In most tutoring centres, mock exam analysis starts with a spreadsheet exported from an optical mark reader or an online exam platform, and it fills a coordinator’s evenings. Here is where that process breaks down, how an automated flow from results to parent report works, and what to prepare before you start.
Where manual mock exam analysis breaks down
In Turkey, students preparing for the national entrance exams (LGS at the end of middle school, YKS for university) sit mock exams all year, and results are counted in nets: correct answers minus a share of the wrong ones. After each exam the centre has a file, and the same steps follow by hand.
- Matching students: finding everyone who mis-marked their number, left the booklet type blank or is spelled differently on the class list.
- Counting nets: correct and wrong answers totalled subject by subject, with a separate net formula for each exam type.
- Topic and outcome breakdown: each question’s topic looked up in another table. When time runs short, this step goes first.
- Comparison: earlier results sit in other files, so following a student’s trend means opening several side by side.
- Sending to parents: a PDF per student, sent one by one over WhatsApp. One file to the wrong parent shares another child’s results.
So the analysis often ends as a ranked list. The student sees their position and has to work out from the table which topic to study. The parent waits for the next parents’ evening to ask the teacher what the number means.
The automated flow: from results file to teacher dashboard
The results file is uploaded or pulled straight from the online platform, and each row is linked to a student by number. Rows with an unknown number or a blank booklet type go to a separate list for the coordinator.
Matched answers are checked against the key and nets are calculated by subject and topic. Because each question is tied to a learning outcome (the curriculum objective it tests), you also see which outcomes each student missed. The report and the class-level teacher dashboard are built from the same data.
Our own product, Hippotenus, also analyses mock exam results at learning-outcome level and finds each student’s gaps, with admin, teacher, parent and student seeing the same data. It taught us which screen each person needs and where the data gets stuck.
What outcome-level analysis shows that a ranking can’t
Two students with 12 nets in maths sit side by side on the list. One lost marks in a single topic, the other across five. Revising that topic may be enough for the first; the second needs a broader plan. Outcome-level analysis shows the difference at a glance.
It helps at class level too. If most of a class misses one outcome, the cause is probably how the topic was taught, and the teacher can pick it up next lesson. In a spreadsheet that table is extra work; in an automated flow it is already on the dashboard.
Five things a student and parent report should show
Nets
Correct, wrong, blank and net for each subject, with the overall score if there is one.
Weak topics and outcomes
The few outcomes where most marks went, in priority order. A long table wears parents out.
Trend across exams
How nets have moved over recent exams. One bad paper reads more fairly inside a trend.
Position in the group
Standing against the class and centre average. Rank always appears next to the average.
A short comment
Two or three sentences from the teacher on what to work on before the next exam.
The teacher dashboard is the class view of the same data: how many students struggled on each outcome, and whose nets have dropped lately. Reports awaiting approval show up here too. The coordinator gets branch and class comparisons.
What the centre needs to prepare first
- 01
Student list
Name, number, class and branch, plus parent contacts. The numbers must match what students mark on the optical form.
- 02
Answer keys
One per exam and booklet type. Publishers supply the key for bought-in exams; for your own, a teacher enters it.
- 03
Outcome mapping for each question
The subject, topic and outcome of every question. This takes the most effort; if the publisher supplies a table, most of it is done.
- 04
Exam calendar
Which exam runs when, and for which classes, across the term. Trends and comparisons follow this calendar.
Mapping is done once per exam and reused in every branch that sits it. For the centre’s own exams, the teacher who writes a question picks its outcome when adding it to the question pool.
AI drafts the comment, the teacher approves it
The comment is the most-read part of the report and the slowest to write. Two or three sentences per student fills a teacher’s evening after each exam at a busy centre.
AI helps here. It reads the nets, weak outcomes and trend and writes a draft. The teacher reads it, corrects or rewrites it, and approves it. The system never sends an unapproved comment to a parent.
Fine to leave to AI
- Drafting a comment from nets and weak outcomes
- Summing up the recent trend in one sentence
- Consistent wording for students in similar positions
Stays with the teacher
- The final wording and the approval
- What the data can’t show, such as a student being ill on exam day
- Sensitive matters that need a conversation with the parent
How long it takes to build exam results reporting software
The biggest factor is where the results come from. The tiers from our dashboards and reporting service map onto exam data like this:
| Where results come from | What’s required | Time |
|---|---|---|
| Optical reader output or Excel / CSV files | A consistent file format or an upload screen | 1-2 weeks |
| The centre’s existing database | Read-only access | 1-2 weeks |
| Online exam platform with an API | API key and permissions | 1-3 weeks |
| Platform without an API | An export flow or an integration layer | 3-5 weeks |
| Several sources combined | A shared identity field (such as student number) | 4-8 weeks |
These figures cover getting the data in and the dashboard. A first version covering mock exam analysis and the parent report typically goes live in 2-4 weeks, with homework, payments and a mobile app in later releases. Avoid switching during busy exam weeks, and pilot with one class.
One point of contact
In our team of seven, one product owner handles all communication with the centre.
A working version every week
Something to try each week, plus an open task list you can see.
Scope changes in writing
The time and budget impact of any change, in writing, within 24 hours.
Dashboards and reporting
Scope and timelines of our dashboard and report automation service.
Frequently asked questions
Our optical reader produces a non-standard file. Is that a problem?
No. The format is defined once: which columns hold the student number, the booklet type and the answers. If it changes between exams, you match the columns on an upload screen. A flow fed by Excel or CSV files is typically ready in 1-2 weeks.
Can we pull results from online exam platforms too?
If the platform has an API, results are pulled automatically; that scope takes 1-3 weeks. Otherwise we use the platform’s export file. We confirm which route works on the first call.
Who does the outcome mapping?
The centre does, since teachers know the outcome structure best. For bought-in exams, check whether the publisher shares an outcome table you can import. For your own exams, the outcome is picked when the question is written.
Can we keep sending reports over WhatsApp?
The notification can go by WhatsApp or SMS. The report itself is safer behind a parent login, where each parent sees only their own child. Student data is personal data and mostly concerns minors; under Türkiye’s data protection law (KVKK), the privacy notice and any consent needed are the institution’s responsibility.
Why custom software rather than a ready-made analysis package?
If your exam format is standard and the package’s report works for you, a package gets you started faster. If you have your own format, or want parents to see a different report, custom software fits better.
Who owns the software and the data?
You do. On final payment the source code and IP are transferred to you. We fix bugs found in the first 30 days after launch free of charge.