> ## Documentation Index
> Fetch the complete documentation index at: https://support.metaview.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Referral Discovery

> Guide to advanced best practices using Sourcing.

Find potential candidates through the professional networks of people you have already hired, including former colleagues, classmates, and collaborators with a credible connection to someone on your team.

Referral Discovery is designed to identify warmer, referral-quality introduction paths rather than starting another broad cold search. It combines evidence of a plausible professional relationship with an assessment of likely fit for the target role.

## When to use it

Use Referral Discovery when you want to identify relevant candidates who may already have a meaningful connection to someone at your company.

It is especially useful when:

* you want warmer leads for a difficult or high-priority role;
* recent hires came from companies with strong relevant talent;
* employees have worked on relevant teams, products, or projects;
* cold outreach is producing low response rates;
* you want to identify people an employee may genuinely know;
* you need to expand beyond an existing target-company list;
* you want to understand which employees’ professional backgrounds are most relevant to a hiring plan.

For example:

> Find potential candidates for our Senior Product Manager role through the professional networks of Maria Chen and James Wu.

## What you get

The task produces two outputs.

**`potential_referrals.csv`** A structured list of potential candidates, including:

* the employee through whom they may be connected;
* evidence supporting the possible relationship;
* their current role and company;
* an assessment of fit for the target role;
* the apparent strength of the connection.

**Summary**

A concise overview showing:

* which employees were used as network anchors;
* how many relevant leads were identified;
* which anchors produced the most useful results;
* candidates identified through multiple employees;
* gaps in network coverage;
* recommended next steps.

The goal is not to generate the largest possible list. Twenty candidates with credible connection paths and relevant backgrounds are more useful than hundreds of weak overlaps.

## How it works

Start by replacing:

**Target Role**

with either:

* a job description;
* a link to the role;
* or a concise hiring brief.

Replace:

**Specific Employees**

with the names of people whose professional networks should be explored.

You can also replace:

**Additional Context**

with information such as:

* target geography;
* remote, hybrid, or onsite expectations;
* mandatory experience;
* preferred sectors or company stages;
* target or excluded companies;
* whether academic connections should be included;
* how strong the evidence of a relationship should be.

For example:

> Target role: Senior Product Manager for a B2B payments product in London. Specific employees: Maria Chen and James Wu. Additional context: Prior fintech experience is preferred but not mandatory. Candidates must be based in the UK or open to relocating. Prioritize former direct teammates and cross-functional collaborators.

The specified employees are used as **anchors**.

If no employees are named, the task can select a small number of relevant hires from the applicant tracking system based on similarity between their backgrounds and the target role.

Each anchor’s professional history is reviewed to identify:

* former employers;
* teams and products;
* schools and degree programs;
* locations;
* overlapping employment periods;
* public projects;
* publications;
* conferences;
* open-source work;
* other visible professional collaborations.

The task then searches the indexed profile database for people connected to those environments, filters them for role fit, and retains only candidates with an indexed profile and LinkedIn URL.

Results from all anchors are combined, deduplicated, assessed against the target role, and ranked for presentation.

## Understanding the anchor model

An anchor is an employee whose professional background is used as a starting point for candidate discovery.

The most useful anchors often:

* previously performed a similar role;
* worked at companies known for the target talent;
* belonged to a relevant product, regional, or functional team;
* attended a strong feeder institution;
* worked across several relevant companies;
* have visible public collaborations or project credits.

An employee does not need a large public network to be a useful anchor. A smaller, highly relevant professional background may produce better candidates than a broad but unrelated one.

When several anchors are used, each professional history is explored separately before the results are combined. This helps avoid over-relying on one employee’s background.

## What counts as a meaningful connection

A referral-quality connection should indicate that the candidate and anchor have a plausible reason to know one another professionally.

Strong evidence may include:

* working on the same team;
* contributing to the same product or project;
* overlapping in a small company or business unit;
* co-authoring a publication or company article;
* appearing together on a conference panel;
* collaborating on open-source work;
* sharing a specific academic program or cohort;
* holding closely connected roles during the same period.

A connection is weaker when it is based only on:

* employment at a very large company;
* attendance at the same large university;
* working in the same industry;
* being connected on a social platform;
* overlapping at different offices or unrelated business units.

The task should not present broad company or university overlap as proof that two people know one another. It should explain the available evidence and calibrate the connection-strength label accordingly.

## Understanding connection strength

Each lead is assigned one of three connection-strength levels.

**Strong**

There is clear public evidence that the anchor and candidate likely worked or collaborated directly.

Examples include:

* the same named team or product;
* co-authored work;
* shared project credits;
* a confirmed reporting or close cross-functional relationship;
* overlapping roles in a small organization.

**Likely**

There is good evidence of meaningful professional proximity, but direct collaboration cannot be confirmed.

Examples include:

* the same function and office during the same period;
* related leadership roles in the same business unit;
* the same small academic program or cohort;
* complementary roles on a visible initiative.

**Possible**

There is a plausible connection path, but the available evidence is limited.

Examples include:

* overlapping at a medium-sized company in the same function;
* attending the same institution during a similar period;
* appearing within the same broader professional community.

Possible connections may still be useful, but they should not be described as confirmed relationships.

Connection strength measures the evidence that two people may know each other. It does not indicate whether the employee would recommend the candidate.

## How candidates are assessed for role fit

Candidate discovery starts with the anchor’s professional background, but inclusion depends on likely relevance to the target role.

The task considers the person’s full career trajectory, including:

* current and previous titles;
* functional experience;
* seniority progression;
* sectors and business models;
* company stage;
* product or customer exposure;
* technical or domain experience;
* leadership scope;
* geographic fit;
* transferable experience.

A candidate should not be excluded only because their current title differs from the target title.

For example, a current Chief of Staff may still be relevant for a Product role if their earlier experience includes several years of product ownership. Conversely, a person with the exact target title may be a weaker match if their domain, seniority, or location is unsuitable.

Candidates currently employed by the hiring company are excluded.

## Understanding the CSV

Each row represents one potential referral candidate.

**Name** The candidate’s name as shown on their indexed profile.

**Current title and company** Their current professional position.

**LinkedIn/profile URL** The indexed profile used for candidate discovery.

**Connected through** The employee or employees whose professional histories led to the candidate.

**Connection notes** The evidence suggesting that the candidate and employee may know one another.

For example:

> Both worked on Stripe’s payments-infrastructure team between 2020 and 2022 and were credited on the same engineering article.

**Role-fit notes** A concise explanation of why the person may be relevant to the target role.

**Connection strength** Strong, Likely, or Possible, based on the available public evidence.

When a candidate is identified through multiple anchors, the connection evidence should be combined rather than creating duplicate rows.

## Candidates identified through multiple paths

Candidates connected to more than one anchor may be particularly useful.

Multiple paths can indicate:

* stronger overlap within a relevant professional community;
* more than one possible route to an introduction;
* independent employees who may know the person;
* a higher likelihood that someone internally has meaningful context.

For example:

> Connected through Maria Chen and James Wu. Maria worked with the candidate at Stripe; James attended the same graduate program and later appeared with them on an industry panel.

Multiple connection paths should increase a candidate’s referral priority when their role fit is also strong.

They should not override weak fit for the target role.

## How to improve the input

The most useful additional context helps distinguish a relevant connection from a relevant candidate.

### Describe the role beyond the title

Instead of:

> Senior Product Manager

use:

> Senior Product Manager for a B2B workflow product. Must have experience owning a technically complex product and working closely with enterprise customers.

This helps assess people whose titles are not an exact match.

### Specify geography

> Candidates must be based in London or within a practical commuting distance.

Without this, a productive anchor network may produce many otherwise relevant candidates who cannot realistically be hired.

### Identify mandatory and flexible criteria

> B2B SaaS experience is mandatory. HR technology experience is preferred but can be learned.

The search can then consider adjacent backgrounds without weakening the genuine requirements.

### Define preferred connection types

> Prioritize direct coworkers and collaborators. Include academic connections only where they shared the same small program or cohort.

This is useful when you want employees to feel comfortable making an introduction.

### Exclude sensitive networks

> Do not include anyone currently working at Maria’s previous employer.

This may help avoid conflicts, confidentiality concerns, or an overly concentrated target list.

### Add known team context

> Maria worked on Stripe Billing rather than the broader payments organization.

Specific team context can significantly improve the quality of the connection evidence.

## Useful ways to adapt the task

### Use one highly relevant anchor

> Find potential candidates for our Enterprise Account Executive role through the professional network of James Wu. Focus on people who worked with him at Snowflake and Datadog.

Use this when one employee has a particularly relevant background.

### Compare several employee networks

> Explore the professional networks of our five most recent Product hires for candidates suited to a Senior Product Manager role. Show which employees produce the most relevant leads.

Use this to identify the most productive internal referral channels.

### Focus on former teammates

> Include only people who appear to have worked on the same team, product, or regional organization as an anchor.

This produces fewer leads but stronger potential introduction paths.

### Include academic connections

> Include former classmates from small MBA, computer-science, and accelerator cohorts where the overlap is specific enough to suggest they may know one another.

This can work well for early-career, founder, investment, and specialist hiring.

### Consider public collaborators

> Include co-authors, open-source collaborators, conference copanelists, and people credited on shared projects.

This is useful for engineering, research, design, developer relations, and executive searches.

### Prioritize multiple connection paths

> Prioritize people connected to at least two employees, but include high-quality single-anchor candidates where the direct connection is strong.

This creates a narrower list with potentially warmer routes to introduction.

## How to interpret the results

The output should be treated as a list of possible introduction paths, not a list of confirmed personal relationships.

Before asking an employee to make an introduction, confirm:

* whether they genuinely know the candidate;
* how recently they interacted;
* whether they are comfortable making the introduction;
* whether they would recommend the person;
* whether the timing is appropriate.

A Strong connection label indicates strong public evidence of professional overlap. It does not mean the employee has agreed to refer the person or would endorse them.

Pay particular attention to:

* the quality of the connection evidence;
* the candidate’s fit independently of the connection;
* whether the relationship is recent enough to be useful;
* whether many leads come from the same narrow team;
* which employees produced the most relevant candidates;
* where the selected anchors provide limited coverage.

## Common pitfalls

**Treating every former coworker as a referral**

People who worked at the same large company may never have met. Look for team, product, office, function, or project evidence.

**Optimizing for quantity**

A large list of weak overlaps creates work for recruiters and may reduce employee confidence in the process.

**Ignoring role fit**

A strong professional connection does not make someone suitable for the role. Connection strength and candidate fit should be assessed separately.

**Looking only at current titles**

Relevant candidates may have transferable experience earlier in their careers or may currently hold a broader title.

**Assuming an employee will make the introduction**

The map identifies potential connection paths. Employees should confirm the relationship and decide whether they are comfortable making an introduction.

**Using overly broad academic overlap**

Attending the same large university is usually weak evidence unless the individuals shared a small program, cohort, society, laboratory, or project.

**Failing to merge duplicate candidates**

A person identified through several anchors should appear once, with all relevant connection paths preserved.

**Over-concentrating on one former employer**

A productive company network can quickly dominate the results. Review whether other employers, schools, and professional collaborations provide useful additional coverage.

## Privacy and responsible use

Use only professional information available through authorized sources and public professional profiles.

Do not treat inferred connections as confirmed relationships. Employees should verify whether they know a person before any referral or introduction is requested.

Referral decisions should remain voluntary. Employees should not be pressured to contact, recommend, or share private information about anyone identified through the task.

Apply the same job-related criteria consistently to all potential candidates. Do not use inferred personal characteristics, protected characteristics, or non-professional information when assessing role fit.

## Recommended workflow

1. Define the target role, geography, and genuine must-have requirements.
2. Select employees whose professional backgrounds are relevant to the role.
3. Generate the initial referral map.
4. Review the connection evidence before sharing candidates internally.
5. Prioritize candidates who combine strong role fit with Strong or Likely connection paths.
6. Highlight candidates connected through multiple employees.
7. Ask each anchor privately which people they genuinely know and would feel comfortable contacting.
8. Record whether the employee recommends, knows, or only recognizes each person.
9. Use warm introductions for the strongest confirmed relationships and thoughtful direct outreach where the connection cannot be activated.
10. Review which anchor networks produced the most relevant candidates and use those findings when selecting anchors for future roles.

The strongest referral-discovery searches treat employees as starting points for identifying relevant professional connections, not as automatic endorsers. The objective is to uncover credible potential introduction paths, then allow employees to confirm which relationships are genuinely appropriate for outreach.