Before launch, give people who did not build the campaign the explicit job of attacking it as a skeptical journalist, a regulator, a competitor, a harmed customer, and a bad actor would, and fix what they find while it is still cheap to fix.

How to do it

1. 1Appoint reviewers who are independent of the project and whose standing does not depend on its success. Give them the full materials, including targeting, data sources, and the parts of the funnel that come after the advertisement.
2. 2Assign perspectives. The journalist asks what the most damaging accurate story would be. The regulator asks which claims lack substantiation, which disclosures are missing, and which design choices impair consent. The competitor asks what can be challenged. The member of a vulnerable audience asks what this does to someone in a bad week. The bad actor asks how the mechanics (referral rewards, review systems, user-generated content, lookalike targeting) could be abused.
3. 3Run a premortem alongside it. Ask the whole team to imagine that it is a year from now and the campaign has become a scandal or a failure, and to write down, individually and before discussion, the most likely reasons.
4. 4Require findings in writing, ranked by severity, each with a named owner and a decision: fix, accept with a stated reason, or stop. A senior person who is not the campaign's sponsor should own the accept decisions.
5. 5Protect the reviewers. If raising a problem costs people socially or professionally, the exercise will produce reassurance.
6. 6Repeat after significant changes, and compare predictions with what actually happened so that the next review is better.

When to use it

• •Before any high-reach, high-stakes, or sensitive campaign: health, finance, politics, children, crisis communication.
• •When a team has worked closely together for a long time and everyone agrees.
• •When a campaign involves a new tactic, a new data source, or automated generation and optimization.
• •After an incident, to find what the previous review process missed.

Counters

Evidence and how strong it is

Red teaming comes from military and security practice, where an independent group plays the adversary in order to find weaknesses that an organization's own planners cannot see; Zenko (2015) surveys its use and its failure modes across the military, intelligence, and corporate worlds, and concludes that it works only when the red team is genuinely independent, has leadership backing, and is listened to. The evidence is largely case-based. The closely related premortem has somewhat better support: Klein (2007) describes the method, which builds on experimental work by Mitchell, Russo and Pennington (1989) showing that imagining an outcome as having already happened leads people to generate more, and more specific, reasons for it. The need for an outside challenge is well grounded in research on group decision-making, where cohesive groups suppress dissent and converge prematurely (Janis 1982), and where assigned devil's advocacy and structured dissent improve decision quality in experiments, although authentic dissent tends to work better than role-played dissent (Nemeth, Brown & Rogers 2001). No studies evaluate red teaming of marketing or advocacy campaigns specifically. Evidence strength: practitioner consensus with case evidence, and moderate experimental support for the component ideas.

1. Zenko, M. (2015). Red Team: How to Succeed by Thinking Like the Enemy. Basic Books
   A survey of red-team practice across sectors and of the conditions (independence, leadership support, willingness to act on findings) under which it works.
2. Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18-19
   The premortem method of imagining that a project has failed and working backwards to the likely causes.
3. Mitchell, D. J., Russo, J. E., & Pennington, N. (1989). Back to the future: Temporal perspective in the explanation of events. Journal of Behavioral Decision Making, 2(1), 25-38
   Experimental evidence that imagining an event as already having occurred increases the number and specificity of reasons people generate for it.
4. Janis, I. L. (1982). Groupthink: Psychological Studies of Policy Decisions and Fiascoes (2nd ed.). Houghton Mifflin
   The case-study account of how cohesive groups suppress dissent, and the recommendation of assigned critical evaluators and outside experts.
5. Nemeth, C., Brown, K., & Rogers, J. (2001). Devil's advocate versus authentic dissent: Stimulating quantity and quality. European Journal of Social Psychology, 31(6), 707-720
   Experimental evidence that authentic dissent stimulates better thinking than role-played devil's advocacy.