{"page":"/blog/reinforcement-learning-bidding","modified":"2026-11-10","license":"https://creativecommons.org/licenses/by/4.0/","columns":["RL concept","In bidding","Example"],"rows":[["State","What the bidder knows now","Time left in the day, budget left, spend rate, recent win rate, recent cost per acquisition"],["Action","What the bidder controls","A bid price per impression, or a multiplier λ applied to predicted value for the next interval"],["Reward","What the campaign wants","Clicks or conversions won, value minus cost, penalties for missing a cost target"],["Transition","How the market reacts","Competitors' bids, traffic volume and the auction's clearing prices"],["Episode","The planning horizon","One day or one flight of a campaign"]],"sourceUrls":[[],[],[],[],[]]}