ABM does not start with ads.
It starts with the target account list (TAL).
If we get the list wrong, no amount of clever, creative, personalization, or sales follow-up will save the campaign.
This guide is everything I have learned about how to choose target accounts for ABM after running programs at Userpilot and now for ZenABM, and after watching hundreds of customers do the same.
You will get a step-by-step framework for picking accounts, a tiering model, the CRM signals I rely on most, and the intent and engagement layer that finally tells me which accounts deserve more budget.
Quick map of what I cover below: ICP fit, CRM mining (closed-lost, stalled, disqualified, and reactivation candidates), sales input, intent and LinkedIn ad engagement, tiering, and how to keep the list dynamic instead of frozen on a spreadsheet.
If you’re short on time, here is the short version of how to choose target accounts for ABM:
I will say it plainly: target account selection is the highest-leverage decision in any ABM program.
The CTR you obsess over, the creative tests, the personalization, the gifting budget: none of it matters if the wrong companies are in the audience.
I have seen teams burn $40,000 on LinkedIn ads against a list that was put together in 30 minutes from a Salesforce export, then wonder why nothing closed when the list was the problem the entire time.
The account list is, indeed, the foundation of everything that comes after.
Think about it this way: your ICP is the universe of companies that could buy, while your target account list is the subset you believe will buy this quarter or this year, and those are very different lists. The ICP can have 20,000 companies, but the target list is usually 50 to 1,500, depending on whether you are doing one-to-one, one-to-few, or one-to-many ABM.
If you want a deeper read on the underlying program design, I wrote a complete walkthrough in my ultimate guide to ABM on LinkedIn.
Every good account list starts with a tight ICP, and I do not mean a one-line description.
I mean a real definition with concrete attributes that a tool can actually filter on.
The ICP attributes I use:
Tim Davidson summed up the bigger picture nicely when he was thinking about how the funnel is collapsing:
“At some point, it’s just all going to be one bucket called target accounts. But I could be wrong. Prove me wrong. Regardless, the TOFU MOFU BOFU strategy doesn’t work because that’s not how people buy.” – Tim Davidson (founder at B2B Rizz) on LinkedIn
I agree with Tim.
The bucket I care about is “fits my ICP and shows signs of being a real buyer this year.”
Everything else is theatrics.
Pull your closed-won deals from the last 24 months and look at win rate, sales cycle length, ACV, gross retention, and NPS, because the accounts that score well on all of those metrics are your real ICP.
Not the ones a CMO drew on a whiteboard, not the logos you wish had closed, but the ones that actually closed, paid, expanded, and stayed.
From there, the ICP becomes the filter you apply to every list source going forward.
This is the step most ABM guides skip, yet it is the one that consistently creates the most pipeline for me, because cold lists are slow by definition, while the CRM is full of accounts that already had a touchpoint with my company: accounts that already saw a demo, already replied to an SDR, already signed up for a free tool.
Those are warm leads waiting to be reactivated.
The five CRM segments I always pull
Closed-lost accounts already know who you are, because they evaluated you, their objection is documented, and the buying committee is mapped in the CRM.
Compare that to a cold list where there is zero awareness, zero context, and zero relationship.
If you map the loss reasons by deal potential and re-engage the top three, you skip months of prospecting.
As Nick V. from HeyReach was quoted by Alex Fine:
“They map closed-lost reasons by deal potential and build the top three. No market theories. No founder ego. Just fix what’s actually breaking deals.” – Alex Fine on LinkedIn
The same logic applies to property history in your CRM.
If a company filled out a form three years ago, abandoned a trial last year, and visited your pricing page last week, that property history is doing the work of an entire SDR.
Surface it.
If I build the target list without sales input, the outcome is predictable: AEs reject it because “those companies will never buy,” SDRs ignore it because they have their own pet logos, and the list dies before a single dollar of media is spent.
I run a 30-minute working session per AE before locking the list.
I ask three questions:
This step does two things: it surfaces logos that no algorithm would ever pick, and it gets sales bought into the list before a single dollar of media is spent, with the latter mattering more than the former.
ICP plus CRM plus sales gets me a strong static list, and intent and engagement are what turn it into a living one.
The signals I prioritize the most
This is where my program changed the most, because with ZenABM, I track LinkedIn ad engagement at the account level and use that data to move accounts through five stages:
Once that mapping exists, the list is no longer a static spreadsheet but a queue.
Accounts in Considering get the most budget and the most sales attention, while accounts stuck in Identified for 90 days get reviewed: I either push more impressions to them or drop them and replace them with warmer fits.
For more on the engagement layer specifically, my piece on account scoring and engagement signals for ABM goes deeper into the math behind the stages.
Once I have a list of, say, 600 ICP-fit accounts with CRM context and engagement signals, I tier them, not because tiering is fashionable, but because Tier 1 and Tier 3 should not get the same treatment.
Here’s the model I follow:
| Tier | Profile | Treatment | Typical list size |
|---|---|---|---|
| Tier 1 | Highest-fit, highest-revenue, named by sales as priority | One-to-one: bespoke landing pages, custom outbound, ABM ads with company name | 10 to 50 accounts |
| Tier 2 | Strong ICP fit, often grouped by industry or use case | One-to-few: vertical-specific ads, persona landing pages, pod-based outbound | 50 to 300 accounts |
| Tier 3 | Broad ICP, programmatic reach | One-to-many: matched audience ads, lead gen forms, marketing-driven nurture | 500 to 5,000 accounts |
The mistake is to lock in tiers and never change them, because a Tier 2 account that is in Considering deserves more attention than a Tier 1 account that is still in Identified. Engagement is a stronger signal than how excited my CRO was when she first saw the logo, so I overlay the stage on the tier.
The accounts that get the most budget are the ones that are high tier and high stage: Tier 1 in Selecting gets a custom video and a dinner, while Tier 3 in Identified gets a single image ad and a low CPM bid.
If you want a deeper view of segmentation logic, we have written about tiered account segmentation for ABM with concrete budget splits.
The target account list is not a deliverable; it is a living object.
If your TAL has been the same for six months, something is wrong, because you are either not finding new fits, not removing dead weight, or not letting engagement promote and demote accounts.
Here’s the weekly TAL procedure I follow:
That last step is the difference between marketing as a lead factory and marketing as a pipeline machine.
If sales does not get a real-time signal when a Tier 1 account starts engaging, the engagement is wasted.
If you want to take this further into outbound, my guide on intent-based outbound shows the exact handoff workflow.
I see the same mistakes over and over, and if you eliminate them, you are ahead of at least 80% of ABM programs.
The list lives across a few systems, and the following is the minimum stack I run with:
The reason I built ZenABM is that LinkedIn Campaign Manager hides which companies actually engaged with your ads: you see the campaign-level numbers, not the account-level story.
That gap is what kills most ABM programs, because you cannot promote accounts through stages if you cannot see who engaged.
Here is a concrete pattern I have seen play out repeatedly: a team starts with a 1,200-account list built from a Salesforce export filtered on industry and headcount, then runs LinkedIn ads against it for 60 days.
After 60 days, the engagement picture is brutal but useful:
That last 10% is where the pipeline lives, which means the team should now do four things: increase budget on the engaged 10%, retarget the 40% with a different creative angle, run a fresh persona-targeted creative against the unmatched 30%, and cycle in 200 new closed-lost accounts to keep the list fresh.
Without account-level engagement data, none of those decisions is possible, and the team would just keep spending evenly across 1,200 accounts, never understanding why the pipeline stays flat.
The honest answer is that it depends entirely on your motion.
Here are the numbers I use as defaults:
You can run more than one motion in parallel, and in fact most strong programs do, with a common split being 25 Tier 1 accounts, 200 Tier 2 accounts, and 2,000 Tier 3 accounts, with budget split roughly 40/35/25 across the tiers.
The exact size depends on your sales capacity and average deal size.
Every creative decision, every budget split, every sales play you build sits on top of it, which means a weak list does not produce a weak campaign. It produces no campaign at all, because the right message aimed at the wrong companies is just noise with a media budget attached to it.
The framework I have laid out here is not theoretical. ICP filtering, CRM mining, sales alignment, intent layering, tiering with a stage overlay, and weekly list maintenance: these are the exact steps I run for ZenABM and the steps I have watched the best ABM programs run at companies that actually close pipeline from their LinkedIn spend.
The part most teams skip is the dynamic layer.
They build the list, launch the campaign, and then watch the budget disappear into accounts that never move past Identified. The fix is account-level engagement data.
When you can see which companies are seeing your ads, clicking your ads, and engaging with your bottom-of-funnel themes, you can move budget to where the heat actually is instead of spreading it evenly across a spreadsheet that has not been touched in three months.
ZenABM tracks LinkedIn ad engagement at the account level, maps each company through the five ABM stages, and pushes that engagement data directly into HubSpot or Salesforce, so your sales team knows exactly when a Tier 1 account starts moving.
If you are running LinkedIn ABM without that visibility, you are optimizing blind.
Try ZenABM for free now (37-day free trial) or book a demo to know more!
Some common questions about building your TAL for ABM and their answers:
Start with ICP fit using firmographics, technographics, and growth signals from an enrichment tool like Clay or Apollo.
Then run a small one-to-many LinkedIn ad campaign against that list to generate first-party engagement signals. Within 60 days, you will have a real engagement layer to refine the list with.
For one-to-one ABM, 10 to 50 accounts. For one-to-few, 50 to 300. For one-to-many, 500 to 5,000 or more. Most teams run a blended program with all three tiers and split the budget roughly 40/35/25 across Tier 1, 2, and 3.
The exact size depends on your sales capacity and average deal size.
I update mine weekly.
New ICP-fit accounts get added, accounts stuck in Identified for 90 or more days get removed or replaced, and accounts that moved up a stage get promoted in budget and creative.
Treating the TAL as a static spreadsheet is one of the biggest mistakes in ABM.
Yes, especially closed-lost where the loss reason was timing, budget, or a competitor.
Those accounts already know your brand, have evaluated you, and have a documented buying committee.
Map loss reasons by deal potential and re-engage the top-priority ones with a reactivation campaign.
Combine tier and stage. Spend the most on accounts that are high tier and high stage (Tier 1 or 2 in Interested, Considering, or Selecting). Spend the least on accounts that are low-tier and low-stage.
Engagement is a better predictor of close than tier alone, so let the data move the budget for you.
