When LiveRamp ran the ABM campaign that people still cite, it did two very different things and got the order right. First it used data to identify 15 Fortune 500 companies that fit its ideal customer profile. That is a job for software, and the software did it well. Then it concentrated real, coordinated, human effort on that short list, and reported a 33 percent conversion rate in four weeks. Notice the sequence. The machine narrowed the field. People won the accounts. The tools did the finding. Humans did the closing. That division of labor is the entire answer to how you run account-based marketing today without becoming one more automated voice in a muted inbox.
A good hunting dog does not shoot anything. The dog finds the bird, points, and holds. The hunter walks up and takes the shot. The dog is astonishing at the one thing it does and useless at the other, and a smart hunter never once expects it to pull the trigger. Modern ABM software is a very good dog. The trouble started when the industry asked it to do the shooting too.
So here is the practice, the way I would run it now with everything the machines can do and everything they still cannot. It is not complicated. It is just disciplined about which job belongs to the tool and which belongs to you.
Let the machine do what it is genuinely good at, which is finding. Intent data, for all the ways it gets abused, answers one question well: which companies are thinking about my kind of problem right now. That is real information and it is hard to get any other way. Use it. Let the software watch the whole field and raise its hand when an account shows signs of a live need. Let it map the org chart and tell you who the players are. Let it save you the weeks you would otherwise spend just figuring out where to point. This is the dog finding the bird. Take the gift.
Then, at the exact moment the machine wants to fire an automated sequence, stop. That is the fork in the road, and it is the only decision in this whole series that matters. The standard playbook says: signal detected, launch the cadence. The original discipline says: signal detected, now a human does the real work. Everything good hangs on choosing the second path when every tool and dashboard and quota is pushing you toward the first.
What does the human do with the moment? The same three things people did before the software, only now aimed with far better information.
First, you research the person, not just the company, and you research the way the GumGum team did before it built that comic for T-Mobile's CEO. The tool told you a company is in-market and named the executive who owns the decision. Good. Now go do the slow work the tool cannot. Listen to the podcast she was on. Read the talk he gave. Find the offhand comment where she named the thing that actually keeps her up at night, which is almost never the thing the intent signal flagged. GumGum's whole campaign turned on one human detail, a CEO's taste in comic books, that no data feed would ever surface, because it was said out loud to other people and never became a data point. The machine gives you the door. Your attention gives you the key.
Second, you make one thing that could only have been made for this one buyer, and you make it useful before it is a pitch. Not a merge field with her name in it. Something that proves, the instant she encounters it, that a real person spent real hours on her situation and hers alone. Intridea proved it with a billboard aimed at one office window. GumGum proved it with a single comic book. Yours might be a short written breakdown of a specific problem in her operation, a genuinely useful introduction to someone she should know, or a physical thing that references the exact detail you learned from her podcast, so she knows for a fact a computer did not send it, because a computer could not have listened. The test is simple. If the gesture would still make sense sent to a hundred other people, throw it out. It has to be unfakeable, and unfakeable means it dies the moment you try to copy it.
Third, you carry yourself with the patience of someone who is not desperate, because you are not. You have ten or twenty of these going, not a thousand, so no single one has to close this week for you to eat. That freedom shows. Gartner found that 61 percent of buyers now prefer to buy without a rep at all, which tells you how badly the pushy, automated approach has poisoned the well. The seller who follows up on the buyer's timeline, with things worth receiving, and leaves the door open instead of kicking it, becomes the rare exception to everything that soured those buyers in the first place. He is being treated like a person by a person who is in no hurry and has clearly done the homework, which by 2026 reads almost like a miracle.
Now the fair objection. This does not scale, and your leadership wants scale. Here is the answer, and it is the answer to the whole series. It was never supposed to scale, and the refusal to scale is the point. LiveRamp did not win by touching more accounts than its competitors. It won by concentrating on fifteen. Ten accounts worked with real attention will beat a thousand sprayed with automated attention, not by a little, but by the kind of margin that changes a career. The math only looks wrong if you count activity. Count won deals of real size and it flips.
So use the tools without apology, and put them firmly in their place. Let them find. Let them map. Let them save you the search. Then close the laptop and do the part that has been valuable since long before any of this software existed and will be valuable long after the current crop is forgotten. Understand one person better than anyone else bothered to. Prove it with effort no machine could counterfeit. Wait, usefully, until the moment is right.
The machine points. You take the shot. Confuse those two jobs and you become one more automated voice in an inbox that is already full of them. Keep them straight, and you become the thing the whole market has automated itself out of being able to offer, which is a person worth trusting, showing up at the right door, at the right time, having actually done the work.
That is not a tired technique. It is the oldest one there is, and it still works.
Sources: - [LiveRamp's 15-account Fortune 500 ABM results](https://www.cognism.com/blog/account-based-marketing-examples) - [GumGum's "T-Man and Gums" comic for T-Mobile's John Legere](https://www.cognism.com/blog/account-based-marketing-examples) - [Intridea's billboard aimed at Ogilvy](https://www.cognism.com/blog/account-based-marketing-examples) - [Gartner: 61% of B2B buyers prefer a rep-free experience](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-sales-survey-finds-61-percent-of-b2b-buyers-prefer-a-rep-free-buying-experience)