Desk research: finding and evaluating sources you can rely on (Research and Evidence for Better Decisions) | READOS

Before running a survey, commissioning a study or interviewing customers, it is worth asking whether the answer already exists. Often it does, at least in part. Official statistics describe industries, regions and households. Industry associations publish surveys of their members. Government agencies report on markets, regulations and programs. Companies publish annual reports, product information and announcements. Patent registers show what competitors are developing. And a business’s own records, such as sales, quotes, complaints, quality data, maintenance logs and website analytics, often hold answers that nobody has looked for.

Desk research, also called secondary research, gathers and analyses this existing information. It is usually faster and cheaper than collecting new data, and it should normally come first: it may answer the question outright, narrow it to the parts that need primary research, or show what others have already learned. But the ease of finding information is matched by the difficulty of judging it. Search engines return confident answers of very uneven quality, statistics are quoted without definitions or dates, reports are written to sell, and automated summaries can invent sources that do not exist.

This article explains where to look, how to search efficiently, how to evaluate the credibility and relevance of sources, how to check statistics, how to use automated tools safely, how to record and synthesise findings and how to recognise the limits of desk research. It is part of a series on research and evidence for business decisions.

Start inside the business

Internal data is often the richest and most relevant source, because it describes the business’s own customers, products and processes:

Sales and quote records: who buys what, when, at what price, and which quotes are won or lost.
Customer records and correspondence: enquiries, complaints, compliments, returns and warranty claims.
Operational data: production times, quality results, scrap, downtime and maintenance history.
Financial data: margins by product, customer and job, and costs over time.
Website and digital analytics: what people search for, which pages they view and where they leave.
Staff knowledge: what sales, service and production staff hear and see every day.

Internal data has limits: it describes existing customers and current practices, not people who chose competitors or markets the business has not entered. But it is the natural starting point.

External sources

Useful external sources for Australian businesses include:

Official statistics: the Australian Bureau of Statistics publishes data on population, households, businesses, industries, employment, prices, trade and much more, including counts of businesses by industry, size and location.
Government agencies and departments: reports and data on industries, regions, energy, health, transport and regulation.
Regulators: guidance, enforcement information and industry data.
Industry associations and peak bodies: member surveys, benchmarking reports and industry outlooks.
Standards and technical bodies: standards, guidance documents and technical reports.
Company information: annual reports, investor presentations and announcements from listed companies, and websites and catalogues of competitors and suppliers.
Patent and trade mark registers: IP Australia’s searchable databases and international patent databases show what is being protected and by whom.
Tender and contract records: published government tenders and contract notices reveal what buyers purchase and at what scale.
Academic and research publications: journal articles, theses and research reports, many accessible through scholarly search tools.
Trade press and specialist media: news and analysis for particular industries.
Reviews and public conversations: customer reviews, forums and social media, used carefully.

The expanding into a new region article shows how several of these sources combine when assessing a new location.

Searching efficiently

Search engines and databases reward precise queries:

Use specific terms, including technical names, product categories and Australian spelling.
Use quotation marks for exact phrases.
Limit by site, for example to government or university domains, or to a particular organisation’s website.
Limit by file type, such as PDF, to find reports and papers.
Limit by date to find current information.
Follow references: a good report’s references lead to its sources, and newer work citing it leads forward in time.

Keeping a search log, a simple record of where and how you searched, prevents repeating work and makes the research reproducible.

Evaluating sources

Not all sources deserve equal trust. Useful questions for any source are:

Currency: when was it published, and is it still current? Statistics from several years ago may describe a different market.
Relevance: does it address the question, for the right industry, location, size of business and time period?
Authority: who produced it, with what expertise, and are they accountable for its accuracy?
Accuracy and evidence: does it explain its methods, cite sources and show data, or simply assert conclusions?
Purpose: was it written to inform, to persuade or to sell? Vendor white papers and promotional surveys may be useful but are not neutral.

Lateral reading, checking what other credible sources say about a source and its claims, is often quicker and more reliable than studying the source itself in depth.

Primary, secondary and tertiary sources

A primary source is the original: the statistics release, the study, the standard or the company’s own announcement. A secondary source reports or analyses primary sources, such as a news article or industry review. A tertiary source summarises secondary sources, such as an encyclopaedia or an overview article. Each step away from the original adds the possibility of error, simplification or distortion. When a statistic or claim matters, trace it back to its primary source.

Checking statistics

Statistics are easily misused. Before relying on a number, check:

Definitions: what exactly is counted? “Small business” and “manufacturing” are defined differently in different sources.
Base and denominator: a percentage of what? A large percentage of a small group may be a small number.
Time and place: which year and which geography?
Method: is it a census, a survey sample, an estimate or a model? How large was the sample and how was it chosen?
Uncertainty: are margins of error or confidence intervals given?
Whether it has been misquoted: compare with the original.

The testing the numbers you are given article explains practical checks for figures that arrive in proposals and reports.

Automated search and summaries

Search engines and AI tools can summarise large amounts of information quickly. They can also be wrong with complete confidence, combine unrelated facts, present outdated information as current and, in some cases, invent references, quotations or statistics. Automated summaries are useful for orientation and for suggesting where to look, but every fact that matters should be verified in an original source, and any reference should be opened and read before it is cited.

Copyright and use of information

Finding information is not the same as being free to reuse it. Reports, articles, images and data may be protected by copyright or licence terms. Summarising and citing sources for internal decision-making is generally different from republishing them. When research will be shared outside the business or used in marketing, check licence terms, quote briefly, attribute sources and paraphrase in your own words.

Benchmarking against others

Desk research is often used to compare a business with others: prices, margins, productivity, wages, energy use or quality levels. Industry benchmarking reports, association surveys and published accounts of listed companies provide reference points. Benchmarks are only useful when the comparison is like with like: similar size, product mix, location and accounting definitions. A margin that looks low against a national average may be normal for a business with a different mix of products or customers. Using benchmarks as prompts for questions, rather than as verdicts, keeps them useful.

Reviews and online conversations

Customer reviews, forums and social media show what people say about products, services and competitors in their own words. They can reveal frustrations, unmet needs and the language customers use, which is valuable for product design and marketing. But they are not a representative sample: people with very good or very bad experiences are more likely to post, some reviews are fake or incentivised and loud voices can dominate. Treat online conversations as a source of themes and hypotheses to test, not as measurements of how common an opinion is.

Keeping desk research current

Markets, regulations and technologies change. Desk research used for a major decision should be dated, and important figures should be refreshed before later decisions rely on them. Setting a simple review date for key market information, and noting when new official statistics are due, keeps a business from planning on numbers that are several years old.

Recording what you find

A simple source log saves time and protects credibility. For each source, record the title, author or organisation, date, link or location, key findings relevant to the question, and any concerns about quality. A synthesis table, with sources as rows and the key questions as columns, shows quickly where sources agree, disagree or are silent. Gaps in the table are the questions that may need primary research.

Synthesising findings

Synthesis means combining evidence into conclusions, not simply listing what each source says. Look for patterns across sources, explain disagreements, weigh sources by their quality and relevance, and state clearly how confident the conclusions are. Where evidence is thin or conflicting, say so.

Limits of desk research

Desk research describes the past and the general case. It may not reflect a specific market, a new product or recent changes. It often lacks information about customers’ reasons and preferences. And it depends on the quality of others’ work. Recognising these limits shows where primary research, such as interviews, surveys or trials, is needed. The market research on a small-business budget article describes how desk and primary research fit together.

A worked example

This is an illustrative example. The business, sources and figures are hypothetical.

A manufacturer of workshop dust extraction units wants to estimate the potential market among small joinery and cabinet-making businesses in its state. It starts with desk research.

Official counts of businesses by industry provide the number of businesses in the relevant industry classes in the state, broken down by employee size. Industry association reports describe typical workshop equipment and replacement cycles. Regulator guidance on wood dust indicates the controls businesses are expected to have. The company’s own sales records show which sizes of business have bought its units in the past and how often.

Combining these sources, the company estimates that roughly 2,000 businesses fall into its target size range, that perhaps one in eight might replace or add extraction equipment in a given year, and that its typical sale is about $8,000. That suggests an annual market of roughly 250 sales, or about $2 million, in the state. The source log records each figure’s origin and date, and the synthesis table shows that replacement rates are the least certain input. The company commissions a short phone survey of 60 workshops to test that figure before investing in a sales campaign.

How this applies to a small Australian business

Search internal records first, then external sources.
Trace important statistics to their primary source.
Judge sources for currency, relevance, authority, evidence and purpose.
Verify automated summaries against originals.
Keep a source log and a synthesis table.
Compare like with like when benchmarking against other businesses.
Treat reviews and forums as themes to test, not measurements.
Date key market figures and refresh them before major decisions.
Use desk research to focus primary research on the gaps.

Common mistakes

Starting with a survey when the answer already exists.
Quoting statistics without definitions, dates or sources.
Relying on vendor reports as neutral evidence.
Citing references that were never opened.
Mistaking a long list of sources for a synthesis.
Reusing copyrighted material without permission.
Benchmarking against businesses with a different size, mix or location without adjusting.
Ignoring internal data because it is hard to extract.

Frequently asked questions

Is information on government websites always reliable? It is usually authoritative for its own data and responsibilities, but it can be out of date or defined differently from your question. Check dates and definitions.

How can unreliable websites be recognised? Look for missing authors and dates, no cited sources, a commercial motive, claims that contradict credible sources and statistics with no explanation of method.

When should desk research stop? When further sources add little new information, when remaining questions can only be answered by primary research or when the decision deadline arrives.

Can competitor information be used? Publicly available information from websites, catalogues, annual reports, registers and published tenders can be used for research. Confidential information obtained improperly should never be used.

How should conflicting sources be handled? Compare their definitions, dates, methods and purposes, which often explain the difference. Where a real disagreement remains, report the range and the reasons, and give more weight to transparent, independent and recent sources.

Questions to ask

What do our own records already tell us about this question?
Which external sources are authoritative for our industry and region?
Have we traced our key figures to their primary sources?
How confident are we in each source, and why?
Which benchmarks do we use, and are they truly comparable?
When were our key market figures last refreshed?
What gaps remain that need primary research?

Bringing it together

Desk research gathers existing information from internal records and external sources such as official statistics, government and industry reports, registers, company publications and academic work. Efficient searching, careful evaluation for currency, relevance, authority, evidence and purpose, tracing figures to primary sources and verifying automated summaries keep it reliable. A source log and synthesis table turn scattered information into conclusions and show where primary research is needed.

Source: KEVOS editorial notes. Examples and figures in this article are illustrations.