Choose the job
Use a campaign to build recognition, teach a consequential idea, or invite a relevant next step. Keep those jobs distinct in the scorecard. If the objective is immediate qualified demand, make the offer and expected follow-up explicit.Build a useful test
- Decide whom the campaign should reach. Combine a credible business problem with observable targeting criteria. Document who the platform can actually reach, not just the ideal audience in the strategy deck. Check exclusions, geography, and permitted data use.
- Choose a useful invitation for someone browsing. A short demonstration can work before a consultation request. A report download should deliver the report, not quietly promise a meeting the reader did not request.
- Develop distinct creative hypotheses. Test the problem, mechanism, or proof angle. Changing only the background color rarely tests the underlying proposition. Keep each claim within its evidence.
- Match the destination. Repeat the promise, remove unrelated navigation when appropriate, and show what happens after submission. Test native forms as an experience, including delivery and routing, rather than as a cheaper lead counter.
- Set the test’s limits. Set spend and time limits, define the unit of comparison, and avoid changing audience, offer, and creative simultaneously. Small samples should produce tentative conclusions.
- Review attention and progression separately. Use reach and frequency to understand delivery; use appropriate downstream outcomes to evaluate the campaign’s job. Inspect lead quality, duplicates, existing customers, and delayed outcomes before scaling.
Worked example
A fictional security tool targets operations leaders opening new sites. One creative shows the access-removal problem; another demonstrates a review checklist. Both lead to the same ungated checklist with an optional implementation consultation. The checklist creative earns more clicks, but the problem creative yields more relevant consultations. The team retains the checklist for education and investigates the consultation difference before reallocating the acquisition budget. It does not declare one universal winner from two different audience behaviors.Keep a test log
Add actual spend, reach, frequency, valid responses, and the reason for the next decision. Preserve the creative version so a later report refers to the same asset.
Watch for false success
High click rates can reflect curiosity, ambiguity, or accidental taps. Low delivery may reflect a very narrow audience rather than a weak idea. Investigate those causes before rewriting the offer or expanding eligibility indiscriminately.Try it with your own work
Choose one audience and one problem they recognize. Draft two genuinely different explanations of that problem, keeping the offer and destination the same so you can learn from the comparison.Sources and scope
- LinkedIn: Matched Audiences practices describes platform audience options, not a guarantee of campaign performance.
What to read next
Paid media · Proof and claims · Lead magnet Chapter guide · All playbooksCopyright © 2026 Ivan Xu. All rights reserved. See the copyright and reuse terms. Canonical source: github.com/weilun88313/B2B-Playbook