I've been running ABM programs for the last seven or eight years, and I'm running one right now. Most of that work hasn't been campaigns. It's been deciding who the buyer actually is, mapping who sits on each buying committee, and getting four teams to work from the same list of accounts.

The campaigns themselves will look familiar. Webinars and reports - owned and partnered. Short pieces of research written about one account's own situation. Field events, and the follow-up sequences that run after them. Some of it has worked better than I expected. The account research in particular has opened more doors than any clever email. Some of it hasn't. Most teams treat ABM as a type of campaign. It only works when a company treats it as a commitment.

Most of what follows comes out of my own ABM work, my conversation with Mason on the GTM & AI Operators Show, and a discussion with Mohita Nagpal. Mason is the founder of Scrappy ABM, which has helped mostly mid-market SaaS companies source around $100 million in revenue over the past three years. Mohita is VP of Marketing at Sprinto, and previously led marketing at Borneo, Facets Cloud and Hiver.

You can watch my conversation with Mason Cosby on account-based marketing here.

Should you do ABM at all?

Start with product-market fit. ABM means choosing a defined set of companies and marketing only to them. If your first reaction is "but I'll miss out on everyone else," you're not ready, and that's fine.

The second test is money. In absolute dollars, ABM costs more per customer than most other approaches, so it only works when each customer is worth a lot. Most teams check this the wrong way, by looking at the cost to acquire a customer on its own. The better test is lifetime gross profit against that cost.

Take two customers. One pays $10,000 and costs $2,000 to win, a 5:1 return. The other costs $15,000 to win but is worth $100,000 over its life, because it expands in years two and three. On acquisition cost alone, the first looks better. Over the lifetime, the second earns far more. ABM is built for the second kind.

The cheaper customer to win isn't always the better one. Judge ABM on lifetime profit.

The common SaaS rule of thumb is that lifetime value should be more than three times acquisition cost. Mason models ABM at 10:1 over the customer's lifetime, for one practical reason: acquisition costs only ever go up. If the model works at 10:1 on paper, you'll probably land at 3:1 to 5:1 in year one, which is still a good business.

None of this applies to most early-stage companies, and the honest answer for them is no. The exception is a product built for a very small, specific market, where you can name every possible buyer. Mohita agrees, and adds a cost people forget. Even when ABM doesn't need much budget, it needs a lot of time and coordination between teams. Early on, you're better off with SDRs booking as many conversations as possible, so you learn who your best customer is and which message lands. ABM comes after that, once you can say something like "we sell to fintech companies in this region" and size the market behind it.

The 30-second ICP test

Most companies reach for ABM because they want to move upmarket. The risk is moving so far up that the target list no longer looks like anyone who actually buys from you.

Here's a test that takes 30 seconds. Open your CRM and build a list using every criterion in your ICP. Then add one more filter: current customers. If the list comes back empty, you don't have an ideal customer profile. You have a wish list.

"If you have zero customers of your ideal customers, it's not an ideal customer. That's aspirational."

Mason Cosby, founder, Scrappy ABM

A real ICP shouldn't be your whole customer base either. It's the best 20% of it, the customers who pay the most, stay the longest and get the most value. The job of ABM is to go and find more companies like them.

I'd add one thing from my own programs. A good ICP describes people as well as companies. The question isn't which job titles contain the right keyword. It's who inside the account actually owns the outcome your product delivers. Get that wrong and you'll market well to the right companies and still reach the wrong people.

Once the list is real, split it. A small group of must-win accounts gets highly personalized work. The rest get a lighter, steady nurture. Treating every account the same is how ABM quietly turns back into demand gen.

If your ICP doesn't include any current customers, it's aspirational, not ideal.

Why most ABM programs fail

You'll see a figure online that around 80% of ABM programs fail. I couldn't trace it to any study, so treat it as a practitioner's estimate. But anyone who has run ABM knows the direction is right. Most programs fail for one of four reasons.

You'll see a figure online that around 80% of ABM programs fail. I couldn't trace it to any study, so treat it as a practitioner's estimate. But anyone who has run ABM knows the direction is right. Most programs fail for one of four reasons.

1. Sales and marketing aren't actually aligned. ABM asks sales to behave differently. Instead of waiting for an account to come inbound, reps start connecting with people across the buying committee early, so the relationship exists when the account warms up. Mohita goes further: ABM can't be a bottom-up or marketing-only decision. Sales, marketing and strategy or revenue operations need to agree on the market, the target account list and how they'll go after it, and put those commitments in writing.

2. The company isn't a fit. A $10,000-a-year product that customers keep for two years rarely pays back the cost of ABM. A half-million-dollar deal does.

3. Nobody is resourced to run it. Many teams treat ABM like a seasoning. Leadership asks for a few accounts to be "ABM'd," nothing else changes, and close rates stay where they were. Calling it ABM instead of demand gen doesn't change the results.

4. Nobody can measure it. Most marketing tools track people, not accounts. Mason has worked with a $300 million company whose CRM wasn't set up to report on accounts at all. If you can't see that ten people from one company visited your site in the past two weeks, you can't tell whether ABM is working.

"You can't growth hack your way to ABM."

Mohita Nagpal, VP of Marketing, Sprinto

That explains why so many startup programs stall. Startup marketers are trained to find a shortcut, test it and scale it. ABM doesn't reward that. Without sustained effort, the best you'll see is the odd win that nobody can repeat.

I've seen the measurement problem up close. When I audited account ownership on one program, a large share of accounts weren't routed the way we'd agreed, and many had no owner at all. On another occasion, several reps had reached out to the same person at the same target account. None of that shows up in a campaign report. It shows up as an account that stops replying.

Most ABM programs fail for one of four reasons, and none of them is the campaign itself.

ABM is eight jobs, not one

The most common way to set up ABM is to hand it to a marketing manager who already has a full-time job. Mason calls her Susie. Susie already runs social, email and some content. Then the CMO asks her to "ABM" it all and make it more targeted. Now she also has to partner with sales, build enablement and design a targeting strategy, on top of everything she was already doing.

That's a cross-functional program handed to one person with no extra support, and it fails. The opposite fails too: announce to the whole department that "we're doing ABM now," and when nobody owns it, nothing happens.

A working program has eight roles. That's eight jobs, not eight people. One person can hold several of them.

1. Strategist - designs the program.

2. Content - writes the blogs, white papers, outbound sequences and ad copy.

3. Designer - makes it look good.

4. RevOps - builds the workflows and the dashboards.

5. Channel owners - run social, email and other channels, each with their own targets.

6. Outbound - reaches out to the people inside each account.

7. Executive sponsor - usually the CMO, who gets leadership to agree that channel teams will spend part of their time on ABM.

8. Execution owner - the project manager, or as Mason puts it, the cat herder who makes sure everyone does their part.

The last role is the one most teams skip, and it's the one I'd fill first. In my experience, sellers, outbound, webinars, events and LinkedIn all end up working the same account list. Someone has to decide which account gets which touch, and when. One rule I hold to: your hottest accounts never go into an automated outbound sequence. They get a person.

Eight jobs, not eight people. But every seat needs someone in it.

Your buyer decided before the first call

By the time a B2B buyer talks to you, most of the decision is made. 6sense's 2025 study of around 4,000 buyers found that buyers first contact sellers about 61% of the way through their buying process. 94% had already ranked their shortlist before reaching out, and 77% bought from the vendor at the top of it.

By the time a buyer calls you, the shortlist is usually already set.

So the real work of ABM happens in the part of the journey you can't see, and it runs in three stages. First, brand recall, which takes roughly three to six months of the right messages reaching the right people. Second, shaping how buyers think about the problem they have. Only then do they start looking at solutions, and by that point you want to already be on the list.

The buying committee is the other half of this. Mason reaches it three ways:

• Bottom up. Help managers and directors do their jobs better today, with practical, tactical content.

• Top down. Build trust with VPs and the C-suite. They don't need tactics. They need to believe you're credible.

• Around the account. Show up where the account will be, even when you can't target a specific person. You don't know who from an account will attend an event, but you know someone will.

In my own programs, the committee grows with the size of the company. At a small company it might be two or three people who decide. At the largest it can be ten or more, with as many again who influence the decision. We treat the two groups differently. Decision-makers get a direct ask for a meeting. Influencers get invited to webinars and events, so they have something worth forwarding to the people who decide.

Seniority matters too. When an outreach message makes a big claim, it lands better coming from a senior leader than from a sales rep. Senior buyers respond to people at their own level.

Give everything away

The thing that has shortened Mason's sales cycles most is also the most counterintuitive. He tells people exactly how to solve their problems, publicly, without them ever buying. There's no secret method held back for paying clients. People buy because they want someone to apply that knowledge to their company and do the work for them.

This works because teaching shapes thinking. When you explain how to approach a problem, buyers start to think about it the way you do, and you're on the shortlist before they've written one. It also changes how buyers find you. His team now gets inbound from people who asked an AI assistant for a good ABM provider, simply because they've published so much on the topic.

That channel is growing fast. In a November 2025 survey of 350 B2B buyers, generative AI had overtaken traditional search as a research tool for a quarter of them, and nearly two-thirds used it at least as much as search when researching vendors.

Mohita sees the same shift as a question of credibility. Mid-market and enterprise buyers have always checked whether a vendor is credible, through case studies, customer logos and analyst firms like Gartner and Forrester. AI assistants are now one more place they check, and you need to show up there. How much it matters depends on how sophisticated your buyer is.

The same idea works account by account. In my current program, the most effective personal touch hasn't been a clever email. It's been a short piece of research about the account's own situation, shared before any sales call. Most of the deals we won in one month had one attached. One lesson from that work: proof has to match the buyer's size. Sending a 200-person company a list of giant enterprise logos doesn't build trust. It tells them you don't work with companies like theirs.

The catch is time. Content you publish now may take 18 to 24 months to pay off. Most teams only plan for meetings this quarter, and they never build the audience of people who haven't heard of them yet.

The content multiplier

Giving everything away sounds like it needs a large content team. It doesn't. It needs a system that turns one piece of work into many.

Mason thinks of content as entry points. His main one is a podcast. Each episode gets republished, cut into clips, and its transcript turned into posts. All of it runs through Claude, which his team has loaded with more than four million words of his own writing, from social posts and call transcripts to the book he's working on. He says the result talks more like him than he does.

The next step is distribution. Today he's mostly on LinkedIn, with a podcast that has around 20,000 monthly listeners and early work on YouTube. He's hiring someone to take the same material, adapt it for six platforms and post up to 30 times a day. Not new content. The same ideas, reshaped for each place his buyers spend time.

"If you're looking for ABM, it's going to be really tough to not find me."

Mason Cosby, founder, Scrappy ABM

What goes into the system matters as much as the system. A simple rule for building an audience on LinkedIn: post what you wish you'd known two or three years ago. Everyone is a few years ahead of someone, and writing for that earlier version of yourself builds an audience of people who need exactly what you know. It's learning in public.

There's a stronger version once you have the experience. After 70 ABM programs, he now shares what he's seen work across them, not just what he's learning. For a B2B company, that's the version to aim for: patterns from real customer work that nobody else can publish.

One conversation, many surfaces. The ideas stay the same; only the format changes.

In person is the proof

Three years ago, Mason thought conferences were a waste of time. Now he's going to 14 events this year. What changed was AI. Buyers are more sceptical of everything they see online and keep asking whether it's real. In person, nobody can fake it.

He now splits his marketing in two. Social posts and email are the steady drip that says "we're still here, still useful." Events, webinars and full-day virtual workshops are the stamp of legitimacy that proves you're who you say you are.

Mohita reaches the same conclusion for a different reason. AI has made it easy to produce more of everything, at average quality, so every channel is saturated. An event is one of the few places you get a buyer's full attention, without a hundred other things competing for it. They see your brand from every angle at once. Her caveat is worth keeping in mind: as more companies run small events, those will get noisy too, so the format has to keep changing.

The money side is simple. A $30,000 sponsorship makes sense when it guarantees a speaking slot in front of the audience you want. Judge it the way you'd judge any channel: how many meetings did it book? Then extend the window to 90 days after the event, because that's where the follow-up turns into pipeline. In my own programs, sponsorships with no stage time and no real visibility underdelivered. Every event we ran got a threaded follow-up sequence over the following two weeks. And like any channel, events reward repetition. The first year at a conference is rarely the one that pays off.

Podcasts work as an in-person channel at a distance. Mason has seen guest acceptance rates of 50 to 80%, because people rarely say no to being interviewed. One of his clients, with about 2,000 LinkedIn followers and a single customer, launched a podcast to interview the people he wanted to work with. Eight months later he had six clients. Most people picture podcasting as a route to fame. The version that pays is a structured way to have real conversations with your target accounts.

Social and email keep you remembered. Events and live conversations prove you're real.

Nothing is dead. It just looks different

Every few years, someone declares a channel dead. SEO has died. Email has died. Events have died. All of them are still working for the teams that do them well.

His view on SaaS is simple. Before SaaS, companies bought software once and installed it themselves, then paid again for each new version. SaaS changed how software is paid for: ongoing, with the product improving along the way. That billing model may change again. Companies will still buy software. I've written separately about how this is reshaping customer success and post-sales teams, where the shift shows up most.

ABM has its own version of this story. Mohita doesn't see the question as whether ABM is more or less relevant. As long as nobody finds a better way to win mid-market and enterprise accounts, it stays essential. What has changed is the constraint. Personalization at scale and research on every account used to take real human time, and that time limited how much anyone could do. AI has removed that limit, so everyone can now run sophisticated ABM, and the channel gets noisier for it.

When everyone can do the tactics, the edge moves to what's still in your control: your positioning, your messaging and the story you tell about why you're different. Her point is that this story has to hold together across every touchpoint, from ads and events to roundtables, content and visual design, rather than feeling like a set of separate campaigns. Teams used to get away with average work because producing it took effort. Now anyone can produce average work at scale, so the only way to stand out is to rise above that sameness.

Most teams run ABM to book meetings this quarter, and conclude it doesn't work. The ones it works for are planting for a buyer who won't be ready for another 18 months, and will already know who they're calling when they are.

FAQs

What is account-based marketing (ABM)?

ABM is a B2B approach where you choose a defined set of target companies and focus your marketing and sales effort only on them, instead of trying to attract as many leads as possible. It treats each account, and the group of people who buy inside it, as the market.

When should a company start doing ABM?

After it has product-market fit and knows who its best customers are. Most early-stage companies should skip it and use broad outreach to learn their ICP and messaging first. ABM also needs high lifetime value per customer, because it costs more per account than other approaches.

Why do most ABM programs fail?

Usually for one of four reasons: sales and marketing aren't aligned on the target accounts, the business isn't a good fit for ABM, nobody has the time or support to run it, or the company can't measure engagement at the account level. Treating ABM as a quick growth hack rather than a sustained program is the common thread.

Who should own an ABM program?

One accountable person, with support. A working program covers eight roles: strategist, content, design, RevOps, channel owners, outbound, an executive sponsor and an execution owner. One person can hold several of them, but handing all of ABM to a marketing manager who already has a full-time job rarely works.

How do you build an ICP for ABM?

Start from your best current customers, roughly the top 20% by value and fit. Then test it: filter your CRM by your ICP criteria and add "current customers." If nobody shows up, the profile is aspirational, not ideal. A good ICP also names who inside the account owns the problem you solve.

How long does ABM take to show results?

Longer than most teams plan for. Brand recall alone takes around three to six months, sales cycles in mid-market and enterprise can run much longer, and content built to shape the market may take 18 to 24 months to pay off.

How is AI changing account-based marketing?

It makes personalization and account research far easier, which means everyone can do them and they stop being an advantage. Sharp messaging matters more as a result. AI assistants have also become a place where B2B buyers research and check vendors, so publishing useful content openly helps you show up there.