Building a graduate talent pipeline with a business school

Building a graduate talent pipeline with a business school

A graduate talent pipeline is a group of early-career people you know before you have a vacancy, and building one with a business school means joining a cohort that already exists rather than assembling one yourself.

That distinction decides most of what follows, because the standard pipeline playbook does not survive contact with early careers. It was written for experienced hires and it assumes a candidate who has already left a professional trace. A second-year student has not left one.

This is written for the employer who recruits. It is not about the national skills agenda, which uses the same word for something else, and it is not about how a school recruits its own students.

What a talent pipeline means, and where the definition breaks

A talent pipeline is a group of people an employer identifies and stays in contact with before a role opens, so that hiring starts from a known group rather than an open advert.

The industry definition is consistent and easy to find. A pipeline is a pool of passive candidates: sourced from professional networks, tagged in a recruitment CRM, segmented by role family, and kept warm with employer brand content until a requisition appears. Silver medallists from previous rounds go into the same pool. The mechanics are well documented and, for a senior hire, they work.

Now read that against someone who is 20.

• There is no professional history to source against.
• They are not in your applicant tracking system, because they have never applied to anything.
• There is no silver medallist to re-contact, because there was no previous round.
• There is nothing to target on a professional network beyond a course title and a graduation year.

Sourcing, tagging and nurturing all presuppose a person who has already done something you can find. Early-career candidates have not done it yet. So the pool does not pre-exist your outreach, and no amount of software conjures one. It exists in one place, which is wherever the cohort already is.

Most of what is published on this question is published by organisations that sell access to candidates, or software for storing them. That is not a reason to dismiss any of it, but it is a reason to ask who published a framework before you adopt it.

The horizon problem

A graduate pipeline is a two to four year object. You meet someone in their first or second year, they spend months with you on a placement, they finish their degree, and you hire them at the end of it. None of that compresses, because it is governed by a degree rather than by your hiring plan.

Against that, planning horizons in early-career recruiting are short. Surveyed in February and March 2026, only 33.8% of American employers had settled their recruiting plans for the following autumn and 60.4% had not (National Association of Colleges and Employers, Job Outlook 2026 Spring Update, 154 respondents). Commitments made further out are rarer still: employers report that on average just 10% of their internship offers are made more than twelve months before the programme starts (NACE, 2024 Recruiting Compensation Report).

The instrument is shorter than the object. An organisation that has not yet settled next autumn cannot build a thing that pays out in 2029, and when it fails it will conclude that pipelines do not work, when what failed was the horizon.

The direction of travel runs against the planning data, which makes it worth reading closely. The long route is the one that is working. The share of interns converted into full-time hires reached 63.1% for the 2024-25 intern class, the highest in five years and up nearly 13% in a single year, with 88.3% of those offers accepted, and employers expected to take on about 4% more interns the following year (NACE, 2026 Internship & Co-op Report, 284 organisations). Employers are lengthening their commitments faster than they are lengthening their plans.

The benchmarks exist, and most pipelines are not measured against them

The measurement picture is the mirror image of the planning one.

The comparators are published and free to the organisations that help produce them. American employers who measure report an average of 27 days between a student's first interview and an offer, 9 days for the student to answer, and a median early-careers recruiting budget of about $114,000 a year (NACE, 2025 Recruiting Benchmarks Report and 2024 Recruiting Compensation Report). Every one of those is a number you can compute from your own records in an afternoon, and a pipeline that has never been compared against them is being run on impressions.

The conversion rate is the one that decides everything else, because it is the only figure that says whether the object you are funding produces anybody. It is also the one with the clearest published comparator, and the comparator moves: 63.1% for the 2024-25 intern class, against a rate below 51% two years earlier (NACE). Read against that, a pipeline reporting no conversion figure at all is not a pipeline with a bad number. It is a pipeline with no number.

The object that takes longest to pay back is therefore the one least likely to be measured, which is also why it is the one most likely to be cancelled. A commitment nobody has evidence for does not survive the budget round that follows the sponsor's departure.

What the failure looks like: year one, and year five

Retention sorts early-career hires by how much of your organisation they saw before they joined. More than three quarters of new hires who had interned with the employer are still there after one year, and more than half after five. Among hires who interned somewhere else, nearly two thirds remain after a year and nearly half after five. Among hires with no internship experience at all, fewer than half are there after a year and only around a third after five (NACE, 2026 Internship & Co-op Report).

Those are three readings of one problem. A graduate offer is typically signed months before the start date, and in most organisations nothing happens in between. The candidate finishes a degree, collects three other offers, and reappraises a decision made in a different life. The failure sits at the seam between the offer and the job, which is the stretch of the process nobody owns.

A route built on months of real work behaves differently at that seam. Someone who has spent six months in your organisation already knows what a Monday there looks like, and that is a different decision from the one your other offer-holders are making in August.

What a placement reveals that a process cannot

What separates those three groups is not talent, it is exposure. Asked how they choose between candidates for an internship in the first place, employers name three factors above all others: the candidate holds the key skills the position requires, the candidate interned with the company previously, and the candidate interned in the industry previously. Major and grade point average sit below all three (NACE, 2025 Internship & Co-op Report, 247 organisations).

That has a blunt implication for selection. What you most need to know about a 22-year-old is how they behave when a brief is ambiguous and nobody is watching. A CV cannot carry it, and a generated application actively obscures it. It becomes visible over months, which is what a placement is: the longest work sample available to you.

The selection literature has been corrected in a direction that favours observation. Sackett, Zhang, Berry and Lievens showed in 2022 that the standard corrections for range restriction had systematically overstated validity across the field, re-estimating cognitive ability tests at .31 and leaving the structured interview at .42 (Journal of Applied Psychology). Structure and evidence outperform instruments, which is an argument for evidence gathered over months rather than over an afternoon.

The conversion this supports is measured every year. On average, 63.1% of the 2024-25 intern class were converted into full-time hires, the highest rate in five years, with an offer acceptance rate of 88.3% (NACE, 2026 Internship & Co-op Report). How the placement is run moves the number materially: for the 2023-24 class, employers running in-person internships converted 58.5% of their interns against 46.0% for those running hybrid ones (NACE, 2025 Internship & Co-op Report). Time in the building is doing the work.

What a business school actually contributes

An institution holds four things a supplier does not, and they happen to be the four a pipeline needs.

The cohort exists before anyone applies anywhere. The calendar is real and published: placement windows, apprenticeship rhythms and the point at which students actually decide, none of which move to suit a requisition date. Selection has already happened once, at admission. And the exposure is long enough to produce evidence. More than 9,000 SKEMA students take up an internship or apprenticeship each year, for periods of two to 24 months, in France or abroad, and companies reach students on placement through that route.

None of this is preferential treatment, and it is worth being direct about that, because employers are moving the other way. Seventy per cent now use skills-based hiring, up from 65% a year earlier, and 71% of them use it at least half the time (NACE, Job Outlook 2026 Spring Update). So if you select on demonstrated skills rather than on the institution, does a relationship with an institution still mean anything?

It does, because it works somewhere else in the process. A relationship decides who is in the room, months or years before a sift happens. Who gets picked is decided later, by you, on your criteria. Access and time are what the relationship buys, and the judgement stays yours.

The events are entry points rather than the pipeline itself. SKEMA runs Career Fairs and Sector Days across its campuses, the latter built around a single industry, and what a corporate partnership involves is set out in four levels, from recruitment through to research. The analysis published for corporate partners covers what early-career candidates say they want, which is the other half of the same question.

How to enter a pipeline you did not build

The academic calendar sets the entry point, not your vacancy. Placement recruitment runs to fixed windows, and an employer who arrives when the requisition opens arrives after the cohort has committed. Working backwards from graduation to first contact is the planning exercise, and it is longer than most resourcing plans allow for.

Then judge it over the right period. Two academic cycles is the minimum before a partnership tells you anything, because the first cohort is the one you learn on. A pipeline assessed on a single semester has been assessed on its set-up costs.

Measure the sequence rather than the crowd: conversion from placement or apprenticeship into a permanent role, retention at 12 and 24 months, the share of your early-career hires that came through the relationship, and the time from first contact to start. Set the baseline in the first cycle. There is a published comparator for the first of those, currently 63.1%, and none at all for a partnership that has never produced the figure.

To take part, send your participation preferences through the recruitment events form, or contact the corporate relations team to work out which entry point fits the roles you are filling, and the years you are filling them in.