The ROI of Hiring Interns: What the Data Says

Most Indian companies do not have a hiring problem at the senior end. They have a hiring problem at the bottom of the pyramid — the analyst, associate, junior developer and executive roles that make up the majority of headcount growth in a scaling business, and where the economics of recruitment are least forgiving.

The pattern is familiar. A team needs a junior person. The requisition is raised, a consultant is engaged at a fee that is typically a meaningful percentage of first-year CTC, and six to ten weeks later a candidate joins with two years of loosely relevant experience, a salary premium justified by the market rather than the role, and a ninety-day notice period served at a previous employer that has already eroded the urgency the hire was meant to solve. Some of these hires work out. A material proportion do not, and the cost of that failure — recruitment fee, salary paid, manager time, and the vacancy re-opened — is rarely reconciled against the original business case.

The internship route is the obvious alternative, and it is also the one most often dismissed by senior leaders as an administrative burden with uncertain returns. That scepticism deserves a data-driven answer rather than an enthusiastic one. This article sets out what the available evidence actually supports about intern cost, conversion, retention and productivity — and, equally important, where the evidence is thinner than internship advocates typically admit.

What an intern actually costs versus a lateral hire

Begin with the direct outlay. In India, the stipend benchmark has been anchored in part by public policy: the Prime Minister's Internship Scheme, launched as a pilot in October 2024, pays a monthly stipend of ₹5,000, of which ₹4,500 comes from the government via direct benefit transfer and ₹500 from the participating company's CSR funds, plus a one-time joining grant of ₹6,000. Market-rate internships in technology, analytics and finance run well above that figure, but the scheme establishes a floor and a reference point that most employers now negotiate around.

The comparison that matters, however, is not stipend versus salary. It is total acquisition cost per confirmed, retained hire. That calculation includes sourcing spend, screening and interview hours, agency fees, onboarding, the productivity loss of a role sitting vacant, and the probability-weighted cost of a mis-hire.

International benchmarks give a sense of scale. The Society for Human Resource Management has reported an average cost-per-hire of approximately USD 4,129 and an average time-to-fill of 42 days. Indian absolute figures are lower, but the structural composition — agency fees and vacancy cost dominating — is broadly similar.

The table below is a structural comparison rather than a claim about any particular market rate. Populate the cells with your own numbers; the point is which line items exist in each column.

Dimension	Intern-to-hire route	Lateral junior hire
Third-party fee	Usually nil; campus or platform sourcing cost only	Agency fee, typically a percentage of first-year CTC
Pre-offer evaluation	8–24 weeks of observed work output	2–4 hours of interviews plus references
Information quality at decision point	Direct evidence of performance in your context	Inferred from CV, interview and prior employer signal
Time from decision to productive	Near-zero; already ramped and contextualised	Notice period plus onboarding ramp
Downside if it does not work	Programme ends; no separation process	Fee sunk, salary paid, exit process, requisition re-opened
Compensation anchoring	Set at entry band	Anchored to candidate's prior CTC plus market premium
Principal cost	Manager and mentor time	Cash, and the cost of vacancy

The critical asymmetry is in row three. An interview is a sample of behaviour under artificial conditions lasting a few hours. An internship is a sample of behaviour under real conditions lasting weeks. No structured interview process yet devised gives an employer the predictive information that observing someone deliver actual work does. That is the economic argument for internships in a single sentence — not that interns are cheap labour, but that an internship is the least expensive high-fidelity assessment available.

The offsetting cost is real and frequently underestimated: senior time. A well-run internship consumes meaningful mentor and manager hours. Employers who model intern ROI without pricing that in will overstate the return, and employers who refuse to spend it will get the outcome they have paid for.

The conversion advantage: interns who become employees

The most robust longitudinal data on intern conversion comes from the National Association of Colleges and Employers in the United States, which has surveyed employer internship programmes annually for over two decades. Its 2026 Internship & Co-op Survey found the intern-to-full-time conversion rate reached 63.1% for the 2024-25 intern class — a five-year high, with offer and acceptance rates rising in parallel.

NACE reports an intern-to-full-time conversion rate of 63.1% for the 2024-25 intern class, a five-year high — meaning roughly six in ten interns at surveyed employers were converted into full-time hires.

That figure has moved considerably over the cycle. NACE's 2024 report put conversion materially lower. Conversion is therefore cyclical and demand-sensitive; it is not a fixed law of nature, and any employer building a business case on the highest reported number is building on the top of a cycle.

Two further findings from NACE are more actionable than the headline rate. First, acceptance rates are high and rising — 88.3% among 2024-25 interns — which means that when an intern-employer relationship works, the offer usually closes. Second, and most striking, NACE's work on returning versus non-returning interns found that returning interns receive and accept offers at dramatically higher rates than those interning for the first time.

The implication for programme design is direct: a second internship with the same employer is among the strongest predictors of conversion in the dataset. Employers running one-and-done summer programmes are discarding their highest-yield cohort.

A necessary caveat: this is US data. There is no equivalent longitudinal Indian conversion benchmark of comparable rigour, and Indian campus dynamics — placement cell mediation, pre-placement offer conventions, six-month final-year projects — differ materially. The direction of the finding is likely to travel; the precise percentage should not be assumed to.

Retention — why intern-converted hires stay longer

Conversion is only half the argument. A hire who leaves in fourteen months has destroyed most of the value the internship created.

NACE's retention analysis tracked hires five years out and found that employees who had interned at the hiring organisation itself showed materially higher five-year retention than those who had interned elsewhere, and higher still than hires with no internship experience at all.

Five years after joining, NACE found retention was highest among hires who had interned at the same organisation — ahead of those who interned elsewhere, and well ahead of those who had never interned.

A gap of a few percentage points may look modest. Applied across a junior cohort of a hundred over five years it is several additional retained employees, each carrying five years of accumulated firm-specific knowledge that would otherwise have to be rebuilt from scratch.

The mechanism is not mysterious, and it is worth naming because it explains when the effect will not appear. Retention improves because both sides had accurate information before committing. The intern learnt what the work, the manager and the culture are actually like; the employer learnt whether this person can do the job. Attrition in the first two years is very often a mutual information failure, and the internship is the instrument that removes it. Where an internship is unrepresentative of the real job — where interns are given sanitised work and a curated experience — the information advantage disappears and so, predictably, does the retention benefit.

The Indian context makes this valuable. Voluntary attrition at India's largest IT services firms has moderated considerably from its post-pandemic peak but still means that a substantial share of a junior cohort turns over within a few years, per ICRA's analysis of the sector.

The productivity ramp: how quickly interns become net contributors

The honest position here is that there is no credible, generalisable benchmark for when an intern crosses from net cost to net contributor. Any article quoting a precise week number is inventing it. What can be said is structural.

An intern's ramp has three phases: context acquisition, supervised delivery, and independent delivery. The variable that determines the length of phase one is not the intern's ability — it is the quality of the employer's onboarding. Teams with documented systems, a defined first task and an assigned mentor compress it dramatically. Teams that hand an intern a laptop and a vague brief may never exit it.

The more important productivity argument is about the converted hire, not the intern. When an intern converts, the ramp has already been paid for. The lateral hire's first three months are spent learning the codebase, the clients, the internal politics and the tooling; the converted intern's first three months are spent producing. Against a lateral hire who additionally serves a notice period before joining, the effective delta between the two on a calendar basis is substantial.

What separates a programme that works from one that wastes everyone's time

The variance between internship programmes is far wider than the variance between individual interns. Programmes that consistently produce hires share a small number of characteristics.

Real work with real consequences. If the output would not have been produced anyway, it is not work — it is occupational therapy, and it yields no assessment signal.
A named owner. Not the HR function; a line manager whose performance objectives include the intern's output.
Duration sufficient for signal. Very short placements measure enthusiasm. Longer placements measure capability under fatigue, which is the thing you actually need to know.
Explicit conversion criteria communicated at the start. The intern should know on day one what performance would earn an offer.
Structured feedback at defined intervals, documented, so the conversion decision rests on a record rather than a recollection.
A decision date. Programmes without one drift, and the best candidates accept other offers while the employer deliberates.

Structuring an internship that produces hires

Define the target role first. Work backwards from the junior vacancy you expect in nine months, and design the internship as an audition for that specific role.
Scope one deliverable per intern that has a business owner. A named internal customer who wants the output creates accountability that no HR process can replicate.
Set a mentor ratio you can honour. One mentor to two or three interns is workable; one to eight is a programme in name only.
Assess against a rubric, not an impression. Score output quality, learning velocity, communication and reliability separately. Learning velocity is the strongest predictor of trajectory in early-career hires and the one most often overwritten by likeability.
Run a mid-point review. It gives the intern time to correct and gives you a second data point rather than a single end-of-term impression.
Make the offer early and make it clean. NACE's data on acceptance rates suggests that offers extended to interns close at very high rates; the risk is not rejection, it is delay.
Build a returning-intern track. Given the elevated offer and acceptance rates NACE reports for returning interns, inviting strong first-year interns back is among the highest-yield interventions available.

Common mistakes employers make

Treating interns as a cost-reduction device. An internship optimised for cheap output rather than assessment produces neither. The return is in hiring quality, not in labour arbitrage.

Recruiting for the internship rather than for the role. If your intern selection bar is meaningfully below your graduate hiring bar, conversion will disappoint and the programme will be judged a failure on grounds that were determined at intake.

Running the programme out of HR alone. Central coordination is necessary; ownership must sit with the line.

No conversion budget. Programmes that produce excellent candidates and then have no approved headcount to absorb them destroy employer brand on campus with remarkable efficiency, and campus reputation is slow to rebuild.

Ignoring the unconverted majority. Even at a 63% conversion rate, a third of interns leave without an offer. They become alumni, referrers and future lateral candidates. How they are exited is a hiring-brand decision.

The India-specific context

Three structural forces make the internship route more attractive in India than in most markets.

Supply. The All India Survey on Higher Education, conducted by the Ministry of Education, records total higher-education enrolment in the tens of millions with roughly a crore of students passing out annually. No other market offers a comparable early-career funnel.

Screening cost. That same scale is the problem. The India Skills Report 2026, produced by Wheebox with ETS, CII, AICTE and AIU, places overall graduate employability at 56.35%, up from 54.81% the previous year. Improving, but it still means that credentials alone are a weak filter and that some form of demonstrated-work assessment is close to mandatory.

Policy. NEP 2020 embeds internship and hands-on vocational exposure across the education system. The AICTE Internship Policy requires technical institutions to embed internships within the credit structure of approved programmes, which means institutions now have a compliance obligation to place students. Meanwhile the PM Internship Scheme has established both a stipend norm and a national expectation that large employers host structured cohorts.

The practical consequence for an employer is that colleges are now actively seeking industry partners rather than merely tolerating them. That reverses the traditional negotiating position on campus, particularly for mid-market companies that cannot compete with large employers on day-one placement slots but can offer genuinely substantive project work.

How to measure your own intern ROI

Do not benchmark against published conversion rates. Benchmark against your own alternative — what the same hire would have cost through your normal lateral channel.

A workable formula:

Cost per converted hire = (total programme cost) ÷ (number of interns converted and retained at 12 months)

Total programme cost should include stipends for the whole cohort — not only those converted — plus sourcing and campus costs, equipment and licences, and a fully loaded estimate of mentor and manager hours. That last item is the one most often omitted and is frequently the largest.

Compare that figure against your lateral cost per retained hire: agency or sourcing spend, internal recruiter time, interviewer hours, onboarding, the cost of the vacancy over the notice period, and the mis-hire cost weighted by your actual first-year attrition rate for junior lateral hires.

Then track three ratios each cycle:

Offer rate — interns offered ÷ interns who completed. Low here means your intake bar is wrong.
Conversion rate — interns joined ÷ interns offered. Low here means your offer terms, timing or candidate experience are wrong.
12- and 24-month retention of converted hires versus lateral junior hires. This is the number that settles the argument internally, and it is the one almost nobody tracks.

Two cycles of that data will tell you more about your programme than any external benchmark, and it converts an internship from a goodwill activity into a line item with a defensible return. For employers starting from scratch, a dedicated internship platform is the most straightforward way to reach candidates beyond the handful of campuses you already visit.

Frequently Asked Questions

How long should an internship be to give a reliable hiring signal?

Long enough for the candidate to complete a full delivery cycle independently — typically at least eight to twelve weeks. Shorter placements measure attitude and aptitude but rarely capture how someone performs once novelty has worn off, which is the behaviour that predicts the first year of employment.

Should we pay interns, and how much?

Pay. The PM Internship Scheme's ₹5,000 monthly stipend plus ₹6,000 joining grant establishes a public reference point, and market rates in technology and finance sit well above it. Unpaid internships restrict your pool to candidates who can afford to work for nothing, which is a selection bias against exactly the ambitious, resource-constrained candidates who tend to perform well.

What conversion rate should we target?

NACE's 63.1% for the 2024-25 class is a US benchmark at a cyclical high and should be treated as context, not target. A more useful internal test: is your cost per converted-and-retained hire lower than your lateral cost per retained hire? If yes, the programme is working regardless of where the percentage lands.

Is this viable for a company hiring only three or four juniors a year?

Yes, and arguably more so, because the mentor-time constraint binds less at small scale. Run two or three interns against one or two expected vacancies. The risk at small scale is the opposite one — over-converting because you have invested emotionally in the cohort. Hold the bar.

How do we avoid training people who then leave for a competitor?

Partly you cannot, and the expected value still favours running the programme. Partly you address it structurally: make the offer early, before the campus placement cycle creates alternatives; and note NACE's finding that returning interns convert at far higher rates. Bringing strong first-year interns back for a second stint is among the most effective retention mechanisms in the data.

What is the single most common reason internship programmes fail?

No approved headcount to convert into. The programme is judged on conversion, but the conversion decision was foreclosed months earlier at budget. Approve the headcount before the cohort starts, or accept that you are running a brand exercise rather than a hiring channel.

The ROI of Hiring Interns: What the Data Actually Says About Cost, Conversion and Retention