Most B2B account based marketing programmes fail in the same boring way. Marketing builds a list of 200 logos, buys an intent data subscription, runs display ads at those companies for two quarters, and then cannot explain to the CFO why pipeline looks the same as it did before. The tooling worked. The targeting worked. Nobody in sales changed what they did on Monday morning.
That is the thing worth understanding before you spend anything. Account-based marketing is not a channel you switch on. It is an agreement between two teams about which companies matter and who does what when one of them raises a hand. The software is the easy part.
This guide covers how to build that agreement, how to size the list, what to actually make, and how to tell within a quarter whether it is working.
What account-based marketing changes
Traditional demand generation treats the lead as the unit of work. You publish, people convert, the forms fill up, and sales works whatever arrives in descending order of score. It is efficient when your deal sizes are modest and the buying decision sits with one person.
It breaks down at enterprise deal sizes for a structural reason: the person who fills in your form is almost never the person who signs. A typical six-figure software purchase now involves somewhere between six and ten people, and a good number of them will never touch your website before the contract is drafted. Scoring individual leads in that environment produces a queue sorted by who is most curious, not by which company is most likely to buy.
ABM inverts the unit of work. The account is what you target, measure and report on. A trial signup from an unknown Gmail address at a target account is a signal about the account, not a lead to be nurtured in isolation. Three people from the same company reading your pricing page in a week is a much stronger signal than one person downloading four ebooks.
The practical consequence is that your reporting has to change before your tactics do. If your dashboard still counts MQLs, an ABM programme will look like a failure for its entire life, because it deliberately produces fewer of them.
Deciding whether it is worth it
ABM is expensive per account. It only pays when the accounts are worth the expense, and there is a straightforward test.
Take your average contract value, multiply by your realistic win rate against a named target account, and compare the result to what it costs you to run a personalised campaign against that account for two quarters. If your ACV is $80,000, you close one in six named targets, and a tier-one play costs you $2,500 per account in content, media and sales time, the arithmetic is comfortable. If your ACV is $9,000, it is not, and you should be improving your inbound motion instead.
Get honest numbers before you decide. Our customer acquisition cost calculator will give you a defensible blended figure to work from, and the LTV calculator tells you what the ceiling on that spend should be. If those two numbers are currently guesses, start with our guide to getting customer acquisition cost right — an ABM programme built on a wrong CAC will produce confident, wrong decisions for a year.
The second test is sales capacity. ABM generates a small number of high-value conversations that require a senior rep to handle well. If your team is measured on activity volume and has 90 accounts each, the personalised outreach the programme depends on will not happen, and you will have paid for ads nobody followed up.
Building the target list
This is where most of the value is created or destroyed, and it takes longer than people expect. Budget two to three weeks of real analysis, not an afternoon with a filter in your CRM.
Start with your own closed-won data. Pull every deal you have won in the last two years and look for what they share beyond the obvious firmographics. Industry and headcount are the lazy answer. The useful patterns are usually operational: they had just hired a specific role, they were running a particular competing product, they had opened a second location, they had a compliance deadline. Those are the triggers you can actually detect.
Then look at closed-lost, especially the ones that went nowhere. Accounts that took four meetings and then vanished tell you more about bad fit than accounts that said no immediately. If a segment consistently reaches proposal and stalls, it belongs on an exclusion list, not a target list.
Size the list to your capacity, not your ambition. The common error is a 500-account list that gets treated as one undifferentiated audience, which is just advertising with extra steps. A reasonable starting shape for a team with four or five reps:
- Tier one — 20 to 40 accounts. Genuinely bespoke. Named research, custom content, executive involvement, individual plans.
- Tier two — 100 to 200 accounts. Clustered by industry or use case. Content is personalised to the cluster, not the company.
- Tier three — 500 to 2,000 accounts. Programmatic. Same content as your inbound programme, targeted at a defined list.
Each tier is a different economic model, and mixing them up is the most common reason budgets vanish without result. Tier three is cheap per account and should mostly be paid media and email. Tier one should feel closer to a pursuit than a campaign.
What you actually build
The word "personalisation" causes a lot of waste here. Swapping a company logo onto a landing page is not personalisation; it is a mail merge, and buyers have been ignoring it for a decade.
What earns a reply is specificity about the buyer's situation. A few things that consistently work:
A point of view on their problem, in their language. Not a case study from their industry — an argument about the specific operational issue you believe they have, with the reasoning shown. This is where practitioner input matters more than copywriting.
Comparison and migration content. If a target account is running a competitor's product, the highest-value asset you can produce is an honest account of what switching involves: the timeline, what breaks, what it costs, who has to be involved. It converts because it answers the question the buyer is privately worried about.
Landing pages built for one segment. Tier-two clusters justify their own pages. Bottom-funnel pages of this kind should convert at 3–8%; if a segment page sits well under that, the segment definition is probably wrong rather than the copy. Track them separately with the conversion rate calculator so you can see which clusters are real. Our notes on web design for B2B companies cover the page mechanics in more depth.
Anything a rep can send that is not about you. A short teardown, a benchmark, a piece of analysis of their public data. The test is whether the recipient would forward it internally.
For distribution, LinkedIn remains the only mainstream platform where company-level targeting is precise enough to matter for tier one and two, though the CPMs demand discipline about who you show ads to — see our breakdown of LinkedIn marketing services and what they should cost. Paid search still has a role for capturing the category and competitor terms your targets will eventually search, and the organic side compounds if you build the segment pages properly, which is largely what a good B2B SEO programme is for.
Running it with sales
The operating rhythm matters more than the assets. Three things need to be true.
A shared account list that both teams can see and edit. Not a marketing spreadsheet that sales was emailed once. If a rep says an account is dead, it comes off the list that week and the media spend stops. If a rep adds an account, marketing picks it up in the next cycle.
A defined response to signals. Decide in advance what happens when a tier-one account shows three or more engaged contacts in a fortnight, or when someone from a target account visits pricing twice. Who reaches out, within how many hours, saying what. Programmes die in the gap between marketing seeing a signal and anyone acting on it.
A weekly meeting that reviews accounts, not metrics. Twenty minutes, tier-one list on screen, what moved and what is stuck. Metric reviews happen monthly. Account reviews happen weekly, because that is the cadence at which you can still change the outcome.
Coordinating this across content, media, sales and research is a real project management problem once you pass a couple of dozen accounts. Teams that run it out of email lose track by week six; some structure around marketing project management is worth setting up before you launch rather than after it gets messy.
Measuring it without lying to yourself
ABM metrics are easy to game, and the ones that feel best are usually the least informative. A workable hierarchy:
Account coverage. What percentage of the buying committee at each target account do you have contact data and engagement for? This is the leading indicator nobody tracks. Coverage of two contacts at a ten-person committee means you are not in the deal, however good the engagement scores look.
Account engagement over time. Aggregate every touch across every known contact at the account into a single trend. You are looking for direction and breadth — more people, more senior, more often. A single champion consuming everything is a weaker position than four people each reading twice.
Pipeline created from target accounts, stated as a share of total pipeline. This is the number the CFO cares about and the one that should move first.
Win rate and cycle length, target versus non-target. The clearest evidence ABM is working is usually not more deals but faster ones with better win rates, because you entered earlier and covered more of the committee.
Cost per opportunity, not cost per lead. Media efficiency at the account level. Run the ROAS calculator on the programme as a whole rather than per channel — attributing an ABM win to a single touch is meaningless when the whole design is multi-touch and multi-person.
Expect the first two quarters to look bad on anything downstream. If your sales cycle is seven months, a programme launched in January produces its first honest read in the autumn. Agree that timeline with your executive team in writing before you start, or the programme gets cancelled at month five while it is working.
The failure modes worth naming
Buying the platform first. ABM software is genuinely useful once you know your list, your tiers and your response rules. Bought before that, it becomes an expensive dashboard that reports on a strategy you have not written.
Treating intent data as truth. Third-party intent signals are directional at best and noisy at worst. Use them to reprioritise a list you already believe in. Do not use them to build the list.
Personalising the wrapper, not the argument. Company name in the headline, generic content underneath. Buyers read the second paragraph.
Letting the list go stale. Target lists decay fast — people leave, priorities change, budgets get cut. Review the tier-one list monthly and the full list quarterly.
Running it as a marketing programme. If sales did not help build the list and does not attend the weekly review, you have bought targeted advertising and called it ABM.
FAQ
How many accounts should a B2B account based marketing programme start with? Start with 20 to 40 tier-one accounts and no more. That is roughly what a team of four or five reps can genuinely work with personalised outreach, and it is enough to produce a readable result in two quarters. Add tier two and tier three once the tier-one motion is running reliably.
How long before ABM shows results? Engagement metrics move in four to six weeks. Pipeline follows in one to two quarters. Closed revenue takes at least one full sales cycle, so if your average cycle is seven months, plan for a nine to twelve month evaluation window and agree that up front.
Does account-based marketing replace inbound? No, and treating it as a replacement is a common mistake. Inbound builds the demand and the content library that ABM then aims at named accounts. Most effective programmes reuse 60–70% of existing content and only build bespoke assets for tier one.
What is the minimum deal size that justifies ABM? There is no universal threshold, but below roughly $20,000 in annual contract value the per-account cost of a tier-one play rarely clears. Below that, a tier-three programmatic approach against a well-defined list is usually the sensible version, and a stronger inbound motion is usually the better investment.