When Concord spots a recurring preference across your chats, for example a naming convention you keep correcting, it suggests a new Skill on the spot. Accept the suggestion and the next conversation already knows the rule.
Line item names follow: country / format / objective / phase.Apply detailed taxonomy only at line item level.Campaign and IO names should mirror the brief exactly.
QA checklist before launch:
Before submitting any DV360 line item:- Frequency cap set- Brand safety lists applied- Geography matches the brief- Pacing is "Even" unless the user asks for "Ahead"
Reporting defaults:
Default report columns: impressions, clicks, CTR, viewability rate, VCR, cost.Always group by line item by default.Compare to previous period unless a different range is requested.
Brand-specific positioning:
Premium athletic wear, target serious runners and fitness enthusiasts.Avoid mass-market language and price-driven messaging.
One topic per Skill. Smaller, focused Skills are easier for Concord to load only when they matter. Avoid a single mega-Skill with every rule.Be specific. Spell out exact thresholds, names, and formats. “Always use the brief budget” is clearer than “match the brief”.Cover what to do when info is missing. Tell Concord to ask, use a placeholder, or apply a default, so it doesn’t guess.Scope tightly. If a rule only applies to one brand, scope the Skill to that brand. Workspace-wide rules end up loaded everywhere.Disable rather than delete. Turning a Skill off keeps the history if you want to re-enable later.
The fastest way to bootstrap good Skills is to let Concord propose them as you work. Every accepted suggestion is one less thing to type tomorrow.