Blog

Preliminary consumer research before you field a survey

Preliminary research is the low-cost step before a survey: start from a goal, list the questions and check what people already said. You then field a survey for the questions public opinions cannot answer, so it is shorter and better aimed.

What is preliminary consumer research?

Preliminary research, also called exploratory research, is the low-cost step that turns a theory into a hypothesis worth funding. It comes before the expensive work: the panel, the fieldwork and the interviews.

Most research starts with a questionnaire. Preliminary research starts one step earlier, with a goal. “Why are buyers moving to a rival brand?” is a goal. “How satisfied are you with the product?” is a survey question.

From the goal, you build the questions. Then you check which of them people have already answered in public, in their own words. The output has three parts:

  • a list of questions, ranked by how much public evidence answers each one;
  • answers to the questions with strong evidence, with quotes and sources;
  • a shorter list of questions that still need a survey or an interview.

The result is a shorter survey that spends your budget on the questions people have not answered in public.

Here is how that looks for one goal: “Why do lapsed buyers stop buying our oat bar?”

Draft questionGradeWhy
Why did you stop buying it?StrongMany people explain this in reviews and forums
What did you buy instead?StrongPeople name the bar they switched to
What do you eat it with, and when?SomeFewer posts, but they are specific
How much would you pay for a larger pack?Needs a surveyPrice needs a controlled question
Would you try a new flavor?Needs a surveyNobody has tasted it yet

The team now fields two questions instead of five. It also starts the survey knowing the likely reasons, in buyers’ own words.

Preliminary research also improves the questions you keep. People often describe a product in words your team never uses. A survey written in their words is easier to answer.

Can you get survey answers without recruiting a panel?

For some questions, yes. People already explain why they chose a product, what disappoints them and what they compared it with. They do it in reviews, videos, comments, forums and articles, without being asked.

Questions that public opinions answer well:

  • Why do people choose this product over others?
  • What do they like and dislike about each feature?
  • What do they compare it with, and what do they switch to?
  • How and when do they use it?
  • What do they ask before they buy?
  • Which words do they use to describe it?

Questions that still need a panel:

  • How many people hold a view, as a share of all buyers?
  • Will people buy, and at what price?
  • How do people react to a product that is not on sale yet?
  • How many people know the brand?

Public opinions show reasons, not rates. A panel gives you rates. Preliminary research tells you which of your questions need which.

Pay attention when public answers differ from what your team expected. If you assumed price drives people away, but posts talk about taste, test that gap before you write the survey. A survey built on the wrong assumption measures the wrong thing very precisely.

How do you turn online reviews and posts into survey-style answers?

Treat each opinion as an answer to a question nobody asked out loud. Then group the answers the way a survey would.

  • Set the scope. Name the products, competitors, market and time window. A loose scope pulls in noise.
  • Gather from many streams. Reviews, videos, posts, comments, forums and articles show different sides of the same product.
  • Read each stream for what it does best. Video reviewers explain, forum users troubleshoot and store reviewers judge after weeks of use.
  • Tie each opinion to a product and a feature. “The battery dies by noon” is about one model’s battery, not the brand.
  • Count distinct people. One person who posts many times is still one person.
  • Group answers into options. For “Why did you stop buying?”, the options might be price, taste, availability and a better rival. Use the words people used.
  • Show the spread. Say how views split, not the most common view alone.
  • Keep every quote linked to its source. A reader should be able to open the original post.

Then compare the streams. If store reviewers and forum users give the same reasons, the answer is stronger. If they disagree, say so, because the split may reflect different kinds of buyers.

The result reads like a survey answer, but it carries a different promise. It says what people who chose to speak said, not what all buyers think.

How do you run preliminary research, step by step?

Use this checklist for each project.

  1. Write the decision. One sentence: what will your team do differently, depending on the answer?
  2. State the goal or hypothesis. For example: “Buyers leave because the refill is hard to find.”
  3. Draft the questions. Write the questions a survey would ask about that goal.
  4. Set the scope. Products, competitors, market, languages and time window.
  5. Gather public opinions. Read across reviews, videos, posts, comments, forums and articles.
  6. Grade each question. Strong, some, thin or needs a survey. Base the grade on how many distinct people answered and whether sources agree.
  7. Answer the strong questions. Give a plain answer, the spread of views, linked quotes and the caveats.
  8. Add the questions you did not ask. People often raise topics your list missed. Those topics can matter more than your draft questions.
  9. Write the shorter survey. Keep the questions that need a count, a price or a reaction to a new idea. Draw the answer options from people’s own words.

A good grade is honest. A question graded “needs a survey” is a useful result, because it tells you where to spend.

Share the graded list with the team that asked the question. It shows what is already known, what is likely and what still needs money.

Can you run a concept test without recruiting a panel?

Not fully. A concept test shows people an idea that does not exist yet and measures their reaction. Nobody can post about a product they have never seen.

Public opinions can still do useful work before the test:

  • Check the need. Are people asking for this? Do they complain about the gap it fills?
  • Find the objections. What do people dislike about similar products today?
  • Borrow the language. Which words do people use for the benefit you plan to claim?
  • Pick the comparison. Which products will buyers compare your concept with?

With that groundwork, the concept test asks sharper questions. You may also find that the need is weaker than you thought, before you pay for the test. To screen many ideas at low cost, read when synthetic respondents fit.

What are the limits of preliminary research?

Preliminary research cannot give you a share of all buyers, and its evidence is thin for new or small products.

  • People choose to post. You do not know how many people stayed silent, so you cannot compute a share of all buyers.
  • A few voices can dominate. Heavy posters and people with strong views appear more often. Counting distinct people reduces this.
  • Not every post is organic. Some are sponsored. Label them and decide what to filter.
  • A post does not prove use. Write “people who say they used it”.
  • Coverage varies by category. Some products get discussed a lot. A small or new product may have thin evidence.
  • Markets and languages differ. People in one country may discuss a product in ways that do not carry to another. Scope the research to the markets you sell in.

Grade each answer against these limits, and say how strong it is.

Where does Honestly fit?

Honestly is a proprietary dataset of product opinions, built from many data streams. They include videos and what is said in them, posts, comments, replies, reviews and forums. Most sources are public. The record is ours: each opinion keeps its source link. Most opinions are tied to a product, and many to a specific feature.

Bring a goal or a hypothesis. No questionnaire is needed. Before any survey, we look for the questions people already answer in public. Next, you field a shorter survey for the rest. Then go deeper with people who used the product.

Our published studies are on the research page.

Frequently asked questions

What is preliminary research in consumer insights?

It is the low-cost step before a survey, also called exploratory research. You start from a goal, list the questions and check which ones people already answered in public.

Does preliminary research mean I can skip the survey?

No. It shortens the survey by answering the questions public opinions can answer. Questions about counts, prices and new ideas still need fielding.

What kinds of questions do public opinions answer best?

Questions about products people already use: why they chose them, what they dislike, what they compare them with and how they use them. They answer poorly when you need a share of all buyers.

Can I test a new product idea from public opinions?

Not directly, because nobody has seen it. You can check the need, the objections and the language before you run a concept test.

How do I know if an answer from public opinions is strong?

Look at how many distinct people gave it, whether sources agree and how specific the opinions are. Grade it strong, some, thin or needs a survey.

Who should run preliminary research?

The insights team, with input from the team that owns the decision. Marketing and product teams can run a light version before they brief a full study.

Bring one product question. See what the dataset already holds on it.

Request a demo