A team that buys data for B2B marketing and sales meets two kinds of product that are easy to confuse: identity platforms, which resolve people, and company-level data, which describes organisations. This guide sets out what each is built for, which jobs each does and where the two meet at the account, so that you can match a need to the right unit before you match it to a vendor. LiveRamp is the identity platform used as the example, and what is said of it comes from its own documentation, read on 4 October 2026.
The unit decides the job
Both kinds of product do some form of entity resolution, at different units. An identity platform resolves identifiers to people. LiveRamp's RampID documentation says its identity resolution turns known and pseudonymous identifiers into individuals or households and produces a person-based, pseudonymised view of customers that carries the identifier RampID. Its Data Marketplace documentation says buyers and sellers of marketing data transact at a person level. The jobs that follow from that unit are about people: reaching an audience, delivering a message to it in a channel and matching exposure to outcomes.
A company index resolves websites, listings and brands to one company. In Fokals each company has one stable company ID. A listed company carries its ticker, MIC, ISIN, LEI and share-class FIGI, and a brand or subsidiary carries the identifiers of its listed parent. Each row states a fact about the company, with its source and the time it was observed, and the facts are firmographic, technographic, hiring and intent data on one index. Fokals data is company-level throughout. The jobs that follow from this unit are about organisations: which to pursue, in what order, and what has changed.
The account is where the two meet. LiveRamp's B2B page describes using unified identity to reach decision-makers at target accounts across channels. Choosing the account is a company-level question, and reaching the people there is an identity question.
What a company-level record looks like
A person-level record identifies an individual or a household. A company-level record is a set of dated facts joined by one key. The table shows four Fokals datasets and what an illustrative row in each would say about Acme Robotics, an invented company. All four join on the company ID, and the data dictionary defines every column.
| Dataset | Columns read | What an illustrative row says |
|---|---|---|
Technology Stack (company_technologies) | domain, technology_name, first_seen_at, last_seen_at | A CRM tool first seen on the website on 1 October 2026, and seen again on 3 October 2026 |
Technology Changes (company_tech_events) | observed_at, change, key | An analytics tool added on 29 September 2026 |
Hiring Activity (company_hiring_daily) | day, open_postings, new_postings, by_function | The open postings on 2 October 2026, how many of them opened that day, and the open postings by function |
Intent Scores (company_intent_weekly) | week_start, topic, score, surge, evidence | A score for one topic in the week of 28 September 2026, the surge flag and the five strongest signals behind the score |
Every row describes the company. Each carries its source and the time it was observed, and a first observation of a website sets a baseline and is never counted as a change. The two kinds of record meet at the account: the company is the account, and the person works at it.
Which source fits which job
The table maps five common jobs to the unit each needs and the kind of source that serves it.
| Job | Unit it needs | Kind of source |
|---|---|---|
| Reach people in advertising channels | A person or household | An identity and audience platform, which is what LiveRamp's pages describe |
| Match campaign exposure to CRM outcomes | Exposure and your CRM records | A clean room, which LiveRamp's B2B page describes |
| Decide which accounts to pursue this quarter | A company | Dated company signals: hiring, technology changes, intent, announcements |
| Time outreach to a change at a company | A company | Dated company signals |
| Describe a company's stack, markets and languages | A company | Technographic and site data: Technology Stack and Website Profile |
The three company-level jobs need no person at all, and the account list they produce is the input to the person-level jobs: company signals choose the accounts, and an identity platform reaches the people at them.
A worked example: one account list, two systems
Acme Robotics, an illustrative maker of factory software, keeps its accounts in a CRM with one website per account. It wants to run an account-based campaign on the companies whose public signals point to a purchase, and to reach the people there through an identity platform. The work has five steps.
- Build a crosswalk. Normalise the website of every CRM account (lower case, no scheme, no
www.) and join it to thedomaincolumn of Technology Stack (company_technologies), which sits besidecompany_id. Keep the account, the company ID and how the match was made. This is domain matching, and a one-to-one match is the only kind to trust without review. - Measure the match. The query below gives the counts behind the match rate, which is matched accounts divided by accounts that have a website.
- Select by signal. Within the matched accounts, choose those with a surge in Intent Scores, a rise in new postings in Hiring Activity or a recent adoption or removal in Technology Changes.
- Hand over the list in the form the identity platform accepts. Which account identifiers it takes is set out in the platform's own documentation.
- Measure the outcome in your own data. Compare the pipeline from the selected accounts with that from a comparable set of matched accounts that were not selected.
-- Match of CRM accounts against the Fokals index (illustrative table names)
select
count(distinct a.account_id) as accounts_with_website,
count(distinct case when s.company_id is not null then a.account_id end) as matched,
count(distinct s.company_id) as distinct_companies
from crm_accounts a
left join (select distinct company_id, domain from company_technologies) s
on s.domain = a.website_domain
where a.website_domain is not null;Read the three numbers together. Matched divided by accounts is the match rate. If distinct companies is far below matched, many accounts share one company, which is the pattern of parent and child accounts, so decide whether a signal attaches to each account or only to the parent. Unmatched accounts fall into three groups: companies outside the index of listed companies, the brands they own and verified private companies; websites on which no technology is recorded yet, which have no row in Technology Stack; and accounts without a usable website. Sort a sample of the unmatched into these groups to see how many of the companies you track the index covers.
What each side is built for
Company-level data is built around the organisation. Coverage is the companies in the Fokals index: listed companies, the brands they own and verified private companies. Datasets refresh daily or weekly, and every observation is dated, so a trend can be read as the record stood on any day. An intent score is evidence-backed: every score carries the dated signals behind it.
An identity platform is built around the person. The RampID documentation describes resolution to individuals and households and does not describe company facts such as a technology added or a role opened. LiveRamp's B2B page says its marketplace can supply firmographic, technographic and intent signals from B2B providers, and for any such set the provider is the one to ask for the unit, the fields, the dates and the refresh.
Four questions that settle most cases
- What is the output? A person reached points to identity and audiences. A company chosen points to company-level signals.
- Where must it end up? Activated in an advertising channel points to a platform with destinations. Kept in your own warehouse or CRM points to a feed that you load.
- Must each fact be dated? A trend or a back-test needs the time each fact was observed, and rows that are not revised later.
- What may you do with it? Internal use, embedding in a product and redistribution are separate permissions in a licence, and a marketplace's terms can differ from a direct licence.
Fokals is delivered direct, by REST API and as bulk files, which you load where the account list lives, in your warehouse or your CRM. The intent dataset holds the weekly scores used above, and the guide to account-based marketing segments from signals goes further on step 3. The companion guide to the LiveRamp Data Marketplace describes how a buyer works in it.
Frequently asked questions
What is the difference between account-level and person-level data?
Account-level data describes an organisation: one record per company, keyed on a company identifier, holding facts such as the technologies it runs or the roles it has opened. Person-level data describes an individual or a household, as in the identity resolution LiveRamp documents. The two meet at the account: company-level data chooses which accounts matter, and a person-level source reaches the people at them.
What does a company-level record contain?
One record per company, keyed on a stable company ID, with facts such as the technologies it runs, the roles it has opened, what it announced and weekly intent scores. Each fact carries its source and the time it was observed. A listed company also carries its ticker, MIC, ISIN, LEI and share-class FIGI, and a leadership change is recorded by the role concerned.
Can company-level signals be used alongside an identity platform?
Yes, through the account list. Match your accounts to the company index on website domain, select accounts by signal, and give the list to the platform in whatever form it accepts. Which account identifiers a platform takes is for that platform to say. Fokals is delivered direct, by REST API and as bulk files, and your team runs the matching and the hand-over with its own tools.
Do I need an identity platform if I already have a company-level feed?
Only if the job is to reach people in channels or to measure exposure against outcomes. Choosing accounts, timing outreach and describing a company's stack are company-level jobs, and a feed answers them without an identity platform. If your team must reach people at the chosen accounts in advertising channels, an identity platform such as the one LiveRamp documents is the kind of source that does that.
How do I test whether a company-level feed covers my accounts?
Normalise the website of each account, join it to the feed's domain column and read the match rate by segment. Sort a sample of unmatched accounts into companies outside the index, websites with no technology recorded yet and accounts without a usable website. The Fokals index holds listed companies, the brands they own and verified private companies.
The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.
What this page says about the products it names was checked against their public documentation on 4 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.