Value-based bidding needs values you can trust
Cost per conversion treats a $99 signup and a $50,000 deal as the same win, and Smart Bidding believes it. Here's what value-based bidding actually changes, how to build values you can trust, and the honest checklist for when switching will make things worse instead of better.
Your account hits its CPA target every month, and revenue is flat anyway. That gap is the entire case for value-based bidding: telling Smart Bidding what each conversion is actually worth (a deal amount, a cart total, a predicted lifetime value) instead of counting every conversion as equal, so it optimizes toward revenue instead of volume. Here is the argument of this essay, and the reason to read before you switch: most accounts don't have values worth bidding toward yet, and switching on top of a bad value feed just teaches the algorithm your mistake faster. Fix the feed first.
#CPA is lying to you
Cost per conversion treats a $99 self-serve signup and a $50,000 contract as the same event. That single design choice quietly corrupts every decision built on top of it: the "efficient" campaign in your report is often the one efficiently buying the customers you least want, and the "expensive" one is funding your best deals. An account can hit target CPA every month while revenue goes sideways, because CPA measures the cost of a conversion, not the value of the conversion.
The proof is one table: take the same account and period, rank campaigns by cost per conversion, then rank them again by revenue-weighted value per dollar of spend, and watch which campaigns flip position. The pattern looks like this (illustrative numbers: the shape is what every account we've run this on produces):
| Campaign | Platform CPA | CPA on CRM-matched revenue | What happens |
|---|---|---|---|
| Brand search | $38 | $41 | holds, cheap both ways |
| Generic "audit software" | $96 | $54 | flips up: high CPA, but the leads close |
| Competitor terms | $52 | $187 | flips down: cheap clicks, deals never close |
| Display retargeting | $29 | $210 | collapses: form fills, no revenue |
Ranked by platform CPA you'd cut the generic campaign first and scale retargeting. Ranked by matched revenue, that's exactly backwards. Every account we've run this on has flips, and the campaigns that flip are exactly where the budget conversation goes wrong.
Smart Bidding inherits the same blindness. Feed it equal-weighted conversions and it will dutifully buy the cheapest conversions available, which is the correct answer to the wrong question.
#What value-based bidding actually is
Value-based bidding is the pair of Smart Bidding strategies that optimize toward conversion value instead of conversion count (Google's value-based bidding best practices):
- Maximize conversion value spends your budget toward the highest total value it can find (About Maximize conversion value bidding).
- Target ROAS (tROAS) is Maximize conversion value with a return-on-ad-spend target attached: it trades volume for efficiency at the return you set.
That's the whole menu: the mirror pair of Maximize conversions and target CPA on the counting side. The strategies are the easy part. The question that decides whether they work is upstream: where do the values come from, and are they true?
#The values ladder: four rungs, climb in order
Every stack enters value-based bidding somewhere on a four-rung ladder, and each rung feeds the algorithm something closer to the truth:
Static values, one flat number per conversion action ("a demo request is worth $200"), are better than nothing because they at least rank your conversion actions against each other. Conversion value rules adjust values by location, device, or audience without touching your tags (About conversion value rules); Google applies them at bidding time, and they show up in your value columns so reporting and bidding stay consistent (how value rules affect Smart Bidding). Dynamic values are the real thing: actual cart totals passed by your ecommerce tags, or actual deal amounts imported from the CRM. Predicted LTV is the top rung, modeling what the customer will be worth before the revenue lands, and it has its own set of prerequisites, covered in Predictive LTV and value-based bidding for B2B SaaS.
Where you enter depends on your stack. Ecommerce accounts (Track B) usually have dynamic cart values available on day one and under-trust them. Lead-gen and B2B accounts (Track A) have to build the value feed, which is exactly the offline import pipeline in Offline conversion tracking: send CRM deals back to Google Ads and the hashed-email fallback in Enhanced conversions for leads: CRM matching for B2B.
#Values you can trust: the feed is the product
Here is the position this page exists to state: values are a signal-quality problem before they are a bidding-strategy problem. The strategy switch takes five minutes in the UI; the value feed is the actual work, and its failure modes are the same silent ones that break conversion tracking everywhere else:
- Lag. Values that arrive days late (a slow CRM import cadence) mean the algorithm bids today on last week's truth. Import cadence rules are in Offline conversion tracking: send CRM deals back to Google Ads.
- Duplicates. A purchase counted by both the pixel and a server event, or a deal imported twice, doubles its apparent value and teaches the algorithm to overbuy that segment.
- Currency and unit errors. One feed sending cents while another sends euros is a 100× value distortion that no error message will ever surface.
- Zero-variance values. If every conversion carries the same number, you've rebuilt counting with extra steps: value-based strategies need value differences to optimize toward.
If the platform's own claimed values already look inflated to you, that instinct is worth keeping: platform-reported conversion values have their own overcounting problem, which Every platform claims the same conversion: why ad platforms overreport takes apart.
#When NOT to switch yet
Value-based bidding is not the default right answer, and the honest checklist keeps you off the rung you're not ready for:
- Too little volume. Google's own guidance for target ROAS is at least 15 conversions in the past 30 days for most campaign types (About Target ROAS bidding), and value strategies want comfortably more than the minimum, because value adds a second dimension the model has to learn.
- Broken tracking. If conversions themselves are incomplete or double-counted, values multiply the error. Run the The conversion tracking QA checklist: test it like you'd test code first.
- No value variance. One product, one price, no upsell? Maximize conversions is already optimizing the right thing, and switching buys you model retraining for nothing.
The rule of thumb: fix the feed, then pick the strategy. An account that switches to tROAS on top of a garbage value feed doesn't get a better algorithm; it gets its mistakes executed with more conviction.
#What the algorithm does with your values
Downstream of the feed, Smart Bidding treats your values as the thing it's trying to maximize: it learns which auctions (query, device, audience, time) tend to produce high-value conversions and shifts spend there at auction time. How it weighs everything else it runs on (freshness, consent-modeled data, volume), and how each input degrades bidding when it breaks, is the subject of Smart Bidding is a signals problem: what the algorithm actually runs on. And once you're sending real values, the natural next question is what a customer is worth over time, which is where Customer lifetime value for marketers: how to calculate, benchmark, and use it picks up.
#The strategy was never the hard part
Zoom out once and the whole essay collapses to a line: value-based bidding is a data-quality project wearing a bidding label. The switch to tROAS takes five minutes; earning the right to make it is the work. And the order of operations (fix the feed, then pick the strategy) is not special to Google Ads. It's how to approach any system that optimizes against your data, from Meta's value rules to whatever you point a model at next. Get the values right and every downstream tool gets smarter for free. Get them wrong and each one amplifies the same error.
Which leaves one honest problem: a value feed is never finished. A currency flips, a CRM import lags, a second pixel starts double-counting a purchase, and the values quietly go wrong while every dashboard stays green. That's the part worth handing off. Buron joins your ad spend to your actual revenue (CRM deals, Shopify orders) so you can see what your conversions are really worth before you ask an algorithm to bid on them, and it keeps the CPA-flip table above as a standing view rather than a one-off: the day the ranking flips or a value feed drifts, you get a dated finding instead of a quarter of misdirected spend. *[See what your conversions are worth →]
Frequently asked questions
What is value-based bidding?
Value-based bidding means sending Google Ads a value for each conversion (a deal amount, a cart total, a predicted LTV) instead of counting every conversion as equal. Smart Bidding then optimizes toward revenue rather than volume, using strategies like target ROAS or Maximize conversion value.
What are the two types of value-based bidding strategies?
Maximize conversion value, which spends your budget toward the highest total conversion value, and target ROAS, which is Maximize conversion value with a return-on-ad-spend target attached so it trades volume for efficiency at the ROAS you set. Both require conversion values; without them they degrade to counting.