Decision 1: Is ABM the Right Motion at All?
Account-based marketing costs more per account than anything else in B2B. Research, custom content and paid air cover aimed at one company add up fast. So the first decision is arithmetic, not ambition.
Put it on one line: what one won account is worth in gross margin, against what it costs to run the programme across every account you target.
Some illustrative numbers. A first-year contract of $80,000 at a 40% margin is $32,000 of gross profit. If a tier-one account costs you $3,000 over two quarters in research, content and ads, one win pays for roughly ten accounts. Now try a $5,000 contract at the same margin. That’s $2,000 of profit, and you would need to win more accounts than you targeted just to break even.
The second test is whether you can name the companies. The line on our own account-based marketing page is blunt: if you cannot name 50 companies you want as customers, ABM does not fit. The same list rules it out when one person decides and buys the same week, or when there is no sales team to hand accounts to.
Fail either test and stop here. Demand generation is the better motion for you, and there is nothing wrong with that.
Decision 2: How Many Accounts Can You Actually Work?
Most target lists are sized by the market. They should be sized by capacity. The question is not how many companies could buy from you. It’s how many your team can research, write for and follow up on properly, every month.
Budget the hours first, then divide. An example:
| Tier | Hours per account, per month | Accounts | Hours |
|---|---|---|---|
| One-to-one | 6 | 10 | 60 |
| One-to-few | 1.5, worked in clusters | 40 | 60 |
| One-to-many | Programmatic | 300+ | 20 |
| Total | 350+ | 140 |
That’s 140 hours a month, close to one dedicated marketer, before a rep makes a single call. Swap in your own numbers. If the total is bigger than the people you have, the list is too long. The team is not too small.
There is a platform limit to plan around too. LinkedIn’s company list targeting needs at least 300 rows to upload, and the list has to match at least 300 member accounts before an ad set can run. A ten-account tier-one list cannot be its own uploaded audience. Build the paid layer around the bigger tiers, and treat tier one as mostly human work.
Decision 3: Which Accounts Get Which Tier
Revenue alone is a bad way to tier. The biggest company on your list is not tier one if nobody there will take a meeting.
Score each account from 1 to 3 on five things:
- Deal size at your pricing.
- Expansion headroom: other divisions, sites or product lines.
- Fit with your ideal customer profile.
- Existing relationship: a past customer, a warm contact, a shared partner.
- Reachability: can you name and contact the people who decide?
Here is how three fictional accounts come out:
| Account | Deal size | Headroom | Fit | Relationship | Reachable | Total | Tier |
|---|---|---|---|---|---|---|---|
| A | 3 | 3 | 3 | 2 | 3 | 14 | One-to-one |
| B | 3 | 2 | 2 | 1 | 1 | 9 | One-to-few |
| C | 1 | 1 | 3 | 1 | 2 | 8 | One-to-many |
Account B is the trap. It’s large, and on a revenue sort it looks like tier one, but nobody can get to it yet. Put it in the middle tier and let the programmatic layer tell you when it starts to warm up.
Decision 4: Who Inside Each Account You Need to Reach
An account doesn’t buy. A group of people inside it does. Gartner’s survey of 632 B2B buyers, published in May 2025, puts buying groups at five to 16 people across as many as four functions.
So for every tier-one account, write down the roles before you write a word of copy:
- Who signs it.
- Who wants it, and will argue for it internally.
- Who tests it.
- Who prices it.
- Who can stop it.
Then mark which of them you have actually reached. The gaps on that map are next month’s plan.
This is also where the strategy meets sales. If marketing reaches the finance lead and the rep never follows up, that coverage was wasted. Agree in writing who owns each role, and by when. A shared pipeline view helps here, and our post on building a sales pipeline covers the stage definitions both teams need to agree on.
Decision 5: One Case Per Account, Not One Per Person
The instinct in ABM is to personalise right down to the individual. The research says be careful.
The same Gartner survey found that content tailored to the buying group or the organisation lifted consensus by 20%. Content built for individual-level relevance had a 59% negative impact on consensus. Groups that did reach consensus were 2.5 times more likely to report a high-quality deal, and 74% of buyer teams showed unhealthy conflict during the decision.
Source: Gartner survey of 632 B2B buyers, August to September 2024, published May 2025.
What that means for your strategy: write one case per account, built on the problem the whole group shares. Then give each role the proof it needs for that same case.
- Finance gets the payback on it.
- IT gets the security answers for it.
- The champion gets something short enough to forward.
Five different arguments to five different people is how you hand a buying group something new to fight about.
Decision 6: What Counts as Working, and When
Choose your measures before launch, and put them in the order they can move:
- Coverage: how many target accounts show any engagement at all.
- Committee depth: how many roles you have reached per account.
- Pipeline created and moved across the named list.
- Win rate and deal size on target accounts, against your non-ABM baseline.
On a nine-month sales cycle, revenue cannot show up in month two. Coverage can. In our own programmes, coverage typically shifts within four to eight weeks, and revenue follows the client’s existing sales cycle. Report the early measures early, so the programme survives long enough for the late ones to arrive.
And start smaller than feels impressive. A pilot with one tier, one objective, a fixed window and success criteria agreed in advance tells you whether the list, the message or the follow-up was wrong. Four hundred accounts on day one fails everywhere at once and teaches you nothing.
