Situational Shading Strategies

Buyers often ask about bid shading: should they do it? How aggressively? Is it situational? So I decided to aggregate my thoughts and share.

WHAT IS BID SHADING?

Bid Shading is a buy-side technology that, when enabled, automatically reduces bids for ad slots in first price auction environments where the bidding platform believes that the buyer has overpaid for a given impression.  Buyers love bid shading because it saves money and prevents them from overpaying on an impression-by-impression basis. Thus, there’s an increasingly pervasive theory in the industry that, as more publishers and exchanges move to 1st price auctions, it’s always better to shade as aggressively as possible and many platforms have adopted this approach with a one-size-fits-all “universal” bid-shading switch in their UI. In reality, the amount of shading that a buyer should do is directly correlated with the campaign result that he or she is trying to achieve. 

HOW DOES BID SHADING WORK?

Bid shading tech sits on top of a programmatic bidder’s footprint, collecting pricing information about won and lost impressions for various environments around the internet. As buying takes place, the algorithm learns information about average price at a placement level, while controlling for other available signals: daypart, device, OS, above/below the fold inventory, ad size, and all manner of other factors that we’d expect to be available in programmatic environments. These algorithms observe and predict pricing based on circumstances regarding the absence or presence of various factors: for instance, if a customer bids $8 and the bid shading tech believes that the combination of variables for a given placement should warrant a $5 bid, that’s what they might return. However, for any number of reasons, the $5 bid may not win, not only because bid shading doesn’t exist in every platform, but also because brands value a given impression very differently based on their needs and goals. 

SHOULD I BID SHADE?

It depends. Bid shading can essentially be viewed as a trade-off between average price paid and average win rate: the more you shade, the greater the chance you have to lose a given auction. However, shading so you forgo an overpriced impression may be advantageous in-of-itself in the right circumstance. The goal of sophisticated bid shading algorithms is to reduce prices on the right impressions vs. manually clear-cutting CPMs across the board which will simply lead to overall lower inventory quality and, likely, lower win rates. In most circumstances, this trade-off is welcome; the question is how much? 

SITUATIONAL STRATEGIES FOR BID SHADING

How much to bid shade depends on a number of factors, but to simplify, I created a rubric below. In general, more narrowly targeted line items might run into delivery issues if they’re shaded too aggressively, where broad targeting can handle more aggressive shading without risking under-delivery. From a KPI standpoint, performance-based campaigns (particularly those with flexible budgets) can shade fairly aggressively where brand and/or reach campaigns may not want to risk missing out on impressions that would limit overall campaign visibility.

EXAMPLE SITUATIONS

Here are some campaign examples:

1. Let’s say you have a mass-market campaign, with a strict performance goal and lots of available inventory (think a blended cross-device CPA campaign for an e-retailer). For this setup, you’re probably ok to lose bids because of aggressive shading, because inventory is plentiful and cost of inventory and campaign KPI performance is the most important result.
2. Alternatively, let’s say you’re running a reach campaign to a very specific audience in a tight geo and on a short timeline (this is typical for political advertising). Ad fulfillment opportunities will likely be scarce and so buyers may only want to shade the most egregiously expensive bids, or potentially not shade at all. 
3. Maybe you’re a brand buyer with multiple goals, say both a viewability % and completed view % objective, but with a medium-sized audience target. This buyer would probably want the chance to shade somewhere in the moderate “goldilocks zone” of saving as much money as possible, without letting the win rate get too low and under-delivering budget.