Bid shading: tests & results to avoid overspending on inventory
In the context of first-price auctions, bid shading is when SSPs share the winning price and the second-highest price of the auction with the RTB participants who’ve lost the auction. They share only the second-highest bid price with the winner of the auction. This addresses the issue with price transparency: the winner paid their bid of $5 but they could have paid $2.03 instead because everyone else had bid $2. Granular information like this allows DSPs such as Adikteev to adjust their own bids accordingly, and ultimately bid more accurately.
Bid shading was conceived with the clear-cut intention of saving RTB participants money when bidding on ad inventory and creating a more level playing field. In a perfect scenario, everyone would like to pay a lower amount for each auction while still winning. In risking less but still given a clearer path to winning, the client metric could be optimized for further bids and for generating more conversions.
Unfortunately, RTB doesn’t always pan out this way. It’s proven that the lower the bid price, the lower the chances of winning— it’s a competition after all, like a game of poker. If the bid price is reduced, campaign performance will not improve and the bidder will most likely lose a lot of auctions. Contrastingly, increasing the bid price too high will result in overpaying for impressions, especially in first-price auctions.
Therefore, bid shading has evolved to be a balancing act between lowering the bid price to maximize ROI and maintaining a competitive bid to increase the chances of winning the auction in a first-price auction environment. In this way, reaching the target audience at the right price is more feasible, especially in retargeting scenarios where the audience size is strictly limited.
Methods to optimize bid shading
To meet the expectations of both the advertiser and the publisher, bid shading has to be meticulously and cleverly adjusted. At Adikteev, we have looked at different approaches— from simple to complex ones, with others rule-based and others derived from machine-learning models. Multiple strategies have been tested during the research phase, including surplus maximization (maximizing the variance between the expected revenue from a prediction and the cost of obtaining said prediction), shading ratio (the proportion of the budget that is reserved to each impression or advertising campaign), and bid capping (the control over the maximum bid an advertiser is willing to pay for an impression to avoid overspending) with diverse mechanisms.
As mentioned in the beginning, the idea is to incorporate feedback from SSPs into our pricing approach. Depending on the SSP and the nature of the bid, we are given the second-highest bid price if we’ve won the auction, or both the second-highest price and the first price (winning price) when losing. Doing so allows us to consider what other RTB participants are spending on the impression, and how valuable an impression is for our competitors.
In a nutshell, we use this historical bidding information to attempt to bid just right above the highest bidder, instead of bidding way higher with little visibility. This allows for more precise optimizations, less wasted budget, and a maximized ROI by paying only the minimum amount necessary to win the impression.
The results we have seen
We have performed various A/B tests, from straightforward parameters to substantially enhanced metrics.
What does this mean?
Changing user expectations and preferences, increasing competition, an evolving market, and a challenging global economy— all these factors call upon innovation. Now, more than ever, app marketers are forced to be creative and efficient, to produce technologies and solutions that offer new ways to reach and engage audiences. Bid shading is exactly one of those solutions, created for a highly competitive market to optimize the way app marketers bid on inventory. For app marketers, bid shading allows cost savings, improved ad relevance and performance, and better targeting. Accurate budget allocation and having ads reach the right audience at the right time can only assure the achievement of marketing goals and maximize ROI.