Data 360 can score every buyer perfectly and still quote one account two different deals.

Scenario: how one company ends up with two prices

Your AE, Priya, is three calls deep with Jane at Acme. To close by quarter-end she’s floated an aggressive 30% discount. What Priya can’t see is that two of Jane’s colleagues are running their own quiet diligence: Raj has hit the pricing page twice and pulled down the ROI whitepaper, and a third has been reading the security docs. Then Raj fills in a “talk to sales” form.

TL;DR — Account-level identity resolution isn’t a Data 360 feature you switch on — it’s three modelling decisions (what counts as one account, what counts as an engaged account, and which of the two segment patterns a job needs). Those decisions, not the clicks, are what stop your own systems from quoting one buying group two different prices.

It lands as a fresh lead. An SDR picks it up cold, has no idea Raj sits at the same company as Jane, and starts from zero — different pitch, list pricing, maybe a different discount entirely. The buying group compares notes internally, because buying groups always compare notes, and now Acme is looking at two numbers and two stories from one vendor. Best case, you look disorganised. Worst case, Raj’s standard quote just became the anchor Acme uses to push Priya’s 30% to 40%.

You competed with yourself — not because your data was dirty or your reps were careless, but because every person in that account was an island. This is the trap of person-level identity resolution: it’s brilliant at the individual and blind to the group. Data 360’s core object is the Unified Individual — every known email, device, and record for one human, stitched into one profile. Raj resolves perfectly; Data 360 can score his intent to two decimal places. None of that tells you the one thing that matters in this deal: that Raj is in the room with Jane.

Person-level resolution answers “who is this individual and what have they done?” It does not answer “which of my open deals does this individual belong to?” — because that question isn’t about the person at all. It’s about the organisation — a different object your model has to build on purpose. A person-level segment can be flawless and still route Raj to a cold SDR: you score every individual and still lose the account.

The fix is account-level identity resolution — resolving the individuals to the organisation so a buying group is treated as one account, not a scatter of strangers. But it isn’t a switch you flip. It comes down to three modelling decisions, and the rest of this post is those three:

• What counts as one account? — where you draw the boundary between companies.
• What counts as an engaged account? — how individual signals roll up, and the threshold that matters.
• Which of the two segment patterns does the job need? — accounts or people as the output.

Get them right and one company gets one conversation and one number; get them wrong and you get a segment that looks fine and quietly misfires.

Decision 1: “One account” is a boundary you draw, not a fact Data 360 hands you

Resolving the organisation is a second, separate identity-resolution job, sitting alongside the individual one — the same discipline you already apply to people (see our piece on identity resolution as the foundation of real-time personalisation), pointed at a different grain. And it’s the harder job, because the boundary is fuzzier. Two email addresses being the same person is usually a clean call — the strong key is right there. Two records being the same account means first agreeing what an account even is:

• Is Acme UK the same account as Acme US?
• Is a newly acquired subsidiary its own account, or part of the parent?
• Is a franchise one buyer, or forty?

Data 360 can match accounts on contact points (address or phone), on the CRM Account external ID, or on a fuzzy company-name match — and each choice draws the boundary in a different place. Over-merge and you fuse two real buyers into one blurred account; under-merge and Jane and Raj stay split, and the problem never moved. The identity-resolution match rules are the strategy — not a default you accept.

Once the boundary’s drawn, be clear what you’ve built — a common misconception wrecks account models here. People hear “unify the account” and picture all the contacts melting into one giant profile — a merged super-person with everybody’s emails and clicks in one record. That is not what a unified account is, and building it that way is how you end up emailing Jane content meant for Raj.

A unified account is a firmographic record — one Acme, with its industry, size, region, and CRM Account ID — plus a set of relationships to the individuals who belong to it. Jane stays Jane. Raj stays Raj. Their profiles are never fused; they’re linked to the same Unified Account through Data 360’s Account Contact relationship. The account is a hub with spokes, not a blender — and everything after this depends on it: signals reach the account (Decision 2) and activate back to people (Decision 3) only because it’s an identity graph of relationships you can traverse.

Decision 2: “Engaged” has no default — you define what the account must show

If you’re not merging the people, how does the account come to “know” that three of its contacts are engaged and one deal is open? By aggregation. In Data 360 that’s the job of Calculated Insights — metrics computed across the data graph and dimensioned by account: engaged contacts in the last 30 days, total web sessions this quarter, whether an open opportunity exists, the agreed discount floor on it. Each individual’s behaviour rolls up to the account grain while the individual records stay exactly where they are. For account and Unified Account segments, Data 360 also exposes a native Hierarchical Aggregation option — sum, count, or average rolled from child accounts up to the parent — so an Acme UK signal counts toward Acme globally without any record leaving its place.

This is where the decision lives. A Calculated Insight doesn’t know what “engaged” means — you define it. Is an engaged account one with two or more contacts active in 30 days? Three in 14? One senior contact plus an open opp? There is no default that’s right for your business, and the definition you choose silently decides which accounts land in the segment. Worse, Calculated Insights are batch by default, so “the account is in active negotiation” can be up to a refresh cycle stale — fine for building an ABM audience, a real problem if you’re trying to stop an SDR cold-quoting a live deal in the next hour, which is when you’d reach for Streaming or Real-Time Insights instead. The aggregation is easy to switch on; choosing the roll-up logic, the thresholds, and the freshness tier it needs is the work.

Decision 3: Two segment patterns — and the wrong one gives you a segment you can’t action

Once behaviour is aggregating to the account, there are two different ways to build an account-level segment, and they’re not interchangeable:

• Accounts-as-population — the segment’s members are accounts. You filter the Unified Account on firmographics (ICP fit), on aggregated engagement (the Calculated Insights and Hierarchical Aggregation above), and on deal state (open opportunity exists, discount floor agreed). The output is a list of companies.
• Individuals-qualified-by-account — the segment’s members are people, filtered by attributes of the account they’re related to. The output is a list of humans who happen to sit inside qualifying accounts.

You need both, because they do different jobs — and reaching for the wrong one leaves you with a segment you can’t action:

• Individuals-qualified-by-account — the one that stops the cold-quote. In plain language it reads: “every contact related to an account that has an open opportunity and an agreed discount floor.” Its members are people, so you can route them, alert the owning rep, or suppress them from cold outreach — exactly the individuals an SDR would otherwise pick up blind.
• Accounts-as-population — the one you hand to ABM. Its members are companies, so it drives a company-targeted ABM campaign or a named-account list for sales. You can’t email it directly, and you’re not meant to.

Same model underneath, two different populations — and the failure mode is mismatching them. Reach for accounts-as-population when the job is to touch people and you get a list of companies with no inbox; reach for individuals when you meant to target accounts and your ABM campaign chases scattered contacts instead of buying groups. That mismatch is how a technically-correct segment ends up doing nothing.

Accounts have no inbox: activation means stamping context back onto people

An account can’t be emailed. It can’t be served an ad or pushed to an SDR’s queue. So every account-level segment has to fan back out to individuals at activation — which works only because you kept the people separate instead of merging them.

Activation takes three practical shapes:

• Alert and route. When a contact belongs to an account with an open opp, the SDR sees the deal, the owning rep, and the floor before they touch the lead — no more cold quotes.
• Export the account list. Push the qualifying accounts to your ABM platform, or back into CRM as a target list.
• Stamp account context onto every related contact — the workhorse. Write the open deal, owning rep, and discount floor down onto each person, so whoever picks up Raj, in whatever system, starts with Jane’s context already attached.

That last one is what turns “one company, two prices” back into one conversation. The honest limit: if you have no CRM opportunity data to stamp, or your reps don’t work leads in a system that can read these fields, the best account model on earth changes nothing. Activation is where account resolution either pays off or dies.

None of this is a toggle — you modelled it

Step back and the trap reappears in a new form. It’s tempting to read all of this as configuration — flip on account resolution, add a Calculated Insight, save a segment. But every part that looks like a setting is a decision someone made:

• The account “knows” about the open deal and discount floor only because someone ingested the CRM Opportunity as a data model object and related it to the account — the number didn’t arrive by itself.
• The account boundary is only correct because someone chose the match rules over the three plausible alternatives.
• The segment actions cleanly only because someone picked individuals-qualified-by-account over accounts-as-population for this specific job.

Every one of those is a decision, not a setting. And since Data 360 is a packaged platform, say it plainly: the tool doesn’t save you from the decisions — it just makes them fast to implement once you’ve made them. You’d face the identical choices building this warehouse-native on your own data; the discipline is the same whichever stack you’re in. It’s a strategy problem wearing a configuration costume. The three steps are trivial; the decisions inside them are the job.

One company, one conversation, one number

Raj got a second quote not because of bad data or a careless SDR, but because the buying group existed in the real world and nowhere in your model. The three decisions above — one account, engaged, which pattern — are how you build that missing object; get any one of them wrong and the segment quietly misfires again.

So don’t start by rebuilding your identity model. Start smaller: pull your open opportunities and count how many have contacts scattered across more than one account, or sitting on no account at all. That number is the size of the hole — the deals where your own systems can currently quote against you — and it’s usually bigger than people expect. For B2B sellers with real buying groups it’s worth closing; if you’re B2C or product-led with no buying committee, it’s overhead you can skip.

Ready to build the account grain properly

If you’ve found that hole and want the modelling done right — the match rules, the Calculated Insights, the two segment patterns wired to real activation — that’s the work we do. Take a look at how we approach account-level identity resolution and segmentation, and bring the messiest account boundary you have; it’s the fastest way to see whether this pays off for your data.

Frequently Asked Questions

What is account-level identity resolution?

It’s the process of resolving individual customer records to the organisation they belong to, so a company is represented as one account with relationships to its people — rather than as a scatter of unconnected individual profiles. It sits alongside person-level identity resolution as a second, separate job, and it’s what lets you treat a B2B buying group as the single decision it actually is.

How is it different from person-level identity resolution?

Person-level resolution stitches all of one human’s records into a single profile; account-level resolution links those individual profiles to a shared organisation node without merging them. The person grain answers “who is this individual?”; the account grain answers “which company and which deal do they belong to?” You need both, because journeys act on people while buying decisions happen at the account.

Can you segment by account in Salesforce Data 360?

Yes. You aggregate individual behaviour to the account using Calculated Insights (or the Hierarchical Aggregation option on account segments), then build segments either on the Unified Account directly (accounts-as-population) or on individuals filtered by their account’s attributes (individuals-qualified-by-account). The two produce different populations for different jobs, so the choice of pattern matters as much as the filters.

What counts as “one account” when the same company appears many times?

That’s a modelling decision, not a given — and it’s the hardest part. You have to decide whether regional entities (Acme UK vs Acme US), acquired subsidiaries, and franchises are one account or many, then encode that in the identity-resolution match rules. Matching on contact points, the CRM Account external ID, or a fuzzy company-name match each draws the boundary differently, so the ruleset is where you set strategy, not a default to accept.

Why do B2B buying groups need account-level resolution?

Because a purchase is made by a committee, not an individual, and person-level data can’t see the committee. Without the account grain, two people from the same company can enter your funnel separately, get handled by different reps, and receive conflicting pricing — letting you compete with yourself. Resolving to the account keeps one company to one conversation and one consistent number.