At one B2B client, LinkedIn data showed an average of 47 ad impressions at a company before a deal, with almost no engagement, and a report built on individual contacts missed most of it. At another client, lost deals averaged 1.3 people from the buyer's side and won deals averaged 4.5. Its win rate, the share of deals it went on to win, was around one in ten with its largest prospects, and the leadership team traced that to deals being worked through a single contact.
In both cases the buyer was a group of people at one company. Account-based marketing, or ABM, starts from that. You choose the companies you want as customers first, then point marketing and sales at the people inside them. Most B2B marketing works the other way around: reach a wide audience, collect leads and sort out the good ones afterward. ABM fits best when the market is short enough to name, such as a few hundred regional banks and credit unions.
Below is how I'd set up an ABM strategy without buying an ABM platform first.
The four decisions in an ABM strategy
Write down four decisions where sales and marketing can both see them, each with an owner and a way to check it.
| Decision | What to write down | Owner | How you'll check it |
|---|---|---|---|
| Which accounts | The list, a tier for each account and the reason it's on the list. | Heads of sales and marketing together | Sales can say why each tier 1 account is there. |
| Who at each account | The roles you need to reach, and which ones you already know by name. | The account's owner in sales | Known contacts by role, per account. |
| What you'll run | The programs for each tier and the message for each role. | Marketing | Every tier 1 account has something planned this quarter. |
| How you'll measure | The account-level numbers you'll report, and what you'll compare them with. | Marketing operations or revenue operations | A monthly report comparing target accounts with everyone else. |
Download the ABM planning sheet (CSV, with these decisions, a tier template and the measures below).
Pick the target accounts before the tactics
Start with the list. Pull your best customers by revenue and by how quickly they bought, and write down what they have in common: industry, size, the systems they already run and the event that made them start looking. Then find companies that match and aren't customers yet.
Put the draft in front of sales before marketing spends anything on it. A rep who has called on half these companies knows which ones just signed with a competitor and which ones are a bad fit for reasons the data doesn't show.
Split target accounts into tiers
I split the list into three tiers, and the tier decides how much custom work an account gets.
- Tier 1, one-to-one. A short list of accounts worth work written for that company alone, such as a custom page, a briefing for its executives or an invitation to a small dinner.
- Tier 2, one-to-few. Groups of similar accounts, such as one industry or one region, that share a message and a set of content.
- Tier 3, one-to-many. The rest of the list, reached with targeted ads and email. An account moves up a tier when more people there start paying attention.
Size tier 1 to what your sales team can work every week. If the reps can't name the accounts from memory, the list is too long.
Check the account data by hand
Account lists are usually built on enrichment data, the company details a third-party tool fills in automatically. Check it before you target anything with it. At one client, when five recent sales-qualified leads (prospects sales had agreed to work) were checked by hand, two had the wrong current vendor in the enrichment field. That field fed LinkedIn targeting, display ad targeting and the lists sales used for outreach.
Before you trust a field for targeting, read 10 or 20 records against each company's own website.
Map the buying committee at each account
A buying committee is everyone at the customer who has a say in the purchase: the person with the problem, the one who signs, the people who will use the product, and the ones who can stop it, usually IT, security or legal.
The 1.3 and 4.5 figures at the top of this page are why I'd treat the number of people engaged at an account as one of the main things an ABM program is trying to move. One interested person is a lead.
The person who fills out the form often isn't the buyer either. In one case, the executive who owned the problem often handed the evaluation to someone in marketing. The CRM, the system where sales tracks contacts and deals, credited that marketer's visit, and a report built on individual contacts missed the account's ad history. Rolling activity up to the company showed what the contact records left out.
For each tier 1 account, list the roles you need and mark the ones you know by name. The empty rows tell marketing who to reach and tell sales who to ask about on the next call.
What to run at target accounts
Pick programs your team can keep running for a whole quarter across the list. These are the ones I'd start with.
LinkedIn ads aimed at named companies
LinkedIn lets you upload a list of companies and show ads to people there by job title or function. To me, that makes it the closest thing in paid media to an ABM channel. Judge it at the account level instead of by clicks. The 47 impressions per company at the top of this page came with almost no engagement, and they were reaching those companies before a deal, on average.
Run the ads at a steady frequency to every tier, change the message by role, and report impressions and engaged people by account next to the pipeline from those accounts, meaning the open deals sales is working there.
Unpaid LinkedIn: posts, comments and messages
You can run much of this without an ad budget. In my experience, posts from the people who run the company reach more of a target list than posts from the company page, and a buyer at a bank or a credit union would rather hear from someone who knows the industry than from a brand. So have two or three executives post about what they see in the market, and point the company page at the topics your target accounts care about.
Sales does its part before it pitches. Reps connect with people across the buying committee at each target account, comment on their posts and on the industry groups they follow, and send a direct message only when they have something useful to offer. LinkedIn's Sales Navigator, its paid tool for sales teams, helps here: reps can save the target accounts and get alerts when someone there changes jobs or posts.
The trade-off is measurement. LinkedIn doesn't report unpaid reach by company, so log engagement from people at target accounts in the CRM, and ask new prospects how they first heard about you, in their own words.
Getting LinkedIn activity to the right seller
Most LinkedIn activity never reaches the seller who owns the account. Split the work so nobody has to watch everything. The executives who post reply to comments on their own posts, since LinkedIn already notifies them. Each seller follows their own accounts in Sales Navigator and answers their own messages. The one job left is matching the people who reacted to executive and company posts against the target list, and that can run once a week and become tasks in the CRM for the right seller.
Automate the matching, never the engagement. A like, comment or connection request sent by software reads as software, and in my view a buyer who notices is less likely to trust the next message from that person.
Content for each person on the committee
Each person on the committee has a different question. The executive wants to know what this costs and what happens if the company does nothing. The person who will run the product wants to know how hard it is to set up. IT and security want their security questionnaire answered before they have to send it.
So write for the role and for the segment. A bank and a credit union can buy the same product for different reasons, and one customer story from a company like theirs, told by someone in the same role, gives a reader more to go on than a general case study.
Give your champion, the person inside the account who wants you to win, something they can forward to their boss without editing. A one-page summary of the problem, the cost and the proof is usually enough.
Sales outreach to target accounts
ABM works best when reps contact the same accounts in the same weeks as the marketing, with the same message. Share the tier 1 list, the roles you're missing and what each account has already seen, and agree on who contacts whom.
Be careful with outreach nobody asked for. At one client, offers made to customers who had asked for one closed at about 40% over two months. Offers sent proactively to accounts that never replied closed at zero. That wasn't an ABM program, but the warning carries over. An account on your list hasn't asked you for anything yet, so lead with something useful to them before you ask for a meeting.
Partners at target accounts
If you sell with partners, check what they publish about you. At one client, a partner's content mentioned the client's largest competitor and left the client out, even though the two companies were planning to sell together. Ask partners which of your target accounts they already work with, and give them material that names you.
How to measure an ABM program
Measure accounts, and compare target accounts with the ones that aren't on the list. I'd stop using lead counts as the scoreboard, because in my view a good ABM quarter can bring fewer form fills and more deals.
Give an ABM program longer than a month before you judge it. At one client, deals closed a few days after the first sales conversation, which made the sales cycle look short. Review-site data showed the same companies researching the category for five or six months before that.
| Measure | What it tells you | Compare with |
|---|---|---|
| Accounts reached | Share of the list that saw anything from you this quarter | The whole list |
| Engaged accounts | Accounts where two or more people visited, replied or attended something. I'd start with two. | Last quarter |
| People engaged per account | Whether you're reaching the committee or one person | The same count on your recent won deals |
| Meetings and opportunities | Target accounts that reached a sales conversation or a qualified opportunity | Accounts not on the list |
| Win rate and deal size | Whether target accounts close more often, or for more | Accounts not on the list |
| Time to close | Whether the program shortens the sales cycle | Accounts not on the list |
For the HubSpot setup behind these numbers, start with How to build a B2B sales funnel in HubSpot.
Do you need an ABM platform?
Platforms such as 6sense and Demandbase add intent data, an estimate of which companies are researching your category, along with account-level ads and reporting. I'd consider one once the list is too big to watch by hand.
For a first list of a few hundred accounts, I'd start with what most teams already pay for: the CRM, LinkedIn's company targeting and a weekly review with sales. If after a quarter you're still guessing which accounts are in market, meaning actively looking to buy, that's the case for intent data.
Common questions about ABM
What's the difference between ABM and demand generation?
Demand generation reaches a wide market and lets buyers come to you. ABM picks the buyers first, and I'd usually run both: demand generation for the broad market, and ABM for the accounts few enough to name and big enough to matter.
How many accounts should be on an ABM list?
As many as your sales team can work. Tier 1 should be small enough that each account gets a plan of its own. Tiers 2 and 3 can be much larger, because their accounts share programs.
How long does ABM take to show results?
Usually longer than a lead campaign, because the accounts may be researching for months before they talk to sales. In the first quarters, judge it on engaged accounts and people per account. Judge deals and win rate once a full sales cycle has passed.
Read Why B2B leads aren't turning into revenue for how to trace one account from first ad to closed deal.