How Does Value-Based Bidding Work in Google Ads?

How Does Value-Based Bidding Work -- and Why Does It Keep Failing?

A B2B SaaS client came to me last quarter spending EUR 18,000 per month on Google Ads. They had switched to Target ROAS two months earlier after reading that value-based bidding would improve their results. It had not. Cost per qualified lead had actually increased by 35 percent, and the sales team was complaining that lead quality had dropped.

The problem was not the bidding strategy. The problem was the data feeding it. Their conversion tag fired on every demo request with a static value of 1. Every lead looked identical to Google's algorithm -- a student filling out a form for a class project and a VP of Engineering evaluating a EUR 80,000 annual contract were worth the same. The algorithm had no value signal to optimize against, so it optimized for volume instead.

This is the pattern I see in roughly two out of three accounts that attempt value-based bidding in Google Ads. The strategy is sound. The tracking underneath it is not.

What Is Value-Based Bidding?

Value-based bidding is a subset of Google's Smart Bidding that optimizes for total conversion value rather than conversion count. Instead of telling Google "get me conversions at EUR 80 each," you tell it "maximize the total revenue from my budget -- and here is what each conversion is actually worth."

The shift matters because not all conversions are equal. A EUR 500 e-commerce order and a EUR 35 impulse buy are both "purchases," but they contribute very differently to your bottom line. A qualified B2B opportunity and a spam form fill are both "leads," but only one generates pipeline. Value-based bidding gives the algorithm the information it needs to tell them apart.

According to Google's own data, advertisers that switch from a target CPA strategy to a target ROAS strategy see a median 14 percent increase in conversion value at a similar return on ad spend. That number assumes the conversion values feeding the algorithm are accurate. When they are not -- and in most accounts I audit, they are not -- the results go the other direction.

The Two Types of Value-Based Smart Bidding Strategies

If you have searched "what are two types of value-based smart bidding strategies," the answer is straightforward. Google offers two:

Strategy	Objective	When to use
Maximize Conversion Value	Spend the full budget to generate the highest possible total conversion value	You want maximum value and are comfortable spending your full daily budget
Target ROAS	Generate maximum conversion value at a specific return-on-ad-spend target	You need efficiency control -- a minimum return per euro spent

Maximize Conversion Value is the less constrained option. Its first priority is spending your budget; its second is maximizing total value. If you set a daily budget of EUR 500, Google will try to spend all EUR 500 while pursuing the highest-value conversions it can find. There is no efficiency floor. This works well for accounts that have a fixed budget and want to extract as much revenue as possible from it.

Target ROAS adds a constraint. You specify a target return -- say 400 percent, meaning EUR 4 in conversion value for every EUR 1 in ad spend. Google's algorithm then predicts future conversion value before each auction and only bids aggressively when it expects to hit your target. If it cannot find enough high-value auctions, it may underspend your budget. Efficiency comes first.

Both strategies rely on the same underlying mechanism: Google's Smart Bidding AI uses your reported conversion values to predict which clicks will generate the most value, then adjusts bids in real time. The algorithm is the same. The difference is whether you prioritize total value or efficiency.

For a deeper look at how ROAS targets interact with your actual numbers, I wrote about the measurement side in ROAS Calculation: The Formula Is Easy, Your Inputs Are Wrong.

What Value-Based Bidding Needs From Your Tracking

Here is where most implementations fall apart. Understanding how value-based bidding works is the easy part. Making it work requires accurate, differentiated conversion values flowing into Google Ads consistently. That means your tracking needs to clear four bars.

1. Conversion values must be real and variable

Static values defeat the purpose. If every conversion sends a value of 1 or 100, every conversion looks identical to Smart Bidding. There is no signal to optimize against, and you have value-based bidding in name only.

For e-commerce, the fix is usually straightforward: pass the actual order total through the value parameter in your purchase event. But even this goes wrong more often than you would expect -- I regularly find tags pulling the value from a data layer variable that returns undefined on certain checkout paths, or sending revenue including tax when the business reports on pre-tax figures.

For lead generation and B2B, the challenge is harder. The real value of a lead is not known at the time of conversion. It becomes clear weeks or months later, when the deal closes or does not. This is exactly the problem that offline conversion tracking solves -- capturing the click ID, storing it in your CRM, and importing the actual deal value back into Google Ads when it materializes.

2. Enough conversion volume for the algorithm to learn

Google requires a minimum of 15 conversions in the past 30 days for Search campaigns using Target ROAS. For Shopping, it is 20 conversions in 45 days per Merchant Center ID. Demand Gen campaigns need 50 conversions with value in the past 35 days.

These are minimums. In practice, I do not recommend switching to value-based bidding until an account has at least 30 to 50 conversions with valid, variable values per month. Below that threshold, the algorithm has too little data to distinguish patterns, and bid volatility tends to spike.

For B2B accounts with long sales cycles, this creates a real problem. If you only close ten deals a month, you may never have enough downstream conversions to feed Target ROAS directly. The workaround is to import intermediate pipeline stages -- qualified lead, opportunity created, proposal sent -- with modeled values based on historical close rates. I walk through this approach in Marketing Measurement for Long B2B Sales Cycles.

3. Conversions must actually reach Google Ads

This sounds obvious, but it is the most common failure point. Value-based bidding cannot optimize on conversions it never sees. And a surprising number of conversions never make it to Google Ads.

Cookie loss. Safari's ITP caps JavaScript-set first-party cookies at seven days. If your conversion window is longer than a week -- and for most B2B and considered purchases, it is -- a meaningful share of conversions on Safari and Firefox will not be attributed back to the click. The algorithm never learns from them.

Consent gaps. In the EU, a user who declines tracking cookies will not have a Google Ads conversion tag fire. That is legally correct, but it means your conversion data is systematically biased toward users who consent. Google's consent mode with modeling partially fills this gap, but modeled conversions carry less signal than observed ones.

Tag failures. Tags that do not fire on certain pages, fire twice on others, or send the wrong value corrupt the signal that Smart Bidding depends on. If your value-based bidding performance is erratic, the first place to look is your conversion tracking setup -- not your bid targets. I cover the most common breakages in the Google Ads conversion tracking guide.

The fix for cookie loss and consent gaps is server-side tracking and enhanced conversions, which use hashed first-party data (email, phone) as a matching key instead of relying solely on cookies. For B2B accounts importing CRM data, enhanced conversions for leads can recover attribution on conversions that would otherwise be invisible.

4. Conversion value rules must be configured correctly

Google Ads offers conversion value rules that let you adjust reported values based on audience, location, or device. For example, you might tell Google that conversions from returning customers are worth 1.5x the reported value, or that leads from a specific geography close at a higher rate.

These rules are powerful but dangerous. If misconfigured, they distort the signal that Smart Bidding uses to allocate budget. I have seen accounts where a value rule intended to boost high-LTV segments was inadvertently applied to all traffic, inflating reported conversion value across the board. The ROAS numbers looked amazing. The actual revenue did not change.

How to Know if Your Tracking Is Ready

Before enabling value-based bidding, run through this checklist:

• Are your conversion actions sending real, variable values (not static placeholders)?
• Do you have at least 30 conversions with value in the last 30 days?
• Is your conversion value within 10 percent of actual revenue when you spot-check against your CRM or payment processor?
• Are you using enhanced conversions or server-side tracking to close the cookie gap?
• Have you checked for duplicate conversion actions (a common culprit when both a Google Ads tag and an imported GA4 event count the same conversion)?
• Are conversion value rules configured intentionally, or were they set up and forgotten?

If you cannot answer "yes" to all six, value-based bidding will not perform as expected. The algorithm will optimize on flawed data and make flawed decisions with your budget. If you are not sure where the gaps are, I can audit your setup and tell you exactly what to fix.

What Changes When You Get It Right

When the tracking is clean and the conversion values are accurate, value-based bidding does what it promises. The algorithm learns which keywords, audiences, times of day, and devices produce the highest-value conversions. It bids up for prospects that look like your best customers and bids down for the rest.

I have seen this play out repeatedly. The B2B SaaS company I mentioned at the top of this post -- after we built a proper offline conversion pipeline and switched their conversion action from "demo request" (static value) to "qualified opportunity" (actual pipeline value imported from HubSpot), their cost per qualified opportunity dropped from EUR 310 to EUR 195 within eight weeks. The algorithm was finally learning from the right signal.

The key insight about how value-based bidding works is that the algorithm itself is not the bottleneck. Google's Smart Bidding is sophisticated. The bottleneck is almost always the data you feed it. Fix the inputs, and the strategy works. Feed it garbage, and it optimizes for garbage -- efficiently.

FAQ

Not sure your conversion tracking is ready for value-based bidding? Get in touch -- I will audit your setup and tell you exactly what needs to change before you flip the switch.