Primary vs Secondary Research: Differences & Examples
Primary research collects new data; secondary research uses existing evidence. Compare methods, costs and examples, then choose the missing answer to collect.
Primary research collects new evidence for your question. Secondary research uses evidence that already exists, including business records and published studies. Start with what is available, check whether it fits the audience and decision and commission new work for a missing answer that could change the plan. The distinction helps a market research team decide which gaps need new collection.
Once existing sources have narrowed your options, discuss a managed comparison with Sapien and see our comparisons with human research.
Agree who needs the expansion recommendation and by when. Check which existing sources answer the question, what important gap remains and whether new work can close it before the meeting. Spend the budget on an answer that could change the plan.
Primary and secondary research compared
Purchase question | The difference that matters |
|---|---|
What are we buying? | Primary research collects new evidence for the current question. Secondary research analyzes existing material, whether it comes from your own business or another source. |
What can we control? | A new study lets you specify the audience, questions and collection period. Existing evidence comes with definitions and coverage you must inspect. Neither route guarantees a useful answer. |
What is the main advantage? | New research can investigate the missing experience or compare your exact offers. Existing sources can establish context and narrow the work before you pay to collect more. |
What can go wrong? | A new study can recruit the wrong people or ask a leading question. An existing source may describe another period, geography or population. A polished report doesn't remove either problem. |
What should we budget? | Compare access, preparation, collection, analysis and delivery. Existing material can require a license and substantial analysis; new fieldwork needs a feasible audience and schedule. Get a quote for the question rather than assuming one route is always cheaper. |
Methods and sources for the missing answer
The U.S. Small Business Administration's market research guidance distinguishes existing sources from direct consumer research. For the repair-service decision, use that distinction to commission a specific task.
Route | A useful task in the repair example |
|---|---|
New interviews or focus groups | Explore recent repair experiences and the language people use. An interview can follow one person's episode in depth; a group can explore discussion around a proposition. Selected accounts don't establish how common an obstacle is. |
New survey | Ask consistent questions of a defined recent-buyer audience when the decision needs a measured distribution. Specify eligibility, recruitment and question wording before commissioning it. |
New observation or task | Watch how people find a provider, arrange a repair or use an invitation. Agree the task and what will be recorded; observing a task doesn't by itself explain every choice. |
Government and industry material | Check population, business context and category information. Record the underlying source, period and geography; a broad category estimate isn't evidence of demand for this particular offer. |
Existing business records | Review inquiries, quotations and recorded outcomes. Keep the original definitions and missing states visible; people who never contacted the business won't appear in its inquiry log. |
Published studies and search trends | Inspect existing category research and search patterns for useful context. Check who and what they cover before using them to set the next study's audience. |
Our qualitative research guide helps scope the exploratory work. Our guides to syndicated research and Google Trends explain two existing-source routes and their limits.
Example: researching a repair-service expansion
In this hypothetical repair-service expansion, population data narrows the area, an inquiry log identifies recorded requests and interviews investigate recent repair experiences. Each source must fit the people, period and question.
Your first decision is whether to investigate a particular service area, rather than declare that the entire city “has potential.”
For existing evidence, review official population and business data, provider locations, services, opening hours and your own inquiries. These show context, competing supply and recorded requests. They do not establish that local households need your service or will choose it.
For new evidence, interview recent repair buyers. Ask what failed, whom they contacted, how long it took and why they chose the provider. If the next decision needs to know how common a problem is, measure it among a defined audience afterward.
The two routes meet in a narrower proposition: “Investigate convenient collection for people whose recent repair involved difficult transport.” You now have a specific hypothesis rather than a population number dressed as demand.
Each source below answers part of the repair decision. The missing category-need evidence still requires investigation.
Existing source | What it answers, what to record and what remains missing |
|---|---|
Official local demographic table | Household characteristics in a defined area describe reachable context. Record collection and publication dates. Investigate recent category need and experience separately. |
Competitor website | Listed locations, services and terms identify visible alternatives. Record access date; listings may be incomplete. Ask whether buyers knew about and could use the offers. |
Internal inquiry log | People who contacted the shop provide expressed requests and recorded outcomes. Record the start/end period. The log excludes unseen buyers and may omit refusal reasons. |
The Census Business Builder is one official starting point for U.S. demographic and business information. Check the underlying dataset and geography rather than treating a map as a market-size calculation. For applying those census descriptors to buyer groups, see our demographic segmentation guide.
Keep the question beside each source. Household counts describe context. Recent repair accounts describe category experience. Listings show advertised services, not whether a buyer could use them. Use the sources together while keeping those meanings separate.
The missing evidence is a recent repair experience
For the repair example, use this interview sequence:
- Tell me about the last item you tried to have repaired.
- What made you decide to repair rather than replace it?
- Which providers did you consider, including doing nothing?
- How did you get the item to the provider?
- What happened from first contact to collection?
- Which part was most difficult? What did you do about it?
- What would have made you choose another provider?
Record buying criteria, recruitment and interview dates. Compare completed repairs with abandoned attempts. Was transport the obstacle, or was the repair poor value or the service unusable? Existing customers alone may make the current offer look too attractive.
Ask about the actual repair before showing the collection offer. Otherwise, you may only collect reactions to your preferred solution. Keep scope, collection window and fee visible in any offer comparison.
Desk research has a stopping point
Continue desk research while another source can change the audience, workable area or alternatives. Stop when the next decision depends on an experience none of the sources measured. A newer population table cannot tell you whether proposed collection times suit recent repair buyers.
Keep the source’s unit, geography, collection dates, definitions and coverage with the brief. Publication date is not the date of observation. An inquiry log excludes people who never found the shop. Use these gaps to choose whom the new study should recruit.
Interview people in the reachable area who recently completed or abandoned a repair. Then compare collection and drop-off on otherwise identical terms. Ask for the actual obstacles, exceptions and a workable offer for a booking pilot.
A research brief for the expansion meeting
Use the source register to make the next purchase specific.
Decision and responsibility: Our commercial lead needs a recommendation on whether to pilot collection in the proposed service area. The marketing team or agency will present the evidence and unresolved questions at an agreed decision meeting. The service owner must first confirm feasible collection and drop-off terms, including scope, window and fee.
Audience and work: Review the existing sources, then investigate recent completed and abandoned repair attempts among people in the reachable area. Agree the eligibility window and recruitment route before fieldwork. Quote the exploratory work separately from any later survey or comparison of complete offers. State what our team must supply and when.
Returned output and next check: Return a source register, the actual questions, audience and collection dates, supported findings and important exceptions. Explain which evidence changes the expansion recommendation. Identify any further comparison, cost evidence or booking pilot required before committing to a rollout. Separate simulated responses from collected human accounts and operational results.
AAPOR's survey guidance recommends checking whether existing evidence or another method can answer the question before conducting a survey. If you need a partner to own the design and interpretation, use our market research services guide to compare that responsibility with fieldwork alone.
Questions before commissioning the work
Is a survey primary research if an agency runs it?
Yes, when it collects new evidence for your study. Commissioning the work doesn't turn it into secondary research. Ask the supplier to document the audience, recruitment, questions and collection period so your team can assess what it supports.
Can our customer records replace new research?
They can answer questions they actually recorded. In the repair example, an inquiry log shows recorded requests and outcomes, while interviews can investigate experiences the log never captured. Start with the records, then buy new work only for a gap that could change the decision.
What if the decision deadline is too close for fieldwork?
Narrow the recommendation to what the available evidence supports. Confirm recruitment and delivery feasibility with the supplier. If the missing experience is essential, recommend a smaller reversible pilot or a later commitment rather than presenting context data as measured demand.
Conflicting sources may describe different opportunities
A competitor can advertise collection while a customer reports difficult transport: the customer may be outside the area, unaware of the service or unable to use its hours. Ask which explanation fits before rejecting either source. A large population and weak response to a specific offer can also coexist.
Compare definitions, dates, audience and behavior. The decision depends on usable service terms and actual bookings, so keep those as the pilot questions rather than averaging unlike numbers.
A simulated offer comparison can explore whether a proposed collection window, fee and scope make sense for the intended audience. See how Sapien builds populations and runs a study to assess the audience data and study fit for that task.
Discuss a managed comparison with Sapien when the source review has narrowed the choice to complete collection and drop-off offers. Use the simulated comparison to select the candidate and investigate its objections; actual route cost and completed bookings need operational evidence.
