A sourcing funnel is the measurable path from a search result to a meeting a candidate actually attends. It connects candidate discovery, contactability, outreach, replies, interest, booking, and attendance so a hiring team can see both volume and loss at every handoff. A complete funnel has seven stages: search results, contactable people, people contacted, replies, positive replies, meetings booked, and meetings attended.

The point is not to celebrate a large list of profiles. The point is to know whether the available market, chosen channels, message, and operating process can create enough worthwhile conversations for one specific hiring need.

What does a sourcing funnel measure?

A sourcing funnel measures how prospective candidates move through active sourcing. A search result is a profile that appears potentially relevant against a defined role brief. It is an input to review, not evidence that the person should be contacted.

A person is contactable only when the team has a usable and appropriate route to reach them. That might be a professional email address, an accessible professional-network message, a referral, or an existing relationship in a talent pool. A publicly visible profile on its own does not make someone contactable.

Contacted should mean a documented message that was delivered through that route. A reply is any substantive response, including a decline. A positive reply signals willingness to explore a next step; it is not a commitment to change jobs. Booking and attendance are separate events because a calendar invitation is not a completed conversation.

Set those event definitions before the campaign begins. Otherwise, two teams may report the same “reply rate” while using different denominators. The guide to AI active sourcing and people search explains why discovery is only the first layer of that workflow.

Where candidates are lost in the sourcing funnel

No single conversion profile applies to every search. Seniority, geography, employer recognition, compensation, work pattern, role scarcity, timing, and the channel all change the result. That makes a transparent operating range more useful than treating one external average as a promise for every vacancy.

The planning model below is current as of 20 August 2026. Its figures are deliberately not presented as an industry benchmark. They are a working scenario that a team should replace with its own consistently measured campaign data:

• Search result to contactable: 55 to 75 percent. Profiles may be duplicated, stale, unsuitable after review, or lack an appropriate route for outreach.
• Contactable to contacted: 80 to 95 percent. A valid route does not mean immediate outreach is always appropriate; priority rules, suppression rules, and approvals create a legitimate gap.
• Contacted to reply: 5 to 20 percent. This range is highly sensitive to the role, channel, timing, reason for contacting the person, quality of message, and sequence design.
• Reply to positive reply: 25 to 45 percent. A reply can be a refusal, objection, question, or expression of interest, so total replies should never be treated as positive interest.
• Positive reply to booked meeting: 50 to 70 percent. Speed, a clear next step, and usable availability matter most at this point.
• Booked meeting to attended meeting: 75 to 90 percent. Clear expectations, easy rescheduling, and considerate reminders are operational levers here.

Top-of-funnel loss is often the largest in absolute terms. From 1,000 search results, a 65 percent contactability rate leaves 650 people before a single message has gone out. This loss is easy to overlook because many outreach dashboards only start counting at send. It is still essential to capacity planning.

Why contactability is the hidden constraint

Contactability is a hidden constraint because it sits between an attractive profile and a credible opportunity to start a conversation. Teams often debate subject lines and follow-ups while a substantial share of researched people have no suitable way to be reached. Guessing an email address does not create reach; it can create failed contact, poor data, and misleading reporting.

The first lever is therefore upstream of messaging: document the search logic, remove duplicates, verify available routes, and choose a plausible first channel for each person. When evaluating AI sourcing tools, database size is not enough. A team should be able to understand a person’s source, data recency, viable channel, and current workflow status.

Contactability is also not consent, relevance, or eligibility. It describes only whether an appropriate contact attempt can be made in practical terms. Privacy, fairness, message relevance, and the right to stop contact need their own controls.

How should each sourcing stage be measured?

A funnel metric is a numerator divided by an unambiguous denominator. Record the channel, source, time window, message version, and reason for exclusion as well as the stage itself. That lets a team distinguish a search-coverage issue from a data problem, an offer problem, or a communication problem.

• Contactability rate: contactable people divided by reviewed search results.
• Outreach rate: people contacted divided by contactable people.
• Reply rate: replies divided by delivered messages, not by all messages planned for a sequence.
• Positive-reply rate: positive replies divided by replies. Also view positive replies against delivered messages so a small response pool cannot look better than it is.
• Booking rate: meetings booked divided by positive replies.
• Show-up rate: meetings attended divided by meetings booked.

A realistic target range is an agreement about the team’s own operating conditions, not a borrowed number. Begin with a transparent planning scenario, separate comparable roles and channels, then replace assumptions after a complete campaign window with observed values. Do not compare a warm referral with a cold email, or a regional sales search with a senior technical search, as if they were the same experiment.

Channel choice multiplies more than reply rate

Channel selection is a multiplier because it affects more than replies. It changes whether a profile is operationally contactable, whether the approach arrives in a credible context, and whether a positive exchange can become a meeting without unnecessary friction. The same person may be hard to reach by professional email and highly reachable through a trusted introduction.

Start with the most plausible route for the individual, not with the channel that can send the most messages. Professional email, professional networks, specialist communities, referrals, and a known talent-pool relationship each have different expectations and tracking limitations. Multiple channels only work when one shared history prevents a candidate receiving disconnected parallel approaches.

After a positive exchange, a structured next step may be useful to gather information that is not visible in a profile. A scheduled voice interview can support that handoff. It does not replace sound outreach or human judgement; it helps ensure that a well-earned candidate conversation has a clear continuation.

Match automation to the stage it can improve

Automation works best when it is broken into controllable steps along the funnel. At the search-result stage, it can sort profiles against agreed criteria, surface duplicates, and assemble research context. It should not convert incomplete evidence into a supposedly certain judgement of fit.

At the contactability stage, automation can consolidate known contact routes, flag missing data, and prepare prioritised worklists. In outreach, it can prepare drafts, schedule messages, and coordinate follow-ups. In the reply stage, it can route routine questions, expose response deadlines, and suggest meeting times. Ambiguous, sensitive, or negative replies need an explicit human escalation route.

At booking and attendance, automation is mostly logistical: checking calendars, confirming details, allowing a reschedule, and writing status back to the shared record. It should never leave a candidate trapped in an automated loop without a clear human owner.

Sprad People Search is designed for this connected path: from search through selectable levels of automated outreach, follow-ups, and initial questions to a booked calendar meeting. Those stages still need conscious approval and oversight. Automation does not decide whether a role is compelling, whether a person is a fit, or how a human conversation should be assessed.

Reverse the funnel to calculate search volume

The most useful funnel calculation starts with the outcome required, not the number of profiles already found. This funnel ends at the attended first meeting, so a hiring plan also needs its own downstream conversion: for this role, how many attended first conversations does the team historically need for one accepted hire?

Assume, explicitly as a planning input rather than a benchmark, that a role needs ten attended first conversations for one hire. With 65 percent contactability, 85 percent outreach, 12 percent reply, 35 percent positive replies, 60 percent booking, and 80 percent attendance, one search result produces 1.11 percent of an attended meeting in the model. The team needs roughly 900 search results for ten attended meetings; 1,000 search results produce about eleven attended meetings.

The information gain is not the number 900. It is the decision rule behind it: rank improvements by the additional attended conversations they could create, not by the amount of activity they add. In this example, lifting contactability from 65 to 75 percent adds roughly 1.7 attended meetings per 1,000 results. Lifting reply rate from 12 to 17 percent adds roughly 4.6, if every other assumption remains constant. Testing determines which lift is attainable; the calculation prevents teams from choosing the loudest lever rather than the most valuable one.

A good person who is not right for the vacancy today is not necessarily a lost result. With the right process and permission, they can move into a maintained candidate portal and talent pool rather than disappearing after one declined opportunity.

What a sourcing funnel cannot tell you

A funnel measures process progression, not quality by itself. A higher reply rate can come from outreach that is too broad. A higher booking rate can come from making the next step too low-commitment. Neither is a success if the conversations are not relevant to the role or do not support fair selection.

Small datasets can also turn random variation into an apparent trend. Do not treat a senior engineering search in one city as directly comparable with a local sales search, and do not combine referrals and cold outreach into one performance figure. A funnel provides orientation, not certainty about a hire.

The important open limitation is this: no level of automation and no reverse calculation can compensate for an unclear role, an uncompetitive offer, or slow decision-making. The funnel can expose those issues; it cannot solve them on its own.

FAQ: Sourcing funnel

What is the difference between a search result and a contactable candidate?

A search result only appears relevant within the search space. A candidate becomes contactable when there is a credible, appropriate route to reach them. Separating the two stops list length from being mistaken for practical reach.

What is a good reply rate in active sourcing?

There is no good rate that applies to every role. Assess reply performance by channel, role, location, and message type. It becomes meaningful only when delivered messages are the denominator and positive responses are tracked separately.

How can a team improve contactability?

Measure search source, data freshness, duplicate records, and the suitable contact route separately. Remove unverified contact attempts from the active sequence and record why a profile was not worked. This turns an invisible loss into an actionable metric.

How many follow-ups should be part of the funnel?

The suitable number depends on the channel, reason for contact, and acceptable frequency. Each follow-up should have a clear purpose, respond to what happened earlier, and be easy to stop. Track follow-ups as part of one candidate sequence, not as newly sourced people.

Can a sourcing funnel predict a hire?

No. It creates a transparent capacity and learning model, not a hiring prediction. Add your own post-meeting conversion data and qualitative review of the role, offer, and candidate experience before using it for planning.