Audience research
Researching B2B buyers with account frames, interviews and surveys that hold up
Research B2B social selling audiences in England with defensible account frames, buyer evidence, interviews, surveys, privacy controls and decision logs.
B2B social selling audience research should identify which organisations can benefit, who participates in their buying decisions, what problem starts a search and which evidence earns a conversation. It should not begin with a platform's estimate of reachable members. England has useful business statistics and public company records, but no official dataset counts people ready to buy through social networks.
A defensible study therefore joins several kinds of evidence. Official data defines the business frame. Account records verify individual organisations. Interviews explain recent decisions. A survey tests the distribution of a tightly worded finding. Platform analytics describe only the activity the platform says it measured.
What to take away
- Start research from a commercial decision, not a platform's estimate of reachable members.
- Choose one account frame, such as ONS local units or DBT population estimates, and write down its exclusions.
- Map work and influence from seller records, then treat each role proposition as needing a source.
- Recruit participants from a defined experience, because recruitment routes create bias.
- Use surveys only after the population, question and response frame are defined.
Write the decision before the audience label
Name the commercial choice that research must change. It may be whether to focus on facilities teams at multi-site manufacturers, whether technical proof should appear before a meeting request, or whether a regional campaign deserves another research round. "Understand our audience" is too loose because almost any observation can appear relevant.
Fix the unit at the same time:
Definitions
- An account is a buying organisation.
- A local unit is a workplace or operating site.
- A participant is one person speaking from a particular role and experience.
- A social response is an action recorded on a platform, not proof of purchasing authority.
For England, state whether location means a head office, the site receiving the service, the participant's workplace or another operational field. These categories often produce different account lists. A London head office can manage sites elsewhere, while an English site may belong to a company registered outside England.
Build an account frame from appropriate records
The Office for National Statistics UK business activity, size and location dataset reports enterprises and local units from a snapshot of the Inter-Departmental Business Register taken on 14 March 2025. Its tables support cuts by industry, employment band and geography. They cover registered organisations in the relevant frame, not a ready-made list of prospective buyers.
The Department for Business and Trade takes a broader approach in its 2025 business population estimates. It combines registered-business data with estimates for unregistered private-sector businesses and assigns regional counts by head office. This is useful when very small or non-employing firms belong in scope. It is unsuitable when the offer requires a staffed English operating site or a defined purchasing team.
Choose one frame and write down its exclusions. For example, a seller of maintenance software might select English local units in two industrial groups and an employment band consistent with on-site operations. That is a hypothetical research rule, not evidence that every selected unit needs the software.
Use the Companies House search service to verify a company's registered name, status, officers, filing history and stated nature of business. A filing is a legal or administrative record. It may be out of date for the sales question and does not reveal an informal buying group. Sole traders and some other organisations will not appear in a company-only frame.
Map work and influence, not just job titles
A business purchase can involve the person experiencing the problem, a manager accountable for the outcome, someone controlling money and colleagues who review security, procurement or implementation. Research which roles exist for the actual offer. Do not assume the most senior person is the user or that one title has the same remit across company sizes.
Start with events already recorded by the seller: won and lost opportunities, support conversations, tender documents, implementation notes and objections. Keep the source date and account context. A procurement document may name mandatory evidence, while an interview can explain why a practical user asked for it.
LinkedIn's current Sales Navigator filter definitions include fields such as current company, function, seniority, title, geography, headquarters, industry and company headcount. Its separate Personas documentation describes product filters based on function, seniority, job title and geography. These records establish what the supplier says the product can filter. They do not confirm that a profile is current, that a title carries a particular decision right or that the person wants contact.
Treat an initial role map as a set of propositions. A card might say, "Operations lead: reports breakdown cost, looks for implementation evidence, may introduce finance." Each clause needs a source or an assumption label. Replace it when interviews or opportunity records disagree.
Recruit people connected to the decision
The Government Digital Service's guide to finding user-research participants recommends recruiting actual or likely users and warns that recruitment routes create bias. Applied to business research, the principle is to recruit from a defined experience rather than from whoever comments most often.
Useful contrasts may include:
- buyers who accepted a meeting and comparable buyers who declined;
- day-to-day users and colleagues who approved or blocked the purchase;
- new accounts and established customers facing the same task;
- organisations inside the English operating rule and near-matches excluded by it.
Recruitment through a seller's followers will over-represent visible, platform-active people. A customer list excludes non-buyers. A professional body may reach a relevant occupation but not the complete market. Record the route, invitation count, refusals, eligibility failures and any incentive. Do not call a small qualitative sample representative of England.
Interview a recent episode
Ask about the last relevant decision, not general attitudes towards professional networking. The GDS page on using in-depth interviews frames interviews as a way to learn about circumstances, problems and behaviour from suitable participants. For this topic, reconstruct what changed at the account, where the participant looked, whose view mattered, what evidence was missing and what happened next.
Delay examples until after unaided recall. "Did you see our post?" encourages agreement and makes the seller central to an episode that may have started elsewhere. "Where did you first look for an answer?" leaves room for a colleague, search engine, tender portal, social post or existing supplier.
Separate the record after each session:
- Observationwhat the participant said or did.
- Interpretationwhat the researcher thinks it means.
- Unknownwhat needs another source or participant.
- Decision implicationwhat might change if the pattern persists.
The GDS research-analysis guidance makes a similar distinction between observations, findings and actions. Preserve awkward cases rather than editing them to fit the current persona.
Use surveys for a narrow measurement job
A survey is useful after the team can define a population, question and response frame. It can test how often a recently observed barrier appears among invited accounts, compare two clearly specified groups or collect structured follow-up after a sales process. An open link posted to followers cannot estimate prevalence among English buyers because the denominator and selection process are unknown.
Before choosing software, decide whether the study needs unique invitations, branching, anonymous responses, exports, collaboration controls or a link to a known account record. Test the questionnaire with people who understand the buying task but were not involved in writing it. Remove double questions and answer choices that assume the conclusion.
Keep platform convenience separate from research validity. Google, Microsoft, SurveyMonkey and Jisc publish different feature records, but none can repair an undefined sample or a leading question. Product access, terms and controls may also vary by account or plan, so verify the selected edition immediately before fieldwork.
Read social analytics within their limits
LinkedIn's audience analytics record says members can review follower trends and top demographics such as title, location, industry, seniority, company size and company. It also says organisational demographics are excluded and the displayed numbers are estimates that may not be precise.
These figures can reveal whether content is reaching a different group from the intended one. They cannot count the English addressable market, identify unseen buyers or prove that a follower influenced a purchase. Compare like periods and preserve the exported definition. Where a demographic is withheld or aggregated, record it as unavailable rather than filling the gap with a guess.
Keep research separate from disguised promotion
The Information Commissioner's Office states that genuine market research is not direct marketing. If the message contains promotion, or the exercise collects information so people can later be marketed to, the direct-marketing rules may apply. Calling an outreach message a survey does not settle its purpose.
Participation consent and a data-protection lawful basis are related but distinct checks. GDS guidance on informed consent for user research covers the research purpose, data collected, use, sharing, recording, retention and withdrawal. The ICO's data-minimisation guidance requires personal information to be adequate, relevant and limited to the stated purpose.
Have a qualified UK privacy reviewer confirm the lawful basis, information notice, processor arrangements, retention and any marketing follow-up. Delete unused recruitment details on the approved timetable. Avoid putting sensitive interview notes into a sales platform merely because the participant already has a public profile.
Turn findings into a controlled choice
Finish with an evidence ledger, not a colourful composite character. For each proposed audience, record the account rule, relevant buying roles, observed trigger, information need, likely research route, contradictory evidence and confidence. Add a review date.
Then make one reversible decision. The team might interview a second role at six existing accounts, revise a technical proof page, or exclude microbusinesses until the offer can serve them. Those are recommendations based on the evidence assembled. They are not observed outcomes.
Do not scale when the English location rule is ambiguous, recruitment comes from one unusually engaged channel, role claims rest on titles alone or a survey has no defined invitation denominator. Resolve the missing evidence first. The next research round should be chosen because it can overturn an important assumption, not because it produces another dashboard.
Before you act
- Name the commercial choice the research must change.
- State whether location means head office, site or workplace.
- Choose one frame and record its exclusions.
- Label each role-map clause as source or assumption.
- Record recruitment route, refusals and eligibility failures.
- Separate observation, interpretation, unknown and decision implication.
Common questions
Why is a platform's estimate of reachable members a weak starting point?
The article says audience research should not begin with a platform's estimate of reachable members. Platform analytics describe only the activity the platform says it measured. No official dataset counts people ready to buy through social networks, so reach does not establish purchasing authority or intent.
When is the DBT business population estimate unsuitable?
The article says the DBT 2025 estimates combine registered-business data with estimates for unregistered private-sector businesses and assign regional counts by head office. It calls this unsuitable when the offer requires a staffed English operating site or a defined purchasing team.
What can Companies House records not tell you?
The article says a Companies House filing is a legal or administrative record. It may be out of date for the sales question and does not reveal an informal buying group. Sole traders and some other organisations will not appear in a company-only frame.
In this guide
- Write testable triggers for a B2B social selling target audienceWrite testable triggers for a B2B social selling audience in England, each with a source date, a verification route and only the personal data needed.
- A B2B buyer persona is a hypothesis with receipts, not a stereotypeBuild evidence-based B2B buyer personas for England around buying work, triggers and proof needs, without turning platform profiles into stereotypes.
- B2B social selling customer interviews: seven questions that rebuild a buying episodeUse seven evidence-led customer interview questions to reconstruct B2B buying episodes in England without disguising sales contact as research.
- Choosing a survey tool for B2B audience research by governance before featuresCompare four survey tools for B2B audience research in England by invitations, logic, exports and governance, using dated first-party records.
- Researching B2B social selling competitors with an evidence ledger, not adjectivesResearch B2B social-selling competitors in England with a dated evidence log covering entities, offers, public posts, adverts and contracts.
