Organizing a conference? Putting together a panel? In charge of the departmental colloquium?

• Strive for balanced meeting rosters: Please feel free to use the data here to determine base rates in your field, and try to recruit as representative a sample of speakers as possible.
• Encourage the representation of under-represented groups: Since base rates are themselves skewed due to myriad factors that hold women and minorities back in scientific fields, please consider a higher than representative ratio of women to men. Your conference is a vessel for scientific discourse, but also an educational tool. The younger generation (which is frequently more diverse than the faculty level) will be encouraged to remain in the field, and their own biases will be reduced, if they encounter many examples of successful women and underrepresented minorities in their field.
• “Share the wealth” of conference invitations: A few prominent women researchers get invited to many conferences (and as a result have to decline some invitations), while others are all but invisible. This is a vicious cycle as not inviting those researchers makes them even more invisible, and thus less likely to be invited in the future. Consider inviting researchers whose work is on-topic and has been recognized as scientifically sound (e.g., use NIH RePORTER to find women R01 recipients in your field, whom you have not heard of), even at the expense of other well-known speakers. More new voices means more new ideas, and a thriving scientific field. 
• Use BWN as a resource: Other conferences in your subfield did well on balancing their speaker rosters? Contact them to find out how they did it, and whether they had others on their list of suggested speakers that could not make it to that conference but can possibly come to yours.

For more ideas about how to reduce implicit bias in your conference, read this from Tin Geber.