Turn a Final-Year Project into a Business Hypothesis Before Building an App

Tom — Singapore Digital Business & Systems Consultant

Tom

Digital Business & Systems Consultant · Singapore

7 min read Article date First published
Turn a Final-Year Project into a Business Hypothesis Before Building an App

Treat a final year project startup idea as a hypothesis, not a business. Your grade shows you can build a solution, not that anyone will change how they work to use it. Before writing more code, check who owns the project, write down who the customer is and what adoption would require, then deliver the outcome by hand to a few real users. If nobody wants the manual version, a polished app will not change their minds.

Key takeaways

  • A grade measures your work against academic criteria. Demand shows when people change what they do, give their time or pay.
  • Ownership is not automatic either way: check your institution's IP policy and any sponsor or industry-partner agreement before commercial use.
  • Write down the customer, the job, today's workaround and what adoption needs: data, training and approvals.
  • Test the riskiest assumption by hand, and decide in advance which result would change your mind.

What your grade proves, and what it does not

Your project was assessed on things like technical difficulty and method: real achievements for your portfolio. A customer asks a different question: is this worth changing my routine for?

Your project showsIt does not yet show
The app works in a demoThat anyone will change their routine to use it
Test users finished the tasks you setThat they would come back without being asked
Your report compared existing solutionsThat today's workaround is not good enough
Your supervisor rated the approachWho would pay for it, approve it or roll it out

Check ownership and permitted reuse first

I cannot tell you who owns your project. It can depend on your institution's intellectual property (IP) policy, any sponsor or industry-partner agreement, and who else contributed, and these terms vary between institutions and projects. Do not assume the code is yours to sell because you wrote it, or that it cannot be.

  • I have read my institution's current IP policy for student work.
  • I know whether the project had a sponsor or industry partner, and I have read that agreement or asked my supervisor about it.
  • I have talked to teammates about reuse and noted what my supervisor or lab contributed.
  • I have checked the terms of any dataset, API or third-party code used.
  • I know who to ask if unsure: my supervisor or the office the policy names.

Do this before commercial use, such as selling or licensing the app or presenting it as an available product. The manual test below needs none of your project code, so you can gather evidence while those questions are answered.

Write the customer and adoption hypothesis

Most final-year projects start from a technology. A business starts from a person with a job to get done. Rewrite your project as one sentence you could prove wrong:

"[Customer] who [situation] currently [workaround]. They would switch to [my outcome] if [adoption conditions], and [buyer or approver] would agree because [reason]."

Two parts trip up student projects. The user is often not the buyer: staff may use a tool the owner must approve and pay for. And adoption has its own cost: someone must supply the data, find time for training, get approvals and fit the tool into a full day.

The project-to-market evidence canvas

Fill in one copy per customer type. Write the change-my-mind result before the test, not after. The third column is from the fictional example below.

BoxQuestion to answerExample (fictional)
CustomerWho has the problem, and who decides or pays?Owners of small neighbourhood minimarts; the owner decides
Job to be doneWhat are they trying to get done?Reorder stock before shelves run empty
Current workaroundWhat do they do now, and what does it cost?A shelf walk at opening, then messages to suppliers
Adoption needsWhat data, training and approvals are needed?Daily shelf photos from staff; no outside approval
Riskiest assumptionWhich belief sinks the idea if wrong?Shops will send shelf photos every day
Manual testHow will you deliver the outcome by hand?Reorder lists written by hand for 3 shops over 4 weeks
Change-my-mind resultWhich result means change or stop?Fewer than 2 shops send photos on most days

For the customer conversations behind each box, use my five neutral interview questions; they suit any course project. For a budget cap and stop rule, borrow the one-page experiment card.

Test a manual alternative before more code

A manual test delivers the outcome your app promises, by hand, to a few real people. If your app schedules, you schedule. You learn whether people want the result, and what adoption really involves, before spending months on software.

  • Deliver the outcome, not the interface. A spreadsheet, a form and messages are enough.
  • Keep it small. My suggestion: three to five users for two to four weeks.
  • Be honest. Call it a pilot done by hand, and take no payment for something you cannot keep delivering.
  • Fit it around your studies. Run it after submission or during a break.

If the manual version is wanted and keeps hitting the same limit, that limit is your first real software requirement. Sometimes the manual version is the business: see starting with a service rather than an app.

A fictional example: the photos never came

(fictional example) Dinesh is a final-year computing student. His project app reads shelf photos and suggests a reorder list; it earned a good grade. His hypothesis: minimart owners lose sales to empty shelves and would send daily shelf photos for a reorder list.

His project had no industry partner, but he still reads his university's IP policy and asks his supervisor what it means for him. His manual test uses no code: after submission, three minimart owners he knows agree to message him shelf photos, and he writes each reorder list by hand, unpaid.

The result changes the plan. Over four weeks, only one shop sends photos on most days; staff are too busy at opening time, so his riskiest assumption fails. Two owners say supplier reps already check shelves and suggest orders. But all three, unprompted, mention chilled items expiring unnoticed.

The upside: a clear answer from four weeks of evenings, with no extra code. The effort and uncertainty: no sale, an unused model, and no proof that expiry checks are worth paying for. His next canvas tests a weekly expiry check done by hand.

When more code makes sense

Write more code when the canvas shows a repeated, wanted outcome, the manual version is straining and you know what adoption needs. If customers or staff would log in to it, such as a booking tool, client portal, dashboard or first product version, that is web app development: I start with a requirements workshop and scope and quote each project individually. If the evidence points to an internal workflow passing through WhatsApp and spreadsheets, operations digitalisation fits better.

You own the idea, the customer relationships and the business decisions. I deliver an agreed technical scope as a separate paid project, not as a co-founder or equity partner. If testing leads you to a graduate job instead, that is a good outcome too.

Your next step

Fill in the canvas this week, starting with the riskiest assumption. Run one manual test after submission, and keep a problem notebook for anything unexpected it turns up.

Frequently Asked Questions

Does a good grade prove market demand?

No. A grade assesses your work against academic criteria. Demand shows when people with the problem change what they do, give their time or pay.

Who owns a final-year project?

It depends on your institution's IP policy, any sponsor or industry-partner agreement and who else contributed, and these vary. Check them, and ask your supervisor if unsure, before any commercial use.

Can I test demand without using my project code?

Usually, yes. A manual test delivers the outcome by hand, so it needs no project code. Check ownership before reusing the code, report or data.

What if the manual test shows nobody needs it?

That is a cheap, useful result. Keep the project as portfolio evidence within your institution's rules, then change the customer or the job, or park the idea.

Do I have to start a business after graduating?

No. A graduate job is a good choice, and the canvas still teaches you why products succeed or fail. You can test in small steps later, within your employment terms.

Filled in the canvas and run a manual test? Send me the canvas, with no customer names, and I'll give you an honest view of whether the evidence points to a web app yet or whether a spreadsheet will do for now. It is a conversation, with no commitment.

No obligation. No sales pitch. Just an honest conversation.

தொடர்பு