benchmark: with 30 accounts, the day-90 readout should say how many of the 30 moved from cold to engaged. It should not say how many closed. Set that expectation with sales and finance on day one, and link it to your pipeline velocity view so later quarters pick up the revenue measures.

Use the readout to check capabilities as well as results. The benchmark report from Momentum ITSMA and the ABM Leadership Alliance drew on a survey of more than 200 ABM leaders and practitioners. It highlights the factors that separate high-performing programs from the rest, and it lists the capabilities you should build and develop for future success. We treat each failed pass test as a capability gap to fix in the second quarter. More budget for the same plays does not close that gap.

What this page does not cover

This page covers one-to-few programs and does not set out full designs for one-to-one or one-to-many programs, specific budget levels or the selection of advertising, intent data or orchestration tools. The example figures illustrate the formulas and are not benchmarks. Nothing here is financial or legal advice.

Related on Take the Market: Self-reported attribution: how to design the field, code the answers and reconcile them with software attribution; Pipeline coverage calculator: the ratio you need from your own win rate and slippage; Demand generation calculator: reverse funnel math from a revenue target to visitors, leads, MQLs, SQLs and opportuniti.