How to Do Market Research That Improves Marketing Decisions

You can finish a market research presentation and still have no idea what to do next.

You know the industry is growing. You have screenshots of competitors, a few interesting survey results and a slide explaining that your audience values quality.

But should you change the offer? Target a different buyer? Explain the price differently? Invest in a new channel?

The presentation does not answer those questions.

When I think about how to do market research, I start here: what decision needs better evidence? That question makes the work more focused. It also makes the findings harder to ignore.

This guide uses a fictional training provider in Amman that wants to sell business English courses to employers. Every business detail, number and finding in that example is illustrative, not a client result or a claim about the Jordanian market.

What is market research supposed to answer?

Start with the decision your business needs to make. Then write the research question beneath it.

“Understand the training market” is a subject. It is too broad to guide the work.

“Should we lead our corporate course offer with presentation skills or general language improvement?” is a decision. It suggests the people to speak to and the evidence to collect.

For our fictional provider, the starting point could be:

Write down your current belief as well: “We think employers want presentation training.” Treat it as something to investigate, not something the research must prove.

Decide in advance what would change your mind. For example, if employers consistently describe customer calls as the urgent problem, you would reconsider a presentation-led offer.

How do you define the market and the people to study?

“Working professionals in Jordan” includes too many different situations to be a useful starting audience.

Narrow the market by geography, use case, organisation type, buying process and the service you can actually deliver. Our provider might begin with Amman-based service businesses whose employees regularly communicate with international clients.

Then separate three roles:

• The user: the employee attending the course.
• The buyer: the person evaluating training providers.
• The approver: the person authorising the budget.

One person may hold several roles. Do not assume they want the same thing. An employee might want confidence during calls; a manager might want fewer misunderstandings; a finance lead might want a clear scope and manageable commitment.

Recruit beyond your easiest contacts. Include recent buyers, people who chose another option and people who considered training but postponed it. Include more than one organisation type if your intended market contains meaningful differences.

Record where participants came from and who is missing. Existing customers can explain their experience with you. They cannot, by themselves, explain everyone who has never considered you.

Ask permission before recording interviews. Explain the purpose of the research, avoid collecting unnecessary personal information and remove identifying details from shared notes unless permission has been given.

Where should you look before commissioning new research?

Start with evidence that already exists.

Your business may have proposal feedback, lost-enquiry notes, sales conversations, support requests and records of what customers actually bought. Outside the business, look for relevant official statistics, sector reports and public information about alternatives.

The U.S. Small Business Administration distinguishes existing research from research conducted directly with customers. Its guidance includes demand, market size, location and the prices of alternatives as useful areas to investigate.[1]

For every external figure, record its source, date, geography, definition and method. A global report about online learning does not automatically describe employer-funded training in Amman.

Reviews can suggest questions worth investigating. Search behaviour can suggest the language people use. Neither should be treated as a complete picture of the market.

Also separate what your analytics can show from what they cannot. A visitor leaving a course page is an observed event. “They thought it was expensive” is an interpretation until you have evidence about why they left.

Should you use interviews, surveys or observation?

Choose the method after you have defined the question.

For an early question, I would often start with interviews, use what they reveal to sharpen the language, and then decide whether a survey or a test is needed. That is a practical sequence, not a rule for every project.

There is no universal number of interviews that makes research “solid”. A few conversations can expose a problem worth investigating; they cannot establish its prevalence in a whole market.

A large survey shared only with your followers can still leave you with a narrow view. More responses do not automatically correct the way people were selected.

What should you ask customers?

Ask about a real situation before asking for an opinion about your idea.

“Would your team benefit from our practical, high-quality training?” makes agreement easy and tells you little about a purchase.

Instead, ask:

1. Tell me about the last time your team needed help communicating in English at work.
2. What happened that made someone consider training?
3. What did you do first, and which alternatives did you consider?
4. Who took part in the decision, and what mattered to each person?
5. What made you choose an option—or decide to do nothing?
6. What did you need to see before approving the spending?
7. Afterward, how did you decide whether it had been useful?

Follow with “Can you give me an example?” or “What happened next?” Do not rush to fill the silence with your explanation.

For surveys, use clear, neutral wording, ask one thing at a time and test the questions before distributing them. Pew Research Center’s guidance explains why wording, order and pretesting matter.[2]

A question such as “How affordable and useful was the course?” combines two different judgements. Split it. Someone may find a course useful and unaffordable.

If you research in Arabic and English, check that both versions ask the same thing naturally. Test understanding with speakers of each language rather than assuming a literal translation preserves the meaning.

Avoid pitching during a research interview. The moment you begin defending the offer, you make it harder for someone to tell you why they would not buy it.

How do you research competitors and estimate demand?

Compare the choices customers actually have

A competitor review should explain the buying decision, not just compare follower counts and colour palettes.

For our provider, the alternatives could include another training company, a private tutor, an online course, internal coaching or postponing training altogether.

For each relevant option, record:

• The audience and problem it appears to prioritise.
• The offer, format, commitment and publicly available price.
• The promise and the evidence offered to support it.
• The steps and friction involved in making an enquiry.
• What customers say they value or struggle with, clearly attributed to the source.

Distinguish a competitor’s claim from verified performance. A website saying “industry-leading results” does not prove the results. A visible advertising campaign does not tell you whether it is profitable.

For the training provider, the useful comparison might be whether an employer can understand the schedule, see a sample session and judge relevance to the team’s daily work. Those details could matter more than how polished the competitor’s Instagram looks.

Build a demand estimate you can explain

A large industry figure is not your sales opportunity.

Start with the customers who fit your definition, then examine need, budget, purchasing frequency, access and your delivery capacity.

Here is a deliberately simplified example, using invented assumptions:

• Suppose you identify 100 eligible organisations.
• Assume each could purchase one JOD 1,500 programme per year.
• That produces a JOD 150,000 annual spending scenario, only if every organisation buys on those terms.

It is not a revenue forecast. You still need evidence about eligibility, budgets, frequency and willingness to buy. You also need to consider how many organisations you can reach and serve, and how many may choose someone else or make no purchase.

If you can initially reach only 20 of those organisations, analyse that smaller group separately. Use clearly labelled scenarios, show where assumptions came from and update them as evidence improves.

How do you turn findings into decisions?

Separate the observation, your interpretation and the action you want to test.

That prevents a tentative explanation from quietly becoming a fact.

For our fictional provider, a working analysis might look like this:

These are hypothetical findings, not conclusions about employers in Amman.

Look for agreement and disagreement across sources. Interview comments, enquiry records and a message test can provide different kinds of evidence about the same issue. If they conflict, investigate the difference instead of choosing whichever supports your preferred idea.

Preserve differences between segments. A pattern among small owner-led firms may not describe organisations with formal procurement teams.

In a small interview study, write “seven of the 12 people interviewed mentioned scheduling” if that is what your records show. Do not turn it into “58% of the market has a scheduling problem”. The wider claim requires evidence your interview study may not provide.

What belongs in the final report?

Make the report usable by the person who must act on it.

Include:

• The decision: what the research was intended to inform.
• The scope and method: who or what was studied, when and how.
• The findings: the most relevant evidence, with sources or anonymised examples.
• The limits: who was excluded, what remains uncertain and where assumptions were used.
• The implications: what you recommend changing, keeping or investigating further.
• The next test: an owner, a measure of success and a review point.

For our provider, success might be more qualified employer enquiries, not simply more clicks. Define what “qualified” means before comparing results: for example, an organisation that fits the target, has an identified need and is considering training within the relevant period.

AI can help organise notes, suggest questions and identify themes for you to verify. Do not treat generated customer opinions as interview evidence, or upload identifiable participant information without appropriate permission and safeguards.

The strongest research report leaves room for uncertainty while making the next decision clearer.

You should be able to finish this sentence:

“Because we learned this, we will change this—and check whether it works by measuring this.”

If you cannot, you may need another conversation, a sharper question or a better connection between the findings and the decision. You probably do not need another decorative slide.

Explore related positioning and communications work: Helpdesk

Suhail Darwazeh

Brand & Communications Strategist — Based in Amman. I help organisations clarify their value and connect strategy to practical communication.