Dynamic Bidding, Autobidders, and Fixed Bids

In ads auctions, using fixed bids may get more overall impressions, but potentially fewer conversions. Promoted solves this by supporting autobidders, a type of dynamic bidding, as an alternative to using fixed bids. There are a few common types:

• Up and down. The autobidder may bid up to 200% of the advertiser's original cost-per-click bid. Marketplaces may also choose a different cap, such as 150%. Bids are increased for higher conversion predictions, and reduced for lower predictions. As long as the advertiser is willing to bid higher, this allows for ultimate flexibility to adapt to conversion predictions.
• Down only. The autobidder only bids up to 100%, and may bid lower. This is a more conservative strategy that avoids overspending but may miss out on some opportunities with high conversion probability.

Ideal ad auctions

The goal of dynamic bidding is to move the cost-per-click (CPC) ad product closer to an ideal ad auction.

In an ideal auction, each advertiser bids exactly their expected value in dollars from winning each ad insertion to maximize advertiser profits. Therefore, the advertiser who wins the auction is the one who values the ad insertion the most. Given no interactions between insertions, this maximizes total platform advertiser value and total advertiser profits, and given a reasonable reserve price, it maximizes platform ad revenue.

Disadvantages of fixed bids

In a fixed bids CPC ad auction, advertisers bid a fixed price per click. here is an example formula:

insertionBid = p(click|insertion) * clickBid

This design has three problems in practice:

1. Advertisers cannot directly bid for each ad insertion. They can only bid for clicks. However, advertisers don’t value clicks directly — they value incremental sales.
2. Some clicks are more valuable than others because some are more likely to convert to a sale, yet a fixed cost-per-click (CPC) bid values all clicks equally.
3. Real auctions are not “dominant strategy” or “bid your true value” auctions. Maximizing the allocation’s (assumed) advertiser value does not necessarily maximize profits, because advertisers may be able to win more profit by bidding lower. Therefore, advertiser bids are not reliable signals of true value.

Dynamic bidding addresses allocation maximization problems (1) and (2), allowing the system to bid closer to the "true value" per insertion on behalf of advertisers. The better the value estimates used in bidding, the more that estimated value will be realized in expectation. Problem (3) relates to pricing and bid incentive theory, which are generally beyond the scope of dynamic bidding.

Bidding closer to the ideal generates more ad value and, therefore, more ad revenue.

Dynamic bidding goals

Dynamic bidding works by making a cost-per-click (CPC) ad auction more similar to an ideal ad auction by bidding for sales using machine learning while keeping the CPC click bid in the insertion bid formula.

The goal of dynamic bidding is to increase ad value. Ad value is the total value (in terms of sales) to advertisers from ads and the maximum sustainable ad revenue that can be billed. Advertisers value sales, so ad value for dynamic bidding is the sum of incremental sales.

Compared to CPC, here are the results of dynamic bidding:

• Dynamic bidding more correctly predicts advertiser value (in sales) per insertion better than any per-ad constant multiplier of any click prediction model.
• The ads allocation will differ. If there is no substantial allocation change of ads, then there can be no difference in ad value generated (prices may differ, but this should be negligible if the allocation is unchanged)
• Total ad value is increased because ads are more correctly allocated to generate more ad value. Assuming ad volume is constant when compared to CPC ad delivery volume, the total ad system must have become more efficient overall and in expectation per advertiser to generate more total sum ad value if total sum ad value increases.
• Dynamic bidding does not penalize specific advertisers for realizing value. If a particular advertiser were to be heavily penalized, that advertiser will rationally refuse to use dynamic bidding. If the additional ad value in an A/B test derives from the penalty of that particular advertiser, then the increase in ad value will not be realized.