Expanding LinkedIn audiences: reach new buyers with a controlled test

Name the bottleneck before expanding

First ask why the existing campaign needs to grow. Limited reach can come from narrow targeting, but also from a bounded market, an unclear offer or insufficient response capacity. More delivery does not automatically solve these constraints. Describe one currently under-reached buyer group and explain why it might suit the same use case. Distinguish lack of access from lack of effect among people already reached. If the current message provides no intelligible value, expansion may merely expose more people to an unclear proposition. The first decision is therefore not which setting to activate, but which justified access hypothesis the team is prepared to examine with accountable evidence.

Distinguish the expansion mechanisms

LinkedIn documents Predictive Audiences as modelled audiences built from suitable source data and describes other audience creation approaches. Check which option is available for the current campaign and how it interacts with existing criteria. Do not use those terms interchangeably. Manually adding a function is a different intervention from selecting a modelled audience. Older lookalike instructions may be obsolete: LinkedIn's official help documents their discontinuation and alternatives. Record the actual mechanism in the test protocol. The team can then reconstruct the question examined, the controls present and the findings that should not be transferred to a different method of expansion simply because both approaches were described internally as “broader targeting”.

Review source data for suitability rather than size

When a modelled audience depends on source data, examine that data's commercial meaning and permitted use. A large list can contain many historically unsuitable contacts. Review origin, recency, duplication, supported markets and the underlying event definition. Do not combine unrelated products or buying situations merely to enlarge the source. Record what genuinely connects appropriate cases and what happens to occur frequently without establishing suitability. Data and platform owners should confirm approved use without copying unnecessary personal details into the strategy document. If the source is poorly understood, technically successful audience construction can reinforce the wrong commercial direction rather than provide a useful route to more relevant buyers.

What is deliberately excluded from the source?

Record unsupported products, outdated relationships and ambiguous events as separate review cases. Remove data only through the agreed procedure. Preserve the business reason for selection without needlessly sharing individual records or expanding access to personal information.

Form an additional-buyer hypothesis

Describe the new group through its task and contribution to buying, not just another job-title word. One role may experience the problem while another owns budget, security or implementation. Expansion needs a message appropriate to the additional role without changing the offer itself. Define exclusions and borderline cases. If another country requires an unsupported language or service arrangement, reachability is not yet a serviceable market. Ask sales to classify a few publicly inspectable company situations as a plausibility exercise. This is a usability check on the hypothesis, not a representative market study. Retain counterexamples rather than assemble only convenient accounts that appear to confirm the desired expansion. Explore this further in the guide B2B display ads: evaluate inventory quality before cheap reach.

Plan a fictional role experiment

The fictional SaaS vendor FieldWork provides a supported planning capability for service teams. All properties and quantities in this example are invented. Its campaign currently addresses operational service leaders; the team wants to examine whether operations managers in comparable business situations provide another useful entry point. It does not simply reuse the same message. The additional role receives an explanation of the organisational decision supported by planning, while product boundaries remain unchanged. Sales labels supported use cases and unserviceable expectations through an agreed process. The pilot can indicate whether the route appears suitable under observation. Without an appropriate comparison, it proves neither an entirely new audience nor additional contracts that advertising caused independently.

Protect the test against uncontrolled mixing

Record the starting audience, additional scope, mechanism, message and measurement period. Examine how overlap is handled in the selected setup and what allocation the report actually supports. Comparing the performance of two differently composed groups is not automatically a causal test of an expansion feature. Keep the offer and qualification rules stable and avoid several major targeting changes at once. Reserve a bounded portion of the learning budget and specify the finding that would justify stopping. A smaller trial can be appropriate, but low volume limits confidence. State this beforehand so the team does not scale simply because an early cost per form submission appears attractive in the platform report. Explore this further in the guide ChatGPT ads for B2B: evaluate access and pilot readiness.

How is overlap assessed?

Ask beforehand whether the same people or accounts can appear in multiple parts of the trial. If the setup does not support sufficient separation, limit interpretation. Unknown overlap is not resolved by retrospectively assuming independent audiences.

Evaluate quality at person and account level

Separate suitable responses from individual people and progress at their company. Three contacts from one account are not automatically three independent buying opportunities. Ask sales to classify enquiries using agreed reasons and review the effort required to handle them. Observe whether the additional role raises different questions requiring extra materials or specialist conversations. Cheaper forms can accompany substantially more unsuitable expectations. Conversely, a few precise technical questions may justify another bounded test without establishing reliable scaling success. Report mature milestones alongside unresolved cases so fast engagement does not replace slower commercial evaluation. The expansion should improve access to supported buying situations rather than reward a higher count of easy-to-record interactions.

Adopt expansion as a reviewable operating rule

Do not retain an expansion merely because it generated more reach. Decide using the defined buyer fit, observed downstream stages, additional handling and remaining economic allowance. A narrower role or cleaner source definition may be the appropriate continuation. Preserve the accepted hypothesis with prerequisites and a trigger for re-examination. Changes in product scope, market support or source composition can invalidate an earlier finding. Retire superseded variants from active use through the agreed process while preserving their decision history. The result is an adaptable route to additional buyers rather than an increasingly opaque collection of broader audience settings whose original purpose and commercial justification can no longer be explained.