A research-style marketing survey works as an internship project when it starts with one clear decision, asks questions that map to that decision, samples a defined group honestly, and ends with a one-page summary that says what was found and how far to trust it. This guide gives you the design template, question types, sampling notes and summary format to do exactly that in 14 days.
Why a research mindset makes a marketing survey more useful
Many student surveys fail the same way: twenty questions, shared in a few chat groups, producing a pile of answers that nobody can use. The problem is not effort. It is that the survey was never tied to a decision. A research mindset reverses the order. You begin with the question the business needs answered, work out what evidence would answer it, and only then write the questions.
For a digital marketing team, useful decisions are concrete: which message to lead with, which channel to try first, what price band feels acceptable, why people abandon a purchase. A survey cannot settle all of them, but it can narrow the choices and reveal language that customers actually use.
What a survey can and cannot tell you
- Can: awareness, preferences, reasons, habits, satisfaction and the words people use to describe a need.
- Cannot: prove what people will actually buy, since stated intent often differs from behaviour.
- Cannot: represent everyone, if the sample is a convenience group such as classmates.
Say this limit in the final summary. It is part of being a researcher.
The survey design template: from decision to question list
Copy this template into a document and fill it in before you open any survey tool.
| Template field | What to write | Example (illustrative) |
|---|---|---|
| Business decision | The choice the results will inform | Which of two messages should a campus coaching brand lead with? |
| Research questions | Two to four questions the survey must answer | What do students worry about most? Where do they look for help? |
| Target group | Who is eligible to answer | Final-year undergraduates preparing for competitive exams |
| Sample plan | How you will reach them and how many | Share through three student groups; aim for 60 or more complete replies |
| Question map | Each question linked to a research question | Q4 and Q5 serve research question 1 |
| Time limit | Maximum length for a respondent | Under five minutes, about 10 to 12 questions |
| Analysis plan | How you will count and compare answers | Percentages by group; coded themes for open answers |
| Ethics note | Consent line, data storage, anonymity | No names collected; responses stored in one protected sheet |
If a question on your list is not linked to a research question, delete it. That single rule removes most of the clutter.
Free 2-Week CareerSwami Internship
Online, 14 days, with a free offer letter and a free completion certificate once you finish the work. Check with your college that it fits your requirement before you enrol.
Question types and when each one earns its place
| Question type | Best use | Watch out for |
|---|---|---|
| Single choice | Facts and clear categories, such as the platform used most | Missing options; always include “Other” or “None of these” where fitting |
| Multiple choice | Behaviours where more than one answer is true | Long lists; keep to about eight options |
| Rating scale (1 to 5) | Agreement, satisfaction, importance | Label both ends and keep the direction the same throughout |
| Ranking | Priorities among a short list | Anything over five items is tiring; consider picking the top two instead |
| Open text | Reasons and wording in the respondent’s own language | Hard to analyse; limit to one or two per survey |
| Screening question | Filtering for the right respondents at the start | Do not reveal the “right” answer in the wording |
| Demographic | Grouping results by year, city type, or course | Collect only what you will actually use |
Writing questions people can answer honestly
- One idea per question. “How satisfied are you with the price and delivery?” is two questions.
- Neutral wording. “How useful did you find our excellent guides?” nudges the answer; remove the praise.
- Specific time frames. “In the last 30 days, how many times did you…” gives better recall than “often”.
- Plain language. Offer a Hindi or Bengali version if your respondents are more comfortable that way.
- Balanced options. Offer equal positive and negative choices on a scale.
- Allow “I don’t know”. Forcing an answer produces noise.
Sampling notes: who you ask matters more than how many
A hundred answers from the wrong people are less useful than forty from the right ones. Be clear about your sampling approach and honest about its weaknesses.
- Convenience sample: classmates and friends of friends. Quick and free, but biased toward people like you. Fine for a first pilot, with a clear warning in the summary.
- Quota approach: decide in advance that you want roughly equal numbers from, say, first and final year, then keep collecting until each group is filled.
- Snowball: respondents pass the survey to people who fit the target group. Useful for niche audiences but clusters around friend groups.
- Customer list: if the brand shares a permitted list of its own customers, that gives a more relevant sample, and the brand controls contact.
Pilot first, then launch
Send the draft to five people and watch for three things: questions they misread, questions they skip and how long it takes. Fix those, then launch. Remove the pilot answers from the final data unless the questions were unchanged.
A rough guide to sample size for an internship survey
There is no magic number. For a two-week project, 50 to 100 complete responses usually allow simple percentages and group comparisons, with the clear caveat that small groups, such as ten respondents from one course, can swing widely. If you split results by group, report counts as well as percentages so readers can see how small the base is.
Cleaning and summarising the responses
Once the survey closes, work in a fixed order.
- Copy the raw data to a second sheet and never edit the original.
- Remove clear duplicates and responses that fail the screening question.
- Check for patterns such as all answers marked the same option, which suggest careless responses.
- Count answers for each closed question and turn them into percentages of valid responses.
- Code open-text answers into three to six themes by reading them all once, then tagging each.
- Compare groups only where each group has a reasonable number of replies.
The one-page insight summary format
| Section | What it contains |
|---|---|
| Decision and question | One sentence each |
| Who answered | Number of valid responses, group breakdown, how they were reached |
| Top three findings | Each finding with the percentage, the question number and a plain-language meaning |
| Voice of the customer | Two or three short, anonymised quotes or paraphrases from open answers |
| Recommendation | One action the team could test next, tied to a finding |
| Limits | Sample bias, small groups, stated versus actual behaviour |
| Next step | A follow-up question or small test to confirm the finding |
Illustrative example: a survey for a campus tutoring service
The following scenario is invented to show how the template works.
A hypothetical tutoring service wants to know whether to lead its social posts with “affordable fees” or “small batches”. The research questions: what do students worry about when choosing a tutor, and where do they look first? The survey has ten questions: one screening question (are you currently preparing for an exam?), two multiple-choice questions on information sources and concerns, a ranking of four factors (fees, batch size, teacher background, flexible timing), a five-point scale on agreement with two statements, one open question and three demographic items.
Suppose 64 valid responses come in. In this made-up data, 48 percent rank batch size in the top two factors and 41 percent rank fees there, and open answers mention “personal attention” in 17 comments. The summary would say: “Batch size was ranked in the top two by slightly more respondents than fees (48 percent against 41 percent, n = 64). The gap is small, and the sample is mostly second-year students from one city, so test both messages in a small post trial before changing the main campaign.” That wording gives a recommendation and states the limit.
Jadavpur University students: aligning a survey project with internship and research expectations
Jadavpur University, in Kolkata, treats internships differently across its faculties, so the first step for any student is to find out which rule applies to their department. The Placement and Training page lists two internship policy documents: a PG long-term policy and a UG and PG short-term policy, both under the engineering and interdisciplinary studies faculties. If your programme falls under either faculty, read both documents in full, because the page itself gives only their titles.
For arts programmes under the NEP framework, the examination rules describe a UG summer internship of 2 credits with a minimum of 7 days and 50 marks split between a report and a seminar or viva, and state that the UGC guidelines for undergraduate internships (2023) are followed as far as practicable. In the engineering faculty, the 2025 UG examination rules weight internship assessment 70 percent external, from a mentor at the host organisation, and 30 percent internal. These are different schemes, so a survey project reported in an arts programme will be judged differently from one in engineering.
The university’s NAAC self-study report says industry internships are encouraged in most programmes, especially engineering. A research-style marketing survey suits students who need a report plus a viva or seminar, because it naturally produces a method, data and a discussion of limits. Rules may have been revised since these documents were published, so treat them as a starting point.
Verification steps for Jadavpur University students planning a marketing survey as internship work
- Ask your department office which internship rule covers your programme and batch, and whether a digital marketing project done online is acceptable.
- Request the current internship guideline: minimum days, credits, report format and the split between report and viva or mentor assessment.
- Ask who must assess the work if the mentor is outside the university, and what signed letter or certificate they need to provide.
- Check whether your department requires ethical approval or a faculty guide’s sign-off before surveying people.
- Confirm how survey data, respondent anonymity and any brand name may be shown in your report.
Ethics and data care in a student survey
Surveys collect information from real people, so keep it simple and respectful. State at the top who is running the survey, why, how long it takes and that answers are voluntary. Do not collect names, phone numbers or email addresses unless you need them, and say how you will protect them if you do. Never share raw responses in public groups. Take extra care with minors: if any respondents may be under 18, check with your supervisor before collecting anything.
Do not mislead respondents about the purpose. If you are working for a brand, say so in plain words. If you offer a reward, state the rules for it clearly, and do not promise something the team has not approved.
Frequently Asked Questions
How many questions should a marketing survey for an internship contain?
About ten to twelve, taking under five minutes. Every question should link to a research question in your template; delete any that do not.
What sample size is enough for a two-week survey project?
Fifty to a hundred valid responses allow basic percentages. Smaller groups within that total should be reported with counts and a warning that they are small.
Is it acceptable to survey only my classmates?
It is acceptable as a pilot or when the target audience is students like you, but the summary must state that the sample is a convenience sample and may not represent other groups.
How do I avoid leading questions?
Use neutral wording, remove praise or hints from the question, balance the answer scale and test the draft with a few people to see whether they read it differently from your intent.
How should I analyse open-text answers?
Read them all once, define three to six themes, tag each answer to a theme and count the tags. Add two or three anonymised short quotes to the summary.
What should I do if the results are unclear or contradict each other?
Report them as they are, explain possible reasons such as a small sample or ambiguous wording, and suggest a follow-up question or small test. Hiding uncertainty weakens the work.
Does survey data replace testing a campaign?
No. A survey shows stated preferences. A small live test of two messages shows behaviour. The two together are stronger than either alone.
Can a short online internship include a customer survey project?
Yes, when it is scoped tightly: design in the first days, a pilot and launch in the middle, and analysis plus the summary at the end. The free two-week online internship at CareerSwami suits that rhythm of a small, finishable research task.
Related Digital Marketing guides
- Indira Gandhi University Meerpur: Podcast and Newsletter Content Repurposing for Interns
- Raja Mahendra Pratap Singh State University: Weekly Reporting Template for Digital Marketing Interns
- Jagan Nath University: Presenting an Internship Campaign Case Study to Recruiters
- All Digital Marketing internship guides
Ready to put this into practice?
Join the free two-week CareerSwami online internship in Digital Marketing. The internship, the offer letter and the completion certificate are free.














