Reviews
Part of Secondary market research: a practical guide
Which secondary market research questions matter most
Common secondary market research questions for 2027 cover source choice, official-data limits, conflicting estimates, competitor claims, reuse rights, and steps.
What to take away
- Six questions do most of the work. They cover what counts as secondary research, when it runs before interviews, whether an official figure fits, why estimates disagree, what reuse rights apply, and what record makes the call reproducible.
- Fitness beats official status. Check the definition, population, period, geography, unit and method before a number enters the report.
- Do not average incompatible estimates. Reconcile them on definition, period and geography first.
- Read the licence page for the exact item. Write the terms into your notes.
- Close with a decision record. A reviewer should be able to rebuild the conclusion from your sources.
Secondary research beyond the search box
Secondary research is not internet research alone. Existing evidence covers internal records, government statistics, international datasets, filings, academic papers and paid commercial research. Web search is a discovery route, not the method.
Important claims should lead back to an original source and its metadata. Teams that stop at a search result skip the check that decides whether new collection is needed. Read secondary market research for how these source types differ in cost and reliability.
Judging whether an official statistic fits
Official data can be authoritative and still describe the wrong population, period, geography, unit or classification. The Census Bureau's guidance for data users tells you to identify the program and dataset behind a result, then review metadata, geography, concepts and formats. Treat that as a fitness check rather than a guarantee.
Official Data Fitness Check
- Identify program and dataset
- Review metadata and documentation
- Check geography, concepts, formats
- Confirm population and period
- Read methodology and flags
Ordering secondary work before interviews
Published evidence should settle definitions, market structure and what is already known. It should also isolate the gaps that direct collection has to close. Direct collection earns its cost when a consequential question stays unanswered. Skipping the desk stage usually means paying to learn things that were already public.
Reconciling estimates that disagree
Sources measure different products, buyers, channels, geographies, years, currencies and price bases. Forecast-based figures add another layer of assumptions.
Write each operational definition down before comparing numbers. Normalize only values that share that definition. Where the boundaries cannot be reconciled, report scenarios rather than one figure.
A documented sourcing routine makes this faster, and how to research secondary market research covers where to check first.
Example: two estimates that disagree
Two estimates that disagree
- Definition. The larger figure counts subscriptions plus advertising; the smaller counts subscriptions only.
- Period. One is a completed calendar year; the other is a forecast for the year ahead.
- Geography and currency. One covers the United States; the other covers North America from Canadian data and is quoted in Canadian dollars.
Align all three and the gap may narrow enough for the decision, or it may stay wide. When it stays wide, publish a range and name the assumptions behind it. Never average the two raw numbers.
Conflicting sources and single-origin numbers
Compare definitions, populations, methods, periods, revisions and shared inputs. Check whether both sources rest on the same underlying dataset, because repeated syndication is not independent confirmation. Keep the disagreement visible in the note and state how it changes the decision.
Reuse rights for public data
Many U.S. federal works carry no copyright, but the exceptions decide the case. USAGov's government works and copyright page lists logos, trademarks, state and local material and contractor work as separate categories. Read the licence page for the exact table or file, then record what it permits: attribution, redistribution, commercial use and storage.
Canada runs on different terms. The Office of the Privacy Commissioner's PIPEDA guidance applies when a dataset touches identifiable people, which limits what a Canadian file can support. Consent and data rights for new collection belong at the point of collection, a subject primary market research covers in full.
Question-to-evidence check
Check first
- Is the source official?
- Program, method and intended use
- Do estimates agree?
- Definitions, period and shared inputs
- Can it be reused?
- Item-level licence and restrictions
- What comes next?
- The decision gap and its cost
Shortcut that fails
- Is the source official?
- Official means suitable
- Do estimates agree?
- Average every number
- Can it be reused?
- Public means unrestricted
- What comes next?
- Commission a survey by default
The decision record fields
- The decision at stake, and the date it is needed.
- Source, table or series, publisher and access date.
- Definition, population, period, geography and unit.
- Method, coverage, revisions and known limitations.
- Licence, attribution, redistribution and privacy terms.
- Transformation applied, plus residual uncertainty.
- What new evidence would change the conclusion.
Two lines close the record: the answer you are giving now, and the evidence gap that would justify collecting your own data.
Note the date each source was last revised, because a revised series can overturn an earlier conclusion. The record also feeds the wider market research strategy, which shows which questions still need testing.
Common questions
Why do market-size estimates disagree?
They measure different products, buyers, channels, geographies, periods, currencies and price bases, and they assume different growth rates.
Should secondary research come before interviews?
Usually yes. Published evidence should settle definitions and isolate the questions that only direct collection can answer.
Do I need a licence check for a public dataset?
Yes. Public access does not remove copyright, contractual or privacy duties, and the terms vary by item.
Is one source enough?
Fit matters more than count, but any claim that carries the decision should rest on two independent origins.
