The Gradient Intent Strategy
Same auction, different clearing prices by bidding on fractional values.
How Value-Based Bidding Works
Google's Smart Bidding algorithms set CPC ceilings based on predicted conversion value, predicted conversion rate (CVR), and your ROAS targets. When advertisers report full order value, the algorithm interprets this as high ability-to-pay and raises bid ceilings. When affiliates report only fractional commission values, the algorithm constrains bid ceilings, keeping the account in a lower clearing price range within the same unified auction.
The bid ceiling is influenced by predicted conversion value (a product of CVR and order value) and your ROAS target. We control two parts of this equation: the value we report and our target ROAS, which is directly tied to commission rate.
High values → high bid ceilings → premium clearing prices.
Fractional values → low bid ceilings → lower clearing prices.
Understanding the “Gradient”
We use “gradient” as a metaphor for value scale steepness: the relationship between the values you report and the bid ceilings that result. This is not a claim about how Google's algorithms work internally. It's a conceptual model that describes observable auction outcomes.
High-Value Signals
When you report $200 order values, you signal high ability-to-pay. This produces high bid ceilings that compete in premium clearing price ranges where CPCs are elevated.
Low-Value Signals
When we report $24 commission values, we signal low ability-to-pay. This produces low bid ceilings that compete in lower clearing price ranges where CPCs remain suppressed.
Same Auction, Different Clearing Prices
All bidders enter the same auction, but they bid with different ceilings. Your $200 tells Google you can pay more. Our $24 tells Google we cannot. Same auction, different likelihood of winning at different price points.
Why Advertisers Cannot Replicate This
Advertisers cannot arbitrarily compress conversion values without violating Google's conversion accuracy requirements and destabilizing their own ROAS modelling.
This isn't a workaround or a hack. It's a structural reality:
- Policy compliance: Google requires advertisers to report accurate business value. Under-reporting (beyond legitimate adjustments like excluding VAT) violates policies.
- ROAS economics: Artificially lowering values destabilizes the data model that value-based bidding uses to optimize campaigns.
- Profitability: Advertisers cannot run on low-value bidding models and maintain profitable campaigns.
- Economic constraint: Advertisers cannot structurally compress their value distribution the way affiliates can.
Affiliates can operate on fractional values because fractional commission is our actual revenue. We're not misrepresenting anything. We're reporting what we actually earn.
Lower Clearing Prices, Same Auction
Fractional-value bidding constrains CPC ceilings, allowing affiliate accounts to win inventory at lower clearing prices that advertisers cannot profitably reach.
This happens because:
- Lower value signals produce lower CPC ceilings automatically
- Low-value accounts receive conservative bid ceiling calculations
- Lower bid ceilings mean we win long-tail inventory where the advertiser's max CPC exceeds efficient clearing prices
- High-value signals trigger more aggressive auction pricing; fractional values avoid this
We don't “unlock hidden segments.” We compete in the same auction with different economics. Advertisers can bid on this inventory, but not profitably at these clearing prices.
Why High-Value Signals Increase CPC Exposure
Recent US DOJ and EU Commission antitrust cases uncovered several auction mechanisms Google used to raise advertiser costs. These mechanisms disproportionately impact accounts reporting high conversion values. The key insight: auction inflation pushes harder where Google detects high value density and revenue potential.
Low-value signals do not trigger aggressive inflation because fractional commission values never hit the thresholds that activate aggressive pricing mechanisms. This is why fractional-value bidding produces structurally lower CPCs. Not because we're gaming the system, but because our reported values constrain what Google's algorithms can extract.
Documented Auction Mechanisms (DOJ/EU Findings)
These are factual findings from court proceedings, not conjecture. They demonstrate how high-value signals expose advertisers to greater cost pressure within the unified auction.
Google uses quality scores to adjust advertiser costs. High-value accounts face more aggressive quality score adjustments that increase CPCs.
US DOJ Legal FilingGoogle's bid shading algorithm reduces advertiser bids to just above the second-highest bidder, extracting maximum revenue. This disproportionately affects high-value accounts.
EU Commission Antitrust CaseGoogle's algorithms adjust pricing based on predicted conversion value. Accounts reporting high values face higher CPCs as the system extracts more revenue.
US DOJ Legal FilingThe connection: These mechanisms inflate CPCs based on predicted value and ability-to-pay signals. Fractional commission values constrain that signal, which constrains the inflation. This is structural economics, not a workaround. We report what we actually earn.
Honest About Overlap
A common concern: will affiliate campaigns cannibalize your existing traffic? We won't claim zero overlap. That would be unrealistic.
Overlap is structurally limited because the affiliate's CPC ceilings are constrained by fractional commission values. Both accounts compete in the same auction, but with different bid ceilings. Your campaigns compete at premium clearing prices. Our campaigns win at lower clearing prices where your high-value signals make competition unprofitable for you.
Yes, there will be some overlap. But it's limited by structural value differences. We publish the economics, not fairy tales. Where your campaigns would need to bid aggressively (and expensively) to win, our campaigns operate efficiently on lower CPC ceilings.
The result: broader market coverage across different clearing price ranges, with overlap limited by the natural separation of bid ceilings.
Advertiser vs. Affiliate Optimization
Two different economic models produce two different optimization environments. Both serve the same objective: maximizing your total revenue.
A Concrete Example
Different ROAS targets reflect different cost tolerances. Here's how the math works at equivalent economics.
Advertiser
($200 × 0.03) ÷ 8.33 = $0.72
RetailerBoost (Affiliate Model)
($24 × 0.03) ÷ 1.1 = $0.65
The insight: Even with equivalent economics (833% ROAS = 12% commission), our max CPC ceiling ($0.65) is lower than the advertiser's ($0.72). This difference comes from fractional value reporting: by reporting $24 commission instead of $200 order value, we operate with lower value signals that constrain CPC ceilings and avoid aggressive auction pricing and Google bid inflation. This gives the affiliate model a slight edge and more fiscal control over spend - the dials are more finely tuned, so Google can't be as aggressive.
Affiliates operate on tighter ROAS margins (110%, sometimes 105%) because we play a volume-weighted portfolio game. We profit at scale, not on individual merchants or sales. Advertisers cannot replicate this economics because they cannot report fractional revenue.
Limitations of the Model
No strategy works everywhere. Here are the constraints we operate within.
Cannot beat brand campaigns
Brand terms have high CVR and high advertiser willingness-to-pay. We exclude brand terms entirely.
Cannot win core head terms in all categories
High-competition head terms often have clearing prices above our ceiling. We focus on mid-tail and long-tail.
Constrained by commission rates
Our bid ceiling is directly tied to commission. Low-commission programs produce very narrow CPC windows, and in some cases can collapse the model and make it unprofitable to pursue.
Not every category yields incrementality
Some verticals have saturated low-CPC positions. Results vary significantly by category.
Lower ceilings can reduce volume in some cases
In high-competition verticals where CPC floors exceed our ceiling, traffic volume is constrained. Lower ceilings do not inherently reduce volume unless the category's clearing prices exceed commission-based ceilings.
Two identical bids can produce different outcomes
Smart Bidding optimizes to predicted value density. Two accounts bidding the same ceiling can still produce different CPCs because of quality score, CVR predictions, and ad relevance.
Common Questions
Whether you're a seasoned PPC professional or new to paid advertising, here are the questions we hear most often.
Isn't this just bidding lower? I can set CPC caps.
You can set CPC caps, but that's not the same thing. When you cap your bids while reporting high values, Google's algorithm sees a mismatch: “This advertiser says they earn $200 but won't pay for it.” The algorithm deprioritizes your campaigns and you lose volume. With fractional-value bidding, the algorithm naturally finds cheaper auctions because it thinks that's all we can afford. We're not fighting the system. We're giving it different economics.
Won't low bid ceilings just get low-quality traffic?
Traffic quality is primarily determined by query intent, not CPC alone. A search for “buy running shoes size 10” has strong purchase intent regardless of the click cost. That said, CPC does correlate with competitive intent in some cases. It's not entirely decoupled. We focus on queries where lower clearing prices don't compromise conversion quality.
This sounds like cannibalization. How do you minimize overlap?
We can't eliminate overlap entirely, and we don't claim to. But we minimize it structurally: we exclude brand terms, focus on non-brand Shopping and Search, and target long-tail queries your campaigns may underserve. Our low bid ceilings mean we naturally win auctions your campaigns wouldn't compete in profitably. And even where overlap exists, you're paying a fixed commission on confirmed orders, not inflated CPCs. We're not stealing clicks. We're finding cheaper ones.
Why do smaller value signals work better?
Smaller signals give Google's algorithms less room to inflate. When you report $200, the algorithm has a wide range to work with and will push CPCs higher. When we report $24, the algorithm has very little headroom. It's forced to find efficient inventory because there's nowhere else to go. Smaller signals mean tighter control over what you actually pay.
Can't I just report lower values to Google?
Google requires advertisers to report accurate business value. Under-reporting would violate their conversion tracking policies and destabilize your ROAS modeling. Note: advertisers can legitimately use cart value excluding VAT, which is accurate. But they cannot arbitrarily compress values. Affiliates report fractional values because commission is our actual revenue. That's compliant.
How do I know these orders are truly incremental?
We measure using post-click, last-click attribution across isolated parallel campaigns. We exclude brand terms and focus on non-brand Shopping and Search. This methodology has limits. Last-click attribution isn't perfect. But running parallel campaigns with brand isolation gives us confidence that conversions are from inventory you weren't efficiently reaching.
Why can't my PPC agency do this instead?
Google's policies require advertisers to report accurate conversion values. Your agency can't under-report your $200 order as $24 without violating those policies and breaking your ROAS modeling. Affiliates are different: our actual revenue is the commission we earn, so reporting $24 is accurate and fully compliant. We're not misrepresenting anything. We're reporting what we actually make.
Doesn't the auction just look at who bids highest?
No. Google's auction runs on black-box algorithms that use dozens of signals: your value reporting, ROAS targets, predicted conversion rates, ad quality, user history, and more. Your bid ceiling is a calculated output, not a direct input you control. Two advertisers targeting the same query can pay completely different CPCs because their value signals tell Google different things about their ability to pay.
Won't lower bid ceilings mean less traffic or worse positions?
Not necessarily. Google serves traffic based on budget, targets, and predicted value. Lower ceilings generally mean you pay less per click, but in high-competition verticals where CPC floors exceed your ceiling, volume can be constrained. Carousel position depends on ad relevance, predicted CTR, and auction dynamics, not just bid amount. Results vary by category.
Incremental Uplift by Category
Percentages represent incremental orders from 3,000+ stores running isolated parallel campaigns. These are conversions from clearing price ranges that advertisers cannot profitably reach due to bid ceiling constraints.
Methodology
- Attribution: Post-click, last-click attribution
- Brand isolation: Brand terms excluded from affiliate campaigns
- Campaign structure: Isolated parallel campaigns (not mixed budgets)
- Measurement window: Consistent 30-day conversion windows
- Category variance: Results vary by vertical. Not all categories perform equally
We cannot share retailer-specific data due to confidentiality agreements. These are aggregate figures across our portfolio.
The Bottom Line
Same Auction, Different Ceilings
Advertisers report full order value. Affiliates report fractional commission. Same auction, but different bid ceilings and different clearing price ranges.
Value-Indexed Bidding
Lower values produce lower CPC ceilings. This constrains where we can compete profitably, but also where advertisers cannot compete efficiently.
Structural Economics
This is not manipulation or gaming. Affiliates report fractional values because commission is our actual revenue. The economics are real, but so are the limitations.
