Persona-level intelligence contact profile analysis is a two-pass read of one stored contact. The first pass reads the persona, which is a type of stakeholder defined by function, level or custom criteria. The second pass reads the person, which is what was recorded about this individual and when. Only the second pass can support a message to a named person, because a persona describes a type, and the type's attributes are not observations of the individual.
Persona-Level Intelligence Contact Profile Analysis: The Short Answer
- Two objects share one name. NMS Consulting defines persona-level intelligence as information about a type of stakeholder: goals, priorities, pressures, buying role, decision criteria and objections. 6sense defines a persona as a job function plus a seniority level. A contact profile is about one person. Analysis that treats the first as the second turns a statement about a group into a claim about an individual.
- Vendor persona scores describe your history. 6sense calculates persona importance from three inputs: the average profile fit of contacts with that persona, the normalized count of such contacts in your CRM, and the average number of sales engagements with that persona before a won opportunity. The output reflects your own deals, not the contact in front of you.
- Some persona scores partly measure record hygiene. 6sense's contact profile fit model lists data cleanliness inputs, including name length, name entropy, phone number length, domain type, valid email and missing name, next to job function and level. A low score can come from how the record looks.
- Persona-level can be inferred, not observed. Intentsify says its persona-level resolution infers from page views and content consumption whether an anonymous visitor is likely a technical practitioner, a marketing leader or an executive.
- Act on Recorded lines and dated events. Use the persona pass to choose whom to look for and the person pass to decide whether and how to write. The Two-Pass Contact Read below sets out the checks and the outcomes: Act, Verify first, Research or Stop.
What the pages that rank for this phrase cover and miss
The pages that rank for this exact phrase use the term for five different objects. Each is useful in its own setting, and none separates the persona from the person.
| Page | What it answers | What it leaves open | What it cites or proves |
|---|---|---|---|
| Conversion Realtor, Persona-Level Intelligence: Contact Analysis for Real Estate | A real-estate model of buyer and seller archetypes (motivations, objections, timeline pressure, communication style) turned into a next action | Which profile lines were recorded and which were modeled; applies to residential transactions, not B2B accounts | Describes its own system; no external data |
| 6sense, Persona Importance Model (August 14, 2025) | A persona is job function plus job level; importance (weak, moderate or strong) is a weighted sum of three inputs; where the persona map appears in the product | It describes a persona's weight on past wins; it says nothing about a given contact's facts or how fresh they are | Vendor documentation of its own model |
| Draup, Persona Analysis AI Agent | An executive profile built from leadership history, responsibilities, strategic priorities, hiring patterns, tech influence and organizational signals, with a structured brief | Which lines are sourced facts and which are generated; it publishes a data-governance statement and says coverage varies by region and role | Vendor page; no per-line sources |
| NMS Consulting, Ideal Customer Profile vs Buyer Persona (October 9, 2026) | Persona-level intelligence as information about a stakeholder type; eight questions to capture, such as objectives, exposure, decision criteria and proof required | How the answers are stored on a contact record, and their freshness | Lists HubSpot, Salesforce, HBS Online, HBR and Gartner as sources |
| Intentsify, Account-Level Intent vs. Persona-Level Intent (October 29, 2024) | Account-level intent versus buying-group, persona-level intent, and why committees matter | Contact-level fields and when a persona label is an inference | Vendor view; cites "recent studies" for a committee size without naming them |
None of these pages says which lines in a persona or contact profile are Recorded and which are inferred, and none gives a decision rule that stops a persona statement from being quoted as a fact about a person. That decision rule is the gap the Two-Pass Contact Read fills.
What each level knows
Three levels of data sit under the same label, and each answers a different question. Intentsify draws the same three-way line for website visitors: account-level resolution identifies the company, persona-level resolution reveals the roles researching, and contact-level resolution matches a visit to an individual, and it adds that a resolved visitor is not a lead.
| Level | Question it answers | Typical content | What it cannot support |
|---|---|---|---|
| Account | Which company? | Firmographics, technology, signals | Who inside the company, or what they think |
| Persona | What type of stakeholder? | Function, level, custom persona criteria, importance score, typical objections | Anything about one named person |
| Contact | Which person? | Name, employer, title, email status, dated events | Why the person would buy, unless the person said so |
The Two-Pass Contact Read: the decision framework
The Two-Pass Contact Read is this guide's model, not a vendor feature or an industry standard. It sorts every statement in a stored profile into the pass that can justify it, and it ends in one of four outcomes. It builds on the Four-Tier Persona Ledger (Recorded, Mapped, Modelled, Guessed) from the persona-level intelligence contact summary guide, which owns the field definitions.
Pass 1: the persona pass (about a type)
| Check | Question | What passes |
|---|---|---|
| 1a Definition | Whose definition of persona is this? | You can name it: function plus level (6sense), title keywords, function, level and custom criteria (Demandbase), saved filters (Cognism), or a profile your team wrote |
| 1b Origin of the weight | What produced the importance or fit number? | The inputs are listed, as 6sense lists them, and you know they are your own history |
| 1c Sample | Does your history cover this offer, segment and region? | There are enough comparable past deals. For a new product, 6sense's listed inputs (CRM counts and pre-win engagement) have nothing to weigh, which is this guide's inference from the model's inputs |
| 1d Coverage | Which persona seats are present at this account, and which are missing? | The persona map in 6sense shows contacts already in the CRM against those that could be acquired, so the pass ends in a gap list |
| 1e Membership | Can one person fall under several personas? | Demandbase allows a person to be associated with multiple personas and up to ten active personas per buying group, so treat membership as a tag, not an identity |
Pass 1 output is a short list: which personas to look for at this account, and the persona note you will keep apart from any person's record.
The handoff rule
A persona-level statement may travel to a contact record only as a labeled persona note, in a field of its own, and never as a line in that person's profile. A persona note reads "Directors of operations in this segment ... " and states its source (your closed-won history, a vendor model, an interview round). It does not read "this person wants ...". If a sales message needs a persona note, write it as a question to the recipient, not as a claim about them.
Pass 2: the person pass (about this individual)
| Check | Question | What passes |
|---|---|---|
| 2a Identity | Is this a real person at this company now? | Name, employer and title carry a source and a date |
| 2b Freshness | When was the title and employer last verified? | A re-verification date inside your own limit, set from how fast your segment changes |
| 2c Reachability | Is the email or phone verified, and by whom? | A verification status with a date, from a check you can name |
| 2d Event | Is there a dated, public event tied to this person or role? | A hire, promotion, posting or announcement with a link |
| 2e Seat | What decision role does the person hold? | The person or the company has stated it; otherwise it stays a hypothesis |
| 2f Inference | Which lines are modeled or guessed, and from what? | Every such line says so and names its basis |
The four outcomes
| Outcome | Condition | Next step |
|---|---|---|
| Act | Person pass is clean, with at least one dated Recorded event | Write the message from the event |
| Verify first | Person pass has a stale or unverified identity or contact field | Re-verify before any message |
| Research | Persona fits, but no dated event or seat evidence | Add to the research queue; do not write |
| Stop | Identity is unverifiable, the person has objected, or the data does not belong in this workflow | Remove or suppress |
Fields: what each analysis line is, where it comes from, and how far to trust it
The table separates what a profile can hold by level. Confidence cells give the kind of evidence each line carries and do not state an accuracy rate. Vendor definitions are from the cited documentation.
| Field | Level | Source type | Freshness evidence | Verification | Confidence | Tier |
|---|---|---|---|---|---|---|
| Job title | Contact | Recorded string from a source such as a profile, a company page or a vendor record | Source date | A second source or a person-confirmed update | Depends on the date | Recorded |
| Employer | Contact | Same | Source date | Same | Depends on the date | Recorded |
| Job function and level | Persona | Classification of the title by a taxonomy or model (Demandbase translates non-English titles and does not show the translated title to users) | Re-run when the title changes | Compare the title with the class | Classification, not observation | Mapped |
| Persona label | Persona | Rule or filter set | Rule date | Review the rule | Follows from the rule | Mapped |
| Persona importance | Persona | 6sense weighted sum of profile fit, CRM contact count and pre-win engagement | Updated by the model | None in the contact record | Describes the seller's past | Modelled |
| Contact profile fit score (0–100) | Contact | 6sense model using background, data cleanliness and derived inputs; updated daily; needs the Predictive add-on | Daily | None in the contact record | Includes hygiene signals such as valid email and missing name | Modelled |
| Email status | Contact | Verification result | Check date | Named method | Per the vendor's status labels | Recorded |
| Hire, promotion or posting event | Contact | Public page or announcement | Event date | Link | High if the link resolves | Recorded |
| Buying role | Persona or contact | Configured by an admin or generated by a model (Demandbase calls its roles pre-defined; it separates them from personas) | Configuration date | The person or the company states it | Hypothesis | Mapped or Guessed |
| Priorities, objections, communication style | Persona | Persona research or a generated summary | Research date | An interview or call note | A statement about a type | Guessed unless the person said it |
Two cautions apply. A fit score is useful for choosing whom to research. It is not a fact about the person. And a hygiene input in a score means a clean-looking record can outrank an accurate-but-sparse one, so read the inputs before trusting the rank.
Step 1: run the persona pass at account level
Start with the account, not the contact. List the personas your own wins point to, and write down where that list comes from. Then compare it with who is in the CRM for the account. The 6sense persona map describes exactly that comparison: contacts tagged to a persona that are already in the CRM against those that could be acquired. Treat the missing seats as the research queue. The B2B buying committee mapping guide covers how to turn a seat into a role, because a persona is not a decision role.
Step 2: run the person pass on the shortlisted contacts
For each person on the shortlist, write the person-pass checks as a short card: identity with date, last verification, contact status, dated event, seat status and a list of inferred lines. If you build the card in a prospect dossier template, keep the confidence rungs defined there and use the tier names above for the persona fields.
Step 3: separate behavior-derived persona claims from facts
Visitor-resolution and intent tools produce persona-level claims from behavior. Intentsify describes persona-level resolution as inference from page views and content consumption, strengthened by matching against known persona profiles. That is useful for deciding which account to prioritize. It is not a record of who the visitor is. If the same session also resolves to a named contact, treat the name as an Identified match to verify, and keep the behavioral persona label as a Modelled line. The website visitor identification platforms comparison covers how match claims differ.
Step 4: set the limits on generated profiles
Generated executive briefs, such as the output of an AI persona-analysis agent, combine sourced facts with narrative. Draup's agent page lists sections from "Career & Leadership Profile" to "Engagement Strategy", and Draup's governance statement says its compensation insights are directional and that it distinguishes modeled from reported data. Apply the same split to any generated profile. Sort each sentence into Recorded, Mapped, Modelled or Guessed, delete what has no basis, and keep the rest labeled. A paragraph that cannot be sorted is a Guessed line.
Step 5: turn the read into a sales action
A profile read is only useful when it ends in an action. For the Act outcome:
- Choose the event. Pick the dated, Recorded line, such as a hire or a published post, that makes this message relevant now.
- Check the account side. Lead Compass is built for that stage: it reads your website, ICP and intent themes, surfaces opportunity picks, cites public sources where available and generates search prompts. It does not pull contacts until you run a search.
- Write one message that names the event and asks a question drawn from the persona note. Do not state the persona note as a fact about the recipient.
- Record the outcome in the profile, including objections, and update the tier of any line the reply confirms or contradicts.
- Feed the persona pass. A confirmed seat or an objection is evidence for the persona note, which is how persona-level statements become sourced.
For the data side, the best B2B contact database comparison covers where named people come from, and Percepture's guide to using intent data to identify sales qualified leads covers the step from a signal to a qualified conversation. How Lead Seeker works describes how a search returns people with the signal that surfaced them.
A worked read (invented)
Ardent Metering is invented for this guide, as are its accounts and contacts, and no result is claimed. It sells meter-data analytics to regional water utilities. The team has one account and a shortlist of three contacts.
Pass 1. The persona list comes from the team's own closed-won notes: operations director, finance director and a data lead. The CRM holds the operations director and the data lead for this account. The finance seat is missing, so it goes to the research queue. Importance scores exist from a vendor model, and the team notes that the model is built on past utilities deals and that the water-treatment segment has no history.
Pass 2 for the operations contact.
| Check | Result | Tier |
|---|---|---|
| Identity | Name, employer and title carry a source and a date | Recorded |
| Freshness | Title last verified inside the team's limit | Recorded |
| Reachability | Email verified, method and date logged | Recorded |
| Event | The utility published a tender for a metering upgrade, with a link | Recorded |
| Seat | No statement about the decision role | Hypothesis |
| Inference | A generated line says the contact "prioritizes cost control" | Guessed, deleted |
Outcome: Act. The message cites the tender and asks which part of the upgrade the contact owns. It does not mention cost control. For the data lead, the title was last verified long ago, so the outcome is Verify first. For the missing finance seat, the outcome is Research.
Where this will not work
- No history. A persona weight built from past deals has nothing to weigh for a new product, a new segment or a new region.
- Sparse records. A record with no recent source cannot pass the person pass. It belongs in the research queue, not in a sequence.
- Fit scores read as facts. A score that includes data-cleanliness inputs is a ranking aid. Do not quote it to the contact.
- One-person personas. In a small account, a persona label can name one person. The persona then identifies an individual, and the persona note becomes a statement about that person, so apply the person pass to it.
- Generated narrative. A fluent paragraph is not evidence. Unsourced lines stay Guessed.
- Privacy law. The GDPR defines profiling in Article 4(4) as automated processing of personal data to evaluate personal aspects, including work performance, interests and behavior. Where personal data was not obtained from the person, Article 14(3)(b) sets the notice no later than the first communication if the data will be used to contact them, and Article 21(2) and (3) give a right to object to direct marketing, including related profiling. Which rules apply to a given workflow is a question for counsel or a data protection officer.
Fact, inference and our view
| Statement | Type | Basis |
|---|---|---|
| A 6sense persona is a function-and-level combination; importance is a weighted sum of average profile fit, normalized CRM contact count and pre-win sales engagement | Fact (vendor documentation) | 6sense, Persona Importance Model |
| The contact profile fit model lists data cleanliness inputs and produces a 0–100 score with Strong, Moderate and Weak buckets, updated daily, with the Predictive add-on | Fact (vendor documentation) | 6sense, Contact Profile Fit Model |
| A person can be associated with multiple Demandbase personas; a buying group can have up to ten active personas | Fact (vendor documentation) | Demandbase, Create and Manage Personas for Buying Groups |
| Persona-level resolution infers likely role from page views and content consumption | Fact (vendor documentation) | Intentsify, website visitor resolution |
| A persona score built on your past wins gives no weight to a segment with no history | Inference | Follows from 6sense's listed inputs |
| A clean-looking record can outrank a sparse but accurate one | Inference | Follows from the data-cleanliness inputs |
| A persona note should live in a field apart from the contact's own lines | Our view | This guide's editorial recommendation |
| Message from a Recorded event, ask a question from the persona note | Our view | This guide's editorial recommendation |
Frequently Asked Questions
What is persona-level intelligence?
NMS Consulting defines it as information about a type of stakeholder that helps a team understand goals, priorities, pressures, buying role, decision criteria, objections, preferred evidence and likely behavior during a purchase. Other vendors use the word for a function-and-level classification or an inferred role of an anonymous visitor, so ask which meaning a page intends.
What is the difference between a persona and a contact profile?
A persona describes a type of stakeholder, such as a function and a level. A contact profile describes one person: name, employer, title, verification status and dated events. The persona can guide whom to look for, and the profile supports a message to the person.
Is a persona score a fact about the contact?
No. 6sense documents persona importance as a weighted sum of inputs from your own CRM and engagement history. The contact profile fit score is documented as a similarity measure to contacts from past opened or won opportunities, with data cleanliness inputs among its factors. Both are model outputs.
How do I use persona-level intelligence without guessing about individuals?
Keep persona statements in a labeled persona note, and write a message only from a Recorded, dated line on the person's profile. Use the persona note to choose a question, not to assert a motive.
Does persona-level analysis raise privacy questions?
It can. Building a profile that evaluates a person's work or interests is profiling as the GDPR defines it in Article 4(4), and notice and objection rules apply where personal data was not obtained from the person. Check the rules for each country and channel with counsel.
How to Read the Tables
The Two-Pass Contact Read separates what a persona can say about a type from what a profile can say about a person. In the field table, "Tier" uses the Recorded, Mapped, Modelled and Guessed scale from the persona-level intelligence contact summary guide, and "Confidence" describes the evidence behind the line, not a rate. "Fact" rows rest on a named vendor's documentation or on regulation text, "Inference" rows follow from them, and "Our view" rows are this guide's editorial recommendations. Ardent Metering and its accounts and contacts are invented. Vendor features, definitions and dates are an October 2026 snapshot.
Sources
- 6sense, Persona Importance Model (published August 14, 2025) and Contact Profile Fit Model (updated July 16, 2025).
- Demandbase, Create and Manage Personas for Buying Groups (updated March 23, 2026).
- Intentsify, Account-Level Intent vs. Persona-Level Intent (October 29, 2024) and Website Visitor Resolution: account vs persona vs contact (June 18, 2026).
- NMS Consulting, Ideal Customer Profile vs Buyer Persona.
- Draup, Persona Analysis AI Agent; Conversion Realtor, Persona-Level Intelligence: Contact Analysis for Real Estate.
- European Union, General Data Protection Regulation: Article 4, Definitions, Article 14 and Article 21.
- Lead Seeker, persona-level intelligence contact summary, B2B buying committee mapping, prospect dossier template, how Lead Seeker works, Lead Compass and best B2B contact database; Percepture, intent data to identify sales qualified leads.
About the Author
Bob Generale is President of Percepture. He works across SEO, AI search, digital PR, sales intelligence and AI-powered revenue systems, with a focus on connecting visibility, buyer intent and sales action.
Next Steps
Take one account from this week's list and run both passes: write the persona list and where it comes from, then a person-pass card for each shortlisted contact. Mark every line Recorded, Mapped, Modelled or Guessed, and write only to the contacts whose outcome is Act. To get verified people at the accounts you chose, talk to sales or start free with five contacts.