When an internship cohort underfills, the diagnosis is usually "the market is tight". In most programmes the real cause is measurable, internal and fixable.
These are the five places candidates leave, in descending order of how much they cost you.
Latency between final interview and offer
This is the largest single source of loss and the easiest to fix. Every week of delay after the final interview reduces acceptance, and in scarce categories it reduces it steeply — a strong candidate is very rarely in only your process.
The fix is administrative, not strategic: pre-clear the approval chain before the process starts, and give the hiring manager a decision deadline rather than a decision request.
Unreviewed take-home assignments
Long take-home assignments that receive no feedback are the fastest way to convert an interested candidate into a detractor. Students talk to each other, and a reputation for wasting people's time propagates through a campus in a single season.
Either keep the exercise short enough that reviewing every submission is realistic, or replace it with a live exercise.
Silence between stages
Early-career candidates read silence as rejection, because in their experience it usually is. A two-week gap with no contact loses candidates who were never formally rejected and would have accepted.
A single templated status email at each stage transition eliminates most of this loss for essentially no cost.
A vague description of the actual work
Candidates decline offers when they cannot picture the job. "You will work with cross-functional teams on impactful projects" tells them nothing, and in a competing offer that describes a specific project with a named mentor, the specific one wins.
Name the team, the project, the tools and the person they will report to. This costs nothing and is one of the highest-leverage changes available.
No conversion story
Students accept internships largely on the strength of what comes after. A programme with no articulated path to a full-time offer — even an honest "roughly this proportion convert, and here is what that depends on" — competes badly against one that has it.
Vagueness here is read as bad news, correctly.
Key takeaways
- Offer latency is the biggest and most fixable source of loss; pre-clear approvals before the process starts.
- Never run an unreviewed take-home — the brand damage outlasts the requisition.
- Templated status emails at every stage transition recover candidates lost to silence.
- Describe the actual work, team and mentor. Specificity beats a better package surprisingly often.
- State the conversion path honestly. Vagueness is read as bad news.
Questions
What acceptance rate should we expect?+
It varies enormously by category and by how fast you move. Rather than benchmark against an industry figure, track your own by institution and by channel — the internal variance will be larger than the external.
Is compensation the main reason candidates decline?+
Less often than teams assume at early-career level. Process speed, clarity about the work and the conversion story routinely outweigh moderate stipend differences.
How do we measure where drop-off happens?+
Instrument each stage transition with a timestamp and a status. Most programmes cannot answer this question at all, which is why the diagnosis defaults to "the market".