“I will study how small and medium enterprises use digital marketing” sounds like a scoped thesis until a committee asks three questions: which size band counts as an SME in your context, how will you actually reach enough owners to say anything generalizable, and where does your data come from if you cannot survey every SME in existence. This guide works through the scoping decisions an SME thesis needs before data collection starts: what “SME” means precisely, the access strategy that works with small-business owners specifically, the national data available, and a worked example of turning a vague SME topic into a researchable one.

What counts as an SME — pick a definition and state it

The U.S. Small Business Administration (SBA) defines “small business” through size standards that vary by industry: a size standard is “usually stated in number of employees or average annual receipts” and represents the largest a business (including subsidiaries and affiliates) may be while still counting as small for federal programs. Critically, there is no single universal threshold — the standard is set per NAICS industry code, so a “small” manufacturer and a “small” retailer are defined by different numeric cutoffs. A thesis that says “SMEs have fewer than 500 employees” without naming which standard or industry that number comes from is using an imprecise shorthand; a thesis that says “this study defines an SME per SBA size standards for NAICS code 722511 (full-service restaurants)” has made a defensible, checkable choice. Outside the U.S., the definition changes again — the EU’s SME definition uses staff headcount and either turnover or balance-sheet total, with different thresholds than the SBA’s — so an internationally-scoped thesis needs to state which jurisdiction’s definition it is using and why. “Medium” enterprises also need their own upper bound stated explicitly, since a thesis that lumps a five-person shop and a 400-employee manufacturer into one “SME” category without acknowledging the management-practice differences between them will draw a challenge on construct validity grounds.

SME research is not entrepreneurship research

The two literatures overlap but ask different questions. Entrepreneurship research typically studies the founding and early-growth process — opportunity recognition, startup financing, the transition from idea to viable firm. SME research studies established small and medium businesses, which may be decades old, family-owned, or multi-generational, and asks different questions: management practice adoption, succession planning, access to credit, response to regulatory change, digital-transformation uptake. Framing an SME thesis using entrepreneurship theory and literature when the actual population is established businesses is a common scoping error that produces a mismatch between the theoretical framework and the empirical setting. If your population genuinely spans both — a mixed sample of new ventures and established firms — say so explicitly and justify why a single framework applies to both, rather than defaulting to whichever literature you found first.

Access: why SME owners are a different gatekeeper problem than large firms

Recruiting SME owners for research differs structurally from recruiting participants inside a large organization:

• The owner is usually the sole gatekeeper. There is no HR department or research-liaison office to route your request through — you are asking the person who runs daily operations to give you their own time, and that person’s calendar has no slack built in for research participation the way a large firm’s employee survey often does.
• Competitive sensitivity is higher, not lower. A large corporation has a communications department managing what information leaves the building; an SME owner has no equivalent and may worry, reasonably, that operational details shared with a researcher could reach a local competitor. Being explicit about confidentiality, data storage and who will see raw responses is more important here than in most organizational research.
• Trade associations and chambers of commerce are often the highest-yield access route. Rather than cold-contacting individual owners, working through a local chamber of commerce, an industry trade association, or a small-business development center that already has owners’ trust tends to improve response rates compared to direct outreach.
• Snowball sampling works better here than in most organizational research. Small-business owner communities are often tightly networked locally; one cooperative owner introducing you to two or three peers frequently outperforms a cold-outreach list of the same size, because the referral carries social trust the researcher cannot manufacture alone.

Confidentiality and ethics for competitively sensitive data

An SME thesis handling operational or financial detail needs an ethics plan that goes beyond standard participant confidentiality. Consider stating explicitly: whether raw financial figures will be reported at all or only as ranges or ratios; whether the business name and specific location will be masked even in a single-case design, where masking is harder because the business may be identifiable from context; and who beyond the researcher and supervisor will see unmasked data. Owners are considerably more likely to agree to participate, and to answer candidly, when this plan is stated up front in plain language rather than buried in a standard consent template — a specific, short explanation of what will and will not be shared tends to outperform a generic confidentiality clause copied from a template built for individual human-subjects research.

Where the national SME data lives

Several U.S. federal sources provide aggregate SME data suitable for a secondary-data or mixed design: the Census Bureau’s Statistics of U.S. Businesses (SUSB) provides firm and establishment counts, employment and payroll by enterprise size and industry; the Small Business Administration’s Office of Advocacy publishes small business profiles by state; and the Federal Reserve’s Small Business Credit Survey covers financing access and use, an increasingly common thesis angle. None of these substitute for firm-level primary data if your question is about management practice inside specific businesses rather than population-level patterns — but they are the right starting point for framing how representative your primary sample is relative to the broader SME population. Our directory of free datasets and open data repositories lists further public sources.

Cross-sectional versus longitudinal design for SME research

Most thesis-length SME studies are cross-sectional out of necessity — a single snapshot of owner interviews or survey responses within the project’s timeline. This is a legitimate and common design choice, but state it as a choice rather than a default, and name what a longitudinal design would have added (tracking the same firms through a specific event — a regulatory change, a funding cycle, a technology adoption decision — rather than cross-sectional variation across different firms at one point in time). Where the timeline genuinely allows two data-collection waves, even a short panel (baseline and one follow-up six months later) substantially strengthens causal claims about a practice change’s effect, and is worth proposing explicitly if your access channel can support a second contact with the same owners.

A worked example

Weak framing: “This study examines how small businesses use technology.”

Stronger framing: “This study examines the adoption of point-of-sale data analytics among independently owned full-service restaurants (SBA size standard, NAICS 722511) in one U.S. state, with a target sample of 20 owner interviews recruited through the state restaurant association, supplemented by SUSB establishment counts to assess sample representativeness against the full population of eligible restaurants in the state.”

The second version names the size standard and industry precisely, states the recruitment channel, gives a target sample size, and names the secondary source that will let the thesis discuss how representative the sample is — everything a committee needs to judge feasibility.

Design notes specific to SME research

Because SME owners are difficult to recruit in large numbers, qualitative and mixed-methods designs are common in this literature, and a purely quantitative design with a large-N survey ambition often proves infeasible within a thesis timeline. Where a survey design is used, a shorter instrument with a strong access-channel partnership (distributed through a trade association’s own newsletter, for example) typically outperforms a longer instrument sent cold. State your realistic sample-size target early and build the design around what is actually achievable through your access channel, rather than designing an ideal study and discovering the access problem after ethics approval. For the operationalisation and instrument-design layer once your population is scoped, our guide to theoretical frameworks for a business dissertation covers the theories (including resource-based view and institutional theory, both common in SME research) that pair well with a small-business setting, and our 60 business and management dissertation topics guide includes several SME-adjacent operations and finance angles worth comparing against your own scope. For the full chapter structure this scope feeds into, see how to write a business management dissertation.

Mistakes that get an SME proposal rescoped

• No size definition stated — “small businesses” or “SMEs” without naming the specific standard, jurisdiction or NAICS code used.
• Entrepreneurship framing applied to an established-business population — a theoretical mismatch between the literature cited and the actual firms studied.
• No realistic access plan — a target sample size with no named recruitment channel (trade association, chamber of commerce, personal network) capable of reaching it.
• Confidentiality left implicit — no explicit plan for protecting competitively sensitive information an SME owner discloses.
• Sample representativeness never addressed — a small qualitative sample presented as representative of “SMEs” broadly without reference to population-level data (SUSB or similar) to contextualize it.
• “Small” and “medium” collapsed into one undifferentiated band — no upper bound stated for what counts as medium, or the two treated as interchangeable despite real management-practice differences.

Frequently asked questions

How does the SBA define a small business?

The SBA uses size standards that vary by industry (NAICS code), stated as either average annual receipts or number of employees. There is no single universal threshold — the definition of small varies by industry.

Is an SME thesis the same as an entrepreneurship thesis?

No. Entrepreneurship research typically studies the founding and early-growth process; SME research studies established small and medium businesses, which may be decades old, and asks different questions about survival, management practice, and access to resources.

Why is SME access harder than studying large firms?

Small business owners are typically the sole gatekeeper for both access and data, are more time-constrained, and are more concerned about competitive information leaking than a large firm’s corporate communications department.

Can I study SMEs using only secondary data?

Yes for aggregate or population-level questions using Census Bureau or state-level small-business data, but firm-level management-practice questions usually require primary data collection since that level of detail is not in public secondary sources.

How many SME owners do I need for a qualitative study?

There is no fixed number; the right number depends on how homogeneous your sample is and how narrowly you have scoped the research question, so justify your target and report how you judged that you had enough interviews.

Should I use snowball sampling for an SME study?

It often works well, since small-business owner communities are tightly networked locally and a referral from a trusted peer carries more weight than a cold approach. Combine it with a named starting channel (a trade association or chamber of commerce) rather than relying on snowball sampling alone from a single initial contact.

Is a cross-sectional design acceptable for an SME thesis?

Yes, and it is the most common design given typical thesis timelines. State it as a deliberate choice, and name what a longitudinal design would add, rather than presenting a single snapshot as though it captured change over time.

Scope your SME thesis with a defensible size definition and access plan

Tesify’s thesis workspace, used by 9,000+ students and 15,000+ chapters, helps you structure an SME thesis around a named size standard, a realistic access channel and population-level context. Every word stays 100% written by you.