The examiners passed it three weeks ago. The corrections were signed off. Then the deposit desk returned it, because the abstract page fields are in the wrong order, the table of contents was typed by hand and no longer matches the pagination, and the file is a standard PDF rather than PDF/A-1.
Your registrar knows exactly how this ends, because it happens every deadline. The candidate is now outside the conferral round they were counting on.
The deposit check is not the examination, and it is not small
It is worth being precise about what the desk enforces, because staff outside the graduate school routinely assume it is a light touch.
MIT’s published specifications for thesis preparation — a mature, public example — run to seventeen separate headings before the copyright sections even begin: pagination, title selection, embedded links, font, spacing, title page, signature page, abstract page, acknowledgments, biographical notes, table of contents, list of figures, list of tables, list of supplemental material, notes and bibliographic references, appendices, and supplemental material.
Several of those are exact rather than advisory. The abstract must generally be no more than 500 words, single-spaced, immediately following the title page, with a prescribed field order: the thesis title, then “by” and the author’s preferred name, then “Submitted to the [Department] on [date] in Partial Fulfillment of the Requirements for the Degree of [name of degree]”, centred — after which the remaining fields are left-aligned rather than centred. The text may be single-, double- or one-and-a-half-spaced, but the abstract, biography, notes, bibliography and acknowledgments must be single-spaced. Notes at the foot of a page must be single-spaced and inside the specified margins.
Then there are the two that defeat people entirely. The deposit format is PDF/A-1, and the guidance is explicit that it begins with accessible source files — which is a statement about how the document was written, not about how it was exported. And the thesis information submitted to the library “must match the title page and abstract of your thesis”, so a title tweaked during corrections and not carried into the submission form is a mismatch the system will find.
Why a return costs a conferral date rather than an afternoon
Because the deposit sits inside a chain of fixed deadlines, and there is no slack in it.
At MIT, thesis information is due to the libraries before the date of graduation; departments must transfer theses to the libraries within 30 days from the last day of class, a requirement carried in Faculty Regulation 2.72; and publication in the repository may begin one week after the later of that transfer deadline or the degree award date. Those intervals are stated numbers, which is good practice — but stated numbers are also unforgiving. A thesis returned two days before the cut-off does not get a two-day extension; it gets the next round.
And a conferral date is not an internal administrative fact. It is the date on which a post-study work visa clock starts, an employment offer conditional on award becomes unconditional, a funding line ends, and a professional registration application becomes admissible. Your desk is not delaying a certificate. It is delaying the events that certificate unlocks — and the candidates for whom the consequence is most severe are usually international, which quietly turns an administrative queue into an equity problem.
The failures are upstream, every time
Look at what actually gets returned and the pattern is consistent. A table of contents that does not regenerate was typed by hand because the headings were never styled. Broken pagination is a missing section break. A list of figures that omits three figures means those captions were typed as body text. Inconsistent references mean the bibliography was assembled manually over four years. A non-archival PDF is a document whose source file was never structured well enough to export cleanly.
None of those is a mistake made at deposit. Every one is a property of how the manuscript was authored, discovered at the last possible moment by the one office with no power to fix it.
That is why the standard responses fail. A longer checklist does not help a candidate who does not know what a section break is. A stricter deadline moves the crisis earlier without reducing it. And pushing the check onto supervisors just relocates the cost onto the scarcest resource you have — which is already happening, as we set out in what your supervisors are actually spending review time on. The writing centre cannot absorb it either; its queue is documented in the waiting list you cannot staff your way out of.
What Tesify for Institutions removes from the desk
The premise is simple: satisfy the specification while the thesis is being written, so that deposit is a transfer rather than an inspection.
- Your specification, applied from chapter one. Section order, front matter, abstract field order and word limit, and the required lists are structural properties of the document from the first draft, not a reformatting job in the final week.
- Generated front matter. Table of contents, list of figures and list of tables are produced from the document’s own structure, so they cannot drift out of step with the pagination.
- A consistent reference apparatus. Maintained as the manuscript grows rather than reconciled against four years of accumulated entries.
- Structured, accessible source. Real heading levels and real captions are what make an accessible, archival export possible at all, which is where the PDF/A-1 requirement actually gets satisfied.
- Metadata that matches. Title and abstract carried from the document into the submission record, so a title changed during corrections does not become a mismatch at the desk.
None of that touches the scholarship, and it should not. The candidate’s argument, method and interpretation stay entirely theirs. What changes is that the manuscript arrives at your desk already shaped like the thing your regulations describe.
Measure it before you buy anything
You almost certainly have the numbers and have never put them together. Three counts, from the last two conferral rounds:
- First-pass acceptance rate at deposit — the share of theses accepted without any return.
- Reason codes for every return, grouped into layout, metadata, file format and other.
- Conferral slippage — how many candidates missed their intended round, and by how long.
If your first-pass rate is above 90% you have an efficient desk and should spend your budget elsewhere. Below 70%, you have a systematic authoring problem and the third number is the one to take to your senior team, because it converts an administrative annoyance into an outcome the institution is accountable for. Your repository deposit rate against theses awarded is a related and equally cheap measure — the method is in how to get a free queryable count of doctoral theses.
Start with one department, not a licence
The next step is a free departmental pilot in one graduate programme for one conferral cycle, with first-pass acceptance rate as the single agreed measure. That is a comparison against your own baseline in your own regulations, and it needs no tender, no capital request and no institution-wide commitment — which is precisely why it clears governance when a purchase order does not. What a full licence should and should not buy you afterwards is set out in what an institutional licence should actually buy.
Request an institutional evaluation and we will work through your specification, your current first-pass rate and a single-department pilot with your graduate school team.
Frequently asked questions
Why do theses that passed examination get returned at deposit?
Because the two checks assess different things. Examiners judge scholarship; the deposit desk enforces section order, layout, metadata consistency and file format against a published specification.
How long is a typical specification?
Longer than most people assume. MIT’s runs to seventeen headings before the copyright sections, several of them exact — including a prescribed abstract field order and a 500-word single-spaced abstract limit.
What is PDF/A-1 and why does it matter?
It is an archival PDF profile required for long-term preservation. It is difficult to produce cleanly from an unstructured source document, which is why the requirement is effectively a requirement about how the thesis was written.
What does a return actually cost the institution?
A conferral date, in the cases that matter. Deposit sits inside fixed calendar deadlines, so a late return moves a candidate into the next round rather than extending the current one.
Is this an equity issue?
Often, yes. The candidates for whom a delayed conferral date is most damaging — visa timing, conditional job offers, funding end dates — are disproportionately international students.
Can we not just publish a better checklist?
A checklist helps a candidate who already knows how to use styles, section breaks and caption fields. It does not help the one who does not, and that is the population generating the returns.
Does this require replacing our repository or student system?
No. The fix sits upstream of both, in how the manuscript is authored. The deposit workflow you already run stays exactly as it is.
What about data protection and student data?
Establish the lawful basis and the processing terms before deployment rather than after, and record them in the same words in your record of processing and your student privacy notice.
Does an authoring platform raise an academic integrity question?
Formatting, front matter and reference consistency are mechanical properties of a document, not intellectual content, and every institution already permits candidates to be helped with them. The argument, method and interpretation remain the candidate’s, and any assistance that goes further should be disclosed under your existing rules.
How long does a pilot take?
One conferral cycle in one programme, measured against your own prior first-pass acceptance rate. Anything shorter cannot produce a comparable number.
What does the pilot cost?
Nothing. It is scoped to a single department precisely so that it can be approved without a procurement process.
What if our first-pass rate is already high?
Then this is not your problem and you should spend the budget elsewhere. Measure before you buy; that is the whole point of the three counts above.