Clients ask for research. You tell them you'll run a survey. Then you open the first guide on page one of Google and it recommends a $99 a month platform with a whole research team behind it. That's not how freelancers work.
This is a working guide written for people who need to deliver real research on a real budget. Six steps to go from a client brief to a delivered report, using tools you can afford and a workflow you can finish in a day. If you've run a survey before but keep hitting walls — bias in the data, low response rates, exports that embarrass you — this will fix that.
What Market Research Surveys Are Actually For
Three jobs. That's the whole list for most freelance work.
- Validate an idea. A client is launching a feature or product. You test demand, pricing, and willingness to switch before they spend money building the wrong thing.
- Understand an audience. Demographics, pain points, buying habits, where they hang out, what they already use. Fuel for positioning, pricing, and content strategy.
- Measure satisfaction. NPS, CSAT, churn signals. For clients who already have customers and want to know whether they're happy.
Everything else — concept testing, naming, price sensitivity, segmentation — slots into one of those three. If your client's brief doesn't fit cleanly into one, push back before you build anything. A survey without a clear job is just a form.
Step 1 — Define What You're Trying to Learn
Before you touch any tool, write the research question in one sentence. One.
Bad: "Learn more about our customers." That's not a research question, it's a vague wish, and a vague wish won't tell you what to ask or who to ask it to.
Good: "Do SMB owners in home services prefer monthly or annual billing, and what's their price ceiling for a scheduling app?" Now you know what to ask, who to ask, and what the client gets at the end.
Three things to agree with the client on paper before you move:
- The research question — one sentence.
- The hypothesis — what you expect to find and why. Even a rough guess forces clarity.
- The decision the data will drive. If no decision is on the line, don't run the survey. Save everyone the time.
Write these at the top of a shared doc. Every question you consider later gets pressure-tested against them. If a question doesn't help answer the research question or inform the decision, it's dead weight.
Step 2 — Choose Your Survey Type
Three structures cover almost every freelance brief. Pick the one your research question demands — don't mash them together.
Screener Surveys
A short filter at the start of a larger study. Five to seven questions that qualify the respondent before letting them continue. Non-smokers can't answer questions about cigarette brands. Non-paying users shouldn't tell you about pricing. Use a screener any time your audience is narrow or your client is paying per completed response.
Exploratory Surveys
Open-ended, qualitative, small sample. You're looking for patterns, language, and pain points you didn't anticipate. Think 30–50 respondents, four or five open-ended questions, and a lot of reading. Useful at the start of a project when you don't yet know what to measure.
Quantitative Surveys
Closed questions, Likert scales, multiple choice, ranking. You already know the hypotheses — you're measuring. In my work I treat 100 completed responses as a working floor for a single segment, more if the client wants cuts by role or geography. Smaller samples still show patterns, but I would not stake a pricing recommendation on them.
Most freelance projects use a screener plus one of the other two. Rarely both exploratory and quantitative in the same survey — that's two studies, not one, and respondents will quit halfway.
Step 3 — Write Questions That Don't Bias Your Data
This is where most freelance surveys fall apart, and the tool is almost never the reason. A biased survey hands the client numbers they can't defend in a boardroom, which is a worse outcome than running no survey at all - at least a no-survey conversation stays about the decision.
Rules I use on every project:
- No leading questions. "How much do you love our checkout experience?" assumes love. Ask "How would you rate our checkout experience?" with a neutral scale.
- One idea per question. "Is the pricing clear and affordable?" is two questions. Split them. If a respondent agrees with one half and not the other, you lose the signal.
- Mix closed and open. Closed for measurement, one or two open-ended at key moments for context. "Why?" after a rating is the most useful question in any survey.
- Use skip logic. Nobody should see a question that doesn't apply to them. A respondent who said "no" to using the product shouldn't be asked which feature they like most. It confuses them and pollutes the data.
- Keep it under ten minutes. Completion rates fall off a cliff after that. Fifteen to twenty well-chosen questions is a good target. If you can't answer the research question in that space, you're asking the wrong things.
Skip logic is the single feature that most freelance survey work needs and most cheap tools don't provide — or lock behind a $99 plan. If you're serious about this work, pick a tool that treats skip logic as a baseline, not an upgrade. Here's the difference between skip logic, display logic, and branching if the terms are getting mixed up.
Step 4 — Build and Test Your Survey
Building the survey is the fast part if you've done Steps 1 to 3 properly. If you haven't, no builder in the world will save you.
My build sequence inside the tool:
- Set up the screener at the top. Three to five questions. Any disqualifying answer routes to a polite end screen that thanks the respondent and ends the survey.
- Add the main questions in a logical order. Easy and demographic questions first, more considered or sensitive ones in the middle, open-ended at the end when respondents are warmed up.
- Use skip logic liberally. Section only visible if a prior answer qualifies. Follow-up only shown if the main answer triggers it.
- Add a short closing — thank-you, estimated results date if the respondent gets any, and a one-line incentive confirmation if you promised one.
Test it on three to five real people before launch - actual people who match your target respondent profile, not colleagues and not the client, both of whom will skim. Watch them take it on their phone too, because an unfortunate number of surveys look fine on desktop and fall apart on mobile.
What you're watching for in the test:
- Any question they have to re-read.
- Any answer option they want to pick but can't.
- How long it actually takes — not the tool's estimate.
- Whether they understand what you're asking at every step.
Ready to build your first screener? Start on the free QPoint plan — 100 responses a month, no credit card, enough to prototype any survey before you commit to a paid plan.
Step 5 — Distribute It (and Get Enough Responses)
A good survey nobody takes is worthless. Freelancers rarely have an audience of their own, so distribution means borrowing one.
Channels that actually work:
- The client's own list. Easiest and best when it fits the research question. Existing customers, newsletter subscribers, past leads. Response rates are higher because the brand is familiar.
- LinkedIn. Reliable for B2B research when the client has a credible profile. On the briefs I have run, a targeted post from the client's account has landed somewhere between fifty and a few hundred responses in a week - the floor is a lot more predictable than the ceiling.
- Prolific or User Interviews. Paid respondent panels. Expensive per response ($2–$4 for screened respondents), but fast and reliable. Use when you need a specific demographic the client can't reach.
- Reddit and Facebook communities. Niche-specific and free, but check community rules first — most ban self-promotion, some allow research posts if you ask moderators.
- QR codes for in-person research. Events, retail locations, trade shows. Shortcut for research briefs tied to a physical location.
Response targets by survey type:
- Exploratory / qualitative: 30–50 completed responses.
- Quantitative: 100+ for a single segment, 300+ if you need to compare segments.
- NPS/CSAT tracking: whatever the client agreed to as the sample size up front.
Keep the survey open until you hit your target or the response rate dies — not both. Clients who say "let it run another week, just in case" are padding the number, not improving it.
Step 6 — Analyse and Deliver
Once the responses are in, you reach the part of the project that actually earns the invoice.
Workflow I use every time:
- Export the raw responses to Excel. Filter and cross-tab there — it's faster than any dashboard for small-to-medium samples.
- Build the client-facing deliverable as a short PDF or deck. Methodology, three to five key findings, one chart per finding, and a recommendation section tied back to the decision you agreed on in Step 1.
- Attach the raw Excel export as an appendix for any client who wants to dig deeper. Most won't. The ones who do appreciate that you included it.
What goes in the report:
- Methodology. One short paragraph. Sample size, dates in field, how respondents were recruited, screening criteria.
- Key findings. Three to five, in priority order. Each one gets a chart and two sentences of context. No more.
- Recommendations. Tied directly to the decision from Step 1. "Launch the annual billing tier at $X" — not "the data suggests a possible preference for annual billing."
- Caveats. One short section on limitations. Sample size, selection bias, anything the client should factor in. Being honest here protects you later.
A survey tool that exports clean Excel and PDF saves hours per project. If you keep reformatting ugly CSVs in Google Sheets late the night before a deliverable, that is a useful signal to pressure-test the tool choice on your next brief.
A Freelance Scenario, Start to Finish
A client sells accounting software to small home-services businesses — plumbers, electricians, HVAC. They want to launch a scheduling add-on. Before building it, they want to know if their customers would pay, and how much.
This is a composite from two scheduling-tool briefs I ran last year - names and the exact numbers changed, workflow is real.
Research question: "Would existing customers pay for a scheduling add-on, and what's the acceptable monthly price range?"
I build a 12-question survey in QPoint. Three-question screener up front: current customer, has at least two field techs, personally schedules jobs (owner or dispatcher - not an admin who only files invoices). Anyone who fails is routed to a polite thank-you page. Everyone who passes gets a quick fork on current scheduling tool use, five quantitative questions on pain points and willingness to pay, two Likert scales on the proposed feature set, and two open-ended questions at the end asking what would make it a must-have.
Skip logic routes respondents who already use a scheduling tool through a different willingness-to-switch question than respondents who don't. Display logic shows a pricing matrix only to respondents who said they'd be interested in principle. The client's list is 4,200 customers; I send to a random 1,500. Response rate landed at around 11% by day four - 167 completed, 112 after the screener - which was enough for the pricing cut the client needed.
Analysis in Excel took closer to three hours than two, because I had misread two of the open-ended questions during the pilot and had to recode the responses before any of the charts made sense. Six-page PDF for the client, three key findings, one price recommendation, raw export attached. Total project time: a day and a half. Tool cost for the month: €20.
The Tool Stack
You don't need an enterprise research platform to deliver work that stands up in a client meeting. Keep the stack lean.
- Survey tool. QPoint Professional at €20/month handles skip logic, display logic, unlimited questions, 3,000 responses, Excel + PDF export. Builds what you need.
- Respondent panel (when needed). Prolific for most B2C work, User Interviews for B2B — pay only for what you use.
- Analysis. Excel. Free, universal, every client can open it. Save the dashboards for in-house teams.
- Deliverable. QPoint PDF export, lightly reformatted. Done.
For the full breakdown of why this stack beats the $99/month options, read Best Survey Tools for Freelancers. Short version: you're paying for a team you don't have.
Six Steps, One Day
Market research is not a mystery. For most freelance briefs it is six repeatable steps and a tool that stays out of your way.
- Define the research question in one sentence.
- Pick the right survey type — screener, exploratory, or quantitative.
- Write unbiased questions, use skip logic, keep it short.
- Build and test on real people before launch.
- Distribute through channels that match the audience.
- Deliver a tight report tied to the decision on the line.
Run the sequence on your next three briefs and most of them will land as research the client actually uses, instead of a deck that sits in a folder nobody reopens. For more on how survey design shapes the quality of your findings, see The Impact of Survey Design on Data Quality.
Build your next client's market research survey this week. Start free on QPoint — upgrade to Professional (€20/month) when you're ready to launch. Skip logic, display logic, and Excel exports are included from day one. No upgrade walls mid-build.