When we started thinking seriously about network quality at Drafted, the first instinct was to measure raw size. How many connections does each employee have? Run the numbers, rank employees by their totals, done. That framing turns out to be almost entirely wrong for hiring purposes.
The difference between depth and breadth in a professional network is not just a theoretical distinction. It determines whether a warm introduction actually carries weight with a candidate, or whether it gets a polite reply and nothing more.
What depth and breadth actually mean in this context
A broad network is one with a large number of connections where most of them are shallow: you met at a conference, you connected after a panel talk, you worked in overlapping companies but never directly. The connections exist and can be traced in a graph, but the social weight behind each one is low.
A deep network is defined by relational density. Fewer total connections, but each one involves actual shared history: worked on the same project, reported to the same manager, went through a recruiting loop together, collaborated closely for at least a year. When someone with a deep network sends an introduction message, the recipient reads it differently. There is enough shared context that the message does not read as a stranger's favor request.
This distinction matters enormously for warm recruiting because the value of a referral path is only as strong as the believability of the introduction. A 2nd-degree connection through someone who barely knows the candidate is not materially different from cold outreach. A direct introduction from someone who knows the candidate well is a fundamentally different ask.
The math of it: one strong connector versus ten weak ones
Consider two employees at the same company. Employee A has 2,200 LinkedIn connections accumulated over twelve years, most from conference attendance, short stints, and large company roles where direct collaboration was shallow. Employee B has 280 connections, all from roles where they worked in small teams and stayed three or more years per position.
For a specific open role, say a senior backend engineer with experience in distributed systems, Employee B's 280 connections probably yield more actionable warm paths than Employee A's 2,200. The question is not how many nodes exist in the graph. It is how many of those nodes would respond positively to a personal message from the connector.
When we model this in Drafted, we apply a warmth score to each potential introduction path that accounts for factors like shared employer overlap, connection recency, and mutual connection density. An employee who shares five mutual connections with a candidate, has a direct LinkedIn connection, and worked at the same company for three years scores much higher on a potential introduction than an employee who connected at an event two years ago with no overlap elsewhere.
Where breadth does matter
We are not saying breadth is valueless. In the right contexts, it is exactly what you want.
For roles where the candidate pool is geographic or community-based rather than purely credential-based, a well-connected networker who attends every industry meetup and knows most of the relevant community faces may be your best source. Local recruiting, community-specific roles, or positions where personality fit is paramount often benefit from the connectors who are just known by a lot of people.
Breadth also matters for diversity of sector exposure. If you are hiring for a role that requires someone with cross-industry experience, an employee with connections across many sectors may surface candidates that your deep-but-narrow connectors would never reach.
And for sourcing volume, breadth wins on sheer probability. If you need to open fifteen roles this quarter and speed matters more than match quality on any individual role, employees with large networks will surface more raw candidates faster, even if conversion rates are lower.
How Drafted weights the two in practice
The current Drafted model scores introduction paths on a composite signal, not a single metric. Connection count feeds in as a minor factor. What carries more weight are indicators of relational depth: co-employment history, LinkedIn mutual connections shared specifically between the potential connector and the candidate (not just anyone in the network), and connection recency.
For a role with specific technical requirements, we surface paths through employees whose professional history overlaps most directly with the candidate's background. A backend engineer who worked in distributed systems for three years at a prior job has more credible warmth toward candidates in that space than a recruiter who has connected with everyone at every relevant company.
For roles where cultural fit or communication style matters more than technical pedigree, we weight differently. There, broader networks and common community connections surface candidates that pure credential-matching would miss.
What this means for how you think about your team's network
The practical implication is that not all employees contribute equally to your referral pipeline, and the right employees to engage depend on the role. The engineering manager who has worked in the same subfield for ten years and maintained close contact with former teammates is your best source for technical roles in that domain. The partnerships lead who has worked at five companies across sales, recruiting, and operations is your best source for generalist and people-facing roles.
Asking every employee to think of referrals for every role is inefficient and often unproductive. It treats the network as undifferentiated. What actually works is matching the role requirement to the employees whose depth of connection puts them in the best position to make a convincing introduction for that specific candidate type.
That targeting is what Drafted tries to do automatically: for a given open role, rank introduction paths by the quality of the connection, not the quantity of names each employee can list. The connector matters as much as the candidate.
The boundary case worth keeping in mind
Depth and breadth are not always cleanly separable. Some employees have both: long tenures at multiple companies, strong ongoing relationships maintained over years, and a wide range of sectors covered. These are the connectors that every talent team should understand are unique assets for referral hiring. When Drafted surfaces one of these employees as the best introduction path for a role, it is usually a meaningful signal. The introduction has both reach and credibility.
The failure mode we see most often is treating network size as a proxy for network value. Large connection counts feel like coverage, but they often mask the fact that most of those connections carry no social weight at all for a recruiting introduction. Running introductions through low-quality paths trains candidates to ignore the channel entirely, which is exactly the opposite of what you want when the whole point is warmth.