What Is Bid Shading and How Does It Affect Your CPM?
Bid shading is a DSP algorithm that reduces your submitted bid below your maximum to avoid overpaying in first-price auctions. Instead of bidding $10 and paying $10, the algorithm estimates the minimum bid likely to win and submits that instead. It exists because programmatic advertising moved to first-price rules, removing the automatic savings that second-price auctions provided. This article covers how the algorithm works, what data it uses, how much it saves, and how it differs from bid capping.
How Does a Bid Shading Algorithm Work?
The algorithm analyzes historical auction data for a specific impression type and predicts the minimum bid likely to win. It then reduces your maximum bid to that predicted clearing price, submitting a lower number while maintaining a target win rate.
DSPs built bid shading in direct response to the shift from second-price to first-price auctions. Under second-price rules, the auction prevented overpayment automatically. Under first-price, that protection disappeared. Bid shading recreates it algorithmically during real-time bidding evaluation.
The algorithm considers five primary inputs:
- Historical clearing prices. What did similar impressions on this publisher, in this ad slot, at this time of day, clear at over the past 7 to 30 days? This is the heaviest-weighted signal.
- Win rate data. How often does the DSP win auctions at various bid levels for this inventory type? The algorithm maps a bid-to-win-rate curve and targets the point where marginal bid increases produce diminishing win rate improvements.
- Floor price signals. If the SSP communicates a floor price in the bid request, the algorithm knows the minimum viable bid. It won't shade below the floor.
- Competitive density. How many other DSPs typically bid on this inventory? Higher competition means less shading room. The algorithm bids closer to maximum in competitive auctions and more aggressively in low-competition ones.
- Campaign constraints. Your target CPA, ROAS goals, pacing requirements, and remaining budget all influence how aggressively the algorithm shades. A campaign behind on pacing will shade less (bid higher) to win more impressions. A campaign ahead of pace shades more aggressively.
Bid Shading Execution Flow
How Much Does Bid Shading Save on CPM?
Typical bid shading savings range from 15 to 30% compared to unshaded first-price bids. The exact savings depend on inventory type, competitive density, and how well-calibrated the DSP's model is for that specific auction environment.
Worked example: your campaign targets in-market auto shoppers on premium news publishers. Your bidding algorithm values an impression at $10 CPM. Without bid shading, you pay $10 for every win. With bid shading, the algorithm estimates the likely clearing price at $6.20 based on historical data, adds a margin for win probability, and submits $6.80.
The savings per thousand impressions: $3.20. On a $50,000 campaign at $10 CPM (5 million impressions), bid shading drops effective CPM to $6.80. Total spend falls to $34,000. That's $16,000 saved.
The Trade Desk reports average bid shading savings of 15 to 25% across its platform. DV360's shading is less transparent in reporting but operates on similar principles. Xandr and Amazon DSP run proprietary models with comparable ranges.
Savings vary by inventory type. Highly competitive inventory (premium publishers, above-fold video) offers less shading room because clearing prices sit closer to maximum bids. Low-competition inventory (long-tail display, off-peak hours) offers more room because the gap between your bid and the likely clearing price is wider.
What Is the Difference Between Bid Shading and Bid Capping?
Bid shading and bid capping both reduce what you pay, but through completely different mechanisms. Confusing them leads to misconfigured campaigns.
| Attribute | Bid Shading | Bid Capping |
|---|---|---|
| Mechanism | Algorithm predicts optimal bid below maximum | Hard ceiling on maximum bid |
| Who controls it | DSP algorithm (automatic) | Advertiser (manual setting) |
| Dynamic? | Yes, adjusts per impression | No, fixed until changed |
| Goal | Pay the minimum needed to win | Prevent spend above a threshold |
| Win rate impact | Maintains target win rate | Reduces win rate if cap is too low |
| Data dependency | High (needs historical auction data) | None |
Bid shading is smart. It adjusts every bid based on real-time auction context. Bid capping is blunt. It sets a ceiling and ignores everything below it.
You can use both simultaneously. Set a bid cap at $12 to prevent outlier bids, and let bid shading optimize within that ceiling. The cap acts as a safety net. The shading does the actual optimization. Most DSPs support this configuration.
How Does Bid Shading Affect Publisher Revenue?
Bid shading reduces what publishers receive per impression. If an advertiser would have paid $10 without shading and pays $6.80 with it, the publisher's revenue drops by 32% on that impression. This is a real revenue impact, not a rounding error.
Publishers have one direct lever against bid shading: floor prices. A well-set floor prevents DSPs from shading below a minimum eCPM threshold. If your floor is $7, a shaded bid of $6.80 gets rejected. The DSP either raises its bid above $7 or loses the impression.
Dynamic floor optimization is the publisher's counterbalance. As DSP algorithms learn clearing prices and shade toward them, dynamic floors adjust upward to capture more of the bid. The result is an ongoing negotiation between buyer algorithms and seller floors, recalibrating every auction.
The equilibrium usually settles at 5 to 15% above where clearing prices sat under second-price auctions. Publishers net more than they did under second-price rules. Advertisers pay less than they would without shading in first-price. Both sides gained from the elimination of the second-price inefficiency, just not equally.
Check your floor performance against CPM benchmarks and model adjustments using the CPM calculator.
What Do Most People Get Wrong About Bid Shading?
Assuming bid shading kills win rates. It shouldn't. A well-calibrated shading algorithm maintains your target win rate while reducing CPM. If your win rate drops after enabling shading, the algorithm is over-shading, which means it needs more auction data or your campaign pacing is ahead of schedule. Typical DSPs need 48 to 72 hours of auction data before shading calibrates accurately for new inventory segments.
Treating bid shading as a "set it and forget it" tool. Shading models learn from historical data. If you change your targeting, enter a new geo, or shift to different publishers, the model resets its learning. CPMs will temporarily rise during the recalibration period. Budget your first week on new inventory at 10 to 15% above target CPM to account for this.
Not checking whether it's enabled. The Trade Desk enables bid shading by default. DV360 applies it automatically across most campaign types. But not all DSPs do. Smaller or self-serve platforms may require manual activation. Verify your DSP's bid shading status before assuming you're getting optimized pricing. The CPM formula doesn't tell you whether you overpaid. Only your DSP's bid shading report does.
Frequently Asked Questions About Bid Shading
How does a bid shading algorithm estimate the optimal bid price?
Bid shading algorithms analyze historical clearing price data, publisher floor prices, win rate probabilities, and competitive bid density across similar inventory to submit the lowest bid likely to win.
Does Google Ads or DV360 automatically use bid shading?
Yes, major DSPs like Google DV360, The Trade Desk, and MediaMath feature built-in machine learning bid shading algorithms that automatically shade bids when buying inventory on first-price exchanges.
What is the difference between bid shading and bid capping?
Bid capping sets a hard ceiling on the maximum CPM a buyer will pay, whereas bid shading dynamically calculates an optimal clearing bid below the buyer's maximum valuation for a specific impression.
How much money does bid shading save advertisers on average?
Bid shading typically reduces advertiser effective CPM costs by 10% to 20%+ compared to unshaded first-price bidding, while maintaining win rates on target inventory.
How does bid shading work in a first-price programmatic auction?
In a first-price auction, the winner pays their exact bid. Bid shading calculates the minimum price needed to clear the auction above the second-highest bid, capturing surplus value for the buyer.