Account-based marketing trends: What's real, what's hype

Most ABM "trends" are new labels on old account-level logic. A few are genuine shifts in how the work gets done. The honest split: rebranding hasn't fixed the core problem. Most ABM programs still can't reliably drive engagement from the target accounts they pick, because the unit of execution never changed. This article draws a hard line between what actually changes what you should do tomorrow and the repackaging that doesn't.

TL;DR

• Rebranding hasn't fixed ABM's core problem: most programs still can't reliably engage the accounts they target.
• Contact-level is the real unit of ABM work: individuals decide, not accounts.
• Dynamic, behavior-driven audiences have replaced static account lists as the operational standard.
• AI intent scores often hide the signal behind a black box teams can't interrogate or improve.
• Lookalikes and account-level blasts relabeled as ABM are still the wrong unit of targeting.

Contact-level is the new unit of ABM work

ABM was built around the account as the atom of execution: pick a list of target companies, run coordinated campaigns against those domains, measure account-level engagement. For a long time, that beat undifferentiated lead gen.

The problem is structural. Buying groups are larger and harder to map than when the account-level model was designed. As Forrester's analysts have put it, unless B2B marketers can pinpoint their individual buyers, no amount of account strategy fully succeeds.

The hardest part of ABM isn't finding the account. It's identifying the right buyer within it, then engaging that buyer at the right time. Both are contact-level problems, and account selection doesn't touch either one.

The gap account selection can't close

Account-level targeting cannot solve a contact-level access problem. Targeting a company domain without knowing which stakeholder is in-market leads to inefficient spending and lower engagement. You might generate an account-level signal (someone at Acme Corp visited your site), but that signal tells sales nothing actionable. Which person? Were they the decision-maker or an intern doing research? Should the SDR call the VP they already know or try to find someone new?

Contact-level precision isn't a feature launch or a vendor positioning play. It's a direct response to a structural gap the engagement data has documented for years. Accounts don't buy. People do.

Dynamic audiences are replacing static account lists

Static lists decay the moment they're built. The VP of Engineering researching data integration last Tuesday may have gone dark or signed a contract by Thursday. A competitor product announcement can trigger re-evaluation across an entire category overnight. You can't track any of that on a quarterly refresh cycle.

Audiences must update based on behavior, not a quarterly schedule. When a contact visits a pricing page, their audience membership updates. When a competitor's brand spikes in search, the segment capturing that behavior expands automatically. Tech providers running ABM programs see pipeline lifts of 11 percent over traditional demand generation, according to Gartner. The difference comes from behavior-triggered signals, not firms running the same fixed list through a new platform.

The timing layer is what makes contact-level precision useful. Finding the right person only matters if you reach them during an active research window, and behavioral data is what identifies that window, per Grand View Research's ABM market analysis.

Cookie deprecation forced a real identity rethink

Third-party cookies were the infrastructure for retargeting and intent enrichment. When a contact visited a review site, cookies carried that signal back to your audience models. That pipeline is largely gone. Firefox and Safari deprecated cross-site tracking years ago, Chrome has repeatedly delayed its own timeline, and regulations keep tightening. Inference built on third-party behavior, the foundation of most legacy intent products, is now far less reliable.

First-party identity resolution fills that gap. On-site de-anonymization (identifying the individuals visiting your own properties) replaces cross-site inference. Direct signal capture from owned touchpoints (site visits, ad clicks, form fills, pricing page views) becomes more valuable precisely because third-party signals have degraded. The ABM market is projected to grow from $1.4 billion in 2024 to $3.8 billion by 2030 at a 17.9 percent CAGR, driven largely by this shift to first-party data strategies (Grand View Research).

AI-powered intent scores that hide the signal

AI-driven intent scoring is real. Models can process behavioral signals at scale and frequency no rules-based system can match.

The transparency problem

The hype is the opacity. Most intent score products surface a number (a contact's score went from 42 to 71) without showing which behaviors drove it. Did it jump because three people at the account visited your pricing page? Because someone searched for a competitor twice? The sales team gets a flag but no causal chain. They can't explain to the prospect why they're calling now, and they can't tell whether the model is firing on signal or noise.

Hidden signals compound the measurement gap. Most ABM teams still track email opens and account engagement rates rather than business impact. Only 52 percent of companies measure ABM ROI at all, according to ITSMA. When you don't track outcomes against business impact, a black-box score never gets interrogated. You never find out whether acting on it produced pipeline.

AI intent scoring adds value when the underlying signals are visible and the model is auditable. The problem is platforms that give you the output without the inputs: a number you can't explain to sales, and a model you can't inspect or improve.

Account-level targeting in new packaging

Targeting a company domain without knowing which individual is in-market is not a trend. It is the original problem ABM was supposed to solve, dressed up in "ABM" language and sold as though the label changes the mechanics.

The bottleneck in ABM isn't finding the account. It's getting from the company to the person. Account-level targeting leaves teams exactly at that bottleneck. Campaigns run against company domains while the decision-maker reads competitor content and stays a ghost in your attribution.

When a vendor offers "account-level personalization at scale," ask what that means at the contact level. If the answer is "we target the company's IP range," the product isn't solving the engagement gap. It's the same mechanics that produced it.

Lookalikes and broad retargeting dressed up as ABM

Lookalike audiences work on opposite logic from ABM. ABM starts with a defined list of accounts and contacts you've deliberately selected. Lookalikes start with a seed audience and statistically expand outward to find similar users.

Probabilistic expansion trades precision for volume. It may find companies that look like your best customers, but it also pulls in companies that match the firmographic profile but aren't in-market. The account list dilutes with every degree of separation from the original seed.

Workbar ran this experiment directly. They dropped platform-native audiences and lookalikes and rebuilt their paid channels around contact-defined segments. Meta and Reddit relaunched using only contact lists layered with geo targeting. The result was measurable lift in paid quality, higher conversion rates, and fewer wasted impressions. Not because they found a new channel. Because they stopped expanding to look-alike noise and started targeting known people.

Seeing contact-level ABM at work

Airbyte: contact-defined intent segments on LinkedIn

Airbyte moved from targeting LinkedIn at the account level to segments built on individual intent signals. These segments identified people actively researching data integration or visiting competitor properties, not just companies that fit a firmographic profile.

One campaign generated about 146,866 impressions and about 12,000 clicks at a $0.64 CPC. Across key campaigns, CPC ranged from $0.50 to $2.00. LinkedIn's platform average for B2B campaigns runs far higher. Airbyte's demand generation lead independently confirmed these results in a public LinkedIn post, attributing the performance to contact-level audience construction rather than account-domain targeting.

The mechanism is the same whether you're running LinkedIn, Meta, or Reddit: identify the individual behind an anonymous signal, then build the ad audience from that contact rather than the company they work for. Ad Reveal does this at paid-media scale, identifying the person behind ad clicks without a form fill. Signal-driven audiences sync those contacts to ad platforms automatically, so they stay current as behavior changes.

What contact-level targeting means for measurement

Contact-level targeting changes what attribution looks like. You know which individual clicked the ad, visited the pricing page, and showed up in the pipeline. Sales gets a name, not a domain. The spend stops disappearing into account-level dashboards that can't tell you which of it reached a real buyer, which is where pipeline anxiety starts.

Teams measuring account engagement rates while buyers make individual decisions will keep filling the same gap. The teams closing it measure contact-level outcomes: did this person engage, convert, or enter the pipeline?

The unit of analysis is the real trend

The engagement gap isn't a technology problem or a budget problem. It's a unit-of-analysis problem, and adding another platform won't close it. The trends that move it (contact-level precision, behavior-driven audiences, first-party identity) all solve the same thing: getting from the account to the person. The trends that don't (black-box scores, account-domain targeting, lookalike expansion) all skip that step and call it progress.

Here's the harder claim: if your ABM motion can name which account is engaged but can't name the individual who visited your pricing page, you're not running ABM. You're running a campaign with account-level branding on it. The label doesn't change the mechanics. The data already proved that.