Loop Optimizations
All posts

Value-Based Bidding: How It Works and How to Start

October 25, 2026
Value-Based Bidding: How It Works and How to Start

Value-based bidding is an automated bidding approach that sets what you pay in each ad auction according to the predicted value of a conversion, such as revenue, profit or lifetime value, instead of treating every conversion as equal. You hand the platform a number for each conversion, and its machine learning bids more for the clicks likely to produce more of that number. This guide covers how it works, the two Google Ads strategies, the data you need, how to set it up for ecommerce and lead generation, and how to tell whether it is paying off.

TL;DR

  • Value-based bidding optimizes for the total value of conversions, not the count, so a $900 order counts for more than a $30 order.
  • In Google Ads the two value-based strategies are Maximize conversion value and Target ROAS.
  • It only works as well as the values you feed it. Flat or inaccurate values turn it back into conversion-count bidding.
  • Expect conversion volume to shift. Judge it on revenue, profit and lead quality, not on cost per conversion alone.
  • Start with clean tracking and a realistic ROAS target, then test changes one at a time.

Table of Contents

What value-based bidding actually is

Most advertisers begin with volume bidding. You tell the platform to get as many conversions as possible for your budget, or at a target cost per conversion. That is a fine goal when every conversion is worth about the same. It breaks down when they are not.

Take a store where one shopper buys a $25 accessory and another buys a $400 jacket. Under volume bidding, both purchases are one conversion. The algorithm has no reason to prefer the second shopper, so it happily chases whichever clicks are cheapest. Value-based bidding fixes this by setting auction prices on the predicted value of a conversion, such as revenue, profit or lifetime value, rather than only the likelihood that a conversion happens.

The same logic applies to lead generation. A demo request from a 5,000-person company and a newsletter signup from a student are both “leads.” If you can tell the platform which one is worth more, it can learn to find more of the first kind.

If you are weighing whether to run this in-house or hire help, it is worth knowing what outside support costs. Paid media agency costs in the UK give a useful picture of typical fee structures and budgets, though your own market will differ.

What value-based bidding actually is — overview diagram

How it works at auction time

Every time someone triggers an ad auction, the platform has milliseconds to decide what to bid. In a value-based setup, two predictions drive that decision. The system estimates the chance of a conversion and the value it is likely to carry, and the strategy then decides how much of that expected value to spend.

In plain terms:

  1. You report conversion values back to the platform through your tracking.
  2. The machine learning studies those values against the signals it has about each click, such as device, location, time of day, audience and query.
  3. For each new auction, it predicts what that click is likely to be worth.
  4. It bids more for high-value predictions and less for low-value ones, within your budget and target.

The key point is that the model learns only from what you tell it. If your values are accurate and varied, it can find patterns you would never set up by hand. If they are missing or identical, there is nothing to learn.

Value-based vs. conversion-based bidding

Conversion-based bidding Value-based bidding
Optimizes for Number of conversions Total conversion value
Treats all conversions as Equal Different, by value
Best for Uniform order values or lead worth Varied order sizes, margins or customer worth
Main input Conversion events Conversion events plus a value for each
Typical Google Ads strategies Maximize conversions, Target CPA Maximize conversion value, Target ROAS
Common risk Cheap, low-quality conversions Fewer conversions if values are noisy or targets are too tight

Neither is better in the abstract. Conversion-based bidding is simpler and needs less data. Value-based bidding is more powerful when your conversions differ in worth and you can measure that difference.

The two value-based strategies in Google Ads

Google Ads offers two Smart Bidding strategies built on value, and they differ in how they treat your budget and your return target.

Maximize conversion value

This strategy tries to extract the most total value it can from a fixed budget. You set the budget, and the system spends it to bring in as much conversion value as possible. It does not aim for a specific return, so it suits campaigns where you want to grow revenue and are comfortable letting the system spend the full budget.

Target ROAS

Target ROAS (return on ad spend) optimizes for value at a specific return you set, expressed as a percentage. If you set 400%, you are asking for roughly $4 in conversion value for every $1 spent. It gives you more control over efficiency, but it also has data requirements, and a target set too high can choke volume.

Which one to start with

If your campaign is new to value bidding, a common path is to begin with Maximize conversion value to build a baseline, then move to Target ROAS once you can see what return your account really delivers. Setting a target before you know your historical ROAS is how many accounts end up with limited delivery. Check the current eligibility notes in Google’s own documentation, since thresholds and options change over time.

Beyond Google: value optimization on Meta, TikTok and others

Value-based bidding is not a Google-only idea. It exists as a paid-media strategy across platforms such as Google Ads and Meta, and other networks, including TikTok, offer their own value-based optimization for web purchases. The shape is the same everywhere: send purchase values back, give the platform enough conversion history to learn, and let it prioritize likely high spenders.

The details differ. Each platform has its own minimum event volumes, its own way of receiving values, and its own naming. Before you switch a campaign, check the current requirements for that platform, because they are the most common reason a value setting is greyed out.

If you run several channels, keep your value definitions identical across them. When Google counts revenue, Meta counts margin and your CRM counts lifetime value, you cannot compare results, and your attribution turns into an argument instead of an answer.

What data you need before you switch

Value-based bidding fails quietly when the inputs are weak. Before changing a strategy, confirm three things.

Values are being passed. You need to pass specific conversion values through tracking tags, data feeds or offline imports. A conversion action that fires without a value, or with a placeholder like 1, gives the model nothing to work with.

There is enough history. The machine learning needs sufficient historical data to learn effectively. A brand-new account or a campaign with a handful of conversions a month will bid erratically.

The values are honest. The strategy depends heavily on data quality: if inaccurate or flat values are passed, it can degrade into conversion-count bidding. Review your values the way you would review a financial report. Do refunds subtract? Is tax excluded? Does shipping count?

Three checks for value bidding readiness

Browser-based tracking also loses signals to ad blockers and privacy settings, which means the platform sees fewer of your real conversions and values. Moving to server-side tracking recovers some of them. Our guide to server-side conversion tracking explains how it works and how to set it up.

Setting it up for ecommerce

Ecommerce is the easiest place to start because every order already carries a number.

  1. Send the transaction value with each purchase. Use the order total, or better, the order total minus shipping and tax. Include the currency.
  2. Decide on revenue or margin. Revenue is simpler. Margin is more accurate if your products vary a lot in profitability. If a $200 item earns you $20 and a $60 item earns you $35, revenue-based bidding will chase the wrong one.
  3. Handle returns. If returns are high, import adjustments so the platform sees net value, not gross.
  4. Pick the strategy. Start with Maximize conversion value for a baseline, then test Target ROAS.
  5. Set the target from history. Look at your actual ROAS over the last 30 to 90 days and set the first target close to it. Tighten gradually.
  6. Leave it alone while it learns. Frequent edits to budgets, targets and creatives reset the learning and muddy your reading.

Setting it up for lead generation

Lead generation is harder because the value usually arrives weeks after the click, if at all. You have to create a value signal. The practical options are to use lead scores, lead-form answers, profit margins or predicted lifetime value.

A workable approach for a B2B team:

  • Assign values by lead type. For instance, a demo request is worth more than an ebook download. Base the numbers on your close rates and average deal size, not on guesses.
  • Use form answers. Company size, budget range or timeline can each map to a different value.
  • Import offline outcomes. When a lead becomes an opportunity or a closed deal, send that back with its real value so the model learns from outcomes, not form fills.
  • Keep the scale sensible. Values do not need to be exact dollars, but the ratios between them should reflect reality. A lead worth ten times more should carry ten times the value.

The benefit here is lead quality. Value-based bidding can bring higher-quality leads and incremental revenue and profitability, but you pay for that with fewer total leads in many cases.

Benefits, limits and common mistakes

What you gain

  • Budget flows toward the clicks most likely to produce revenue or profit.
  • Reporting ties more closely to business results than to raw conversion counts.
  • Lead quality tends to improve when values reflect real worth.

What you give up

A move away from volume bidding may change your total conversion volume or reach, and Target ROAS has data thresholds. If your boss measures success by lead count, set expectations first. A drop in conversions with a rise in revenue is a good outcome, but only if everyone agrees on that going in.

Mistakes that sink campaigns

  • Passing the same value for everything. The model cannot rank what looks identical.
  • Setting an aggressive ROAS target on day one. The system responds by bidding too low to win auctions.
  • Changing too much at once. Strategy, budget, creative and landing page all moved? You cannot tell which one helped.
  • Judging too early. Give the algorithm time to learn before you make calls.
  • Ignoring the creative and the page. Bidding decides who sees your ad. It does not decide whether the ad or the landing page convinces them.

That last point deserves attention. Smart bidding finds the likely buyer, but the message and page still have to close. If you are not testing them, you leave half the gain on the table. Our practical guide to ad creative testing covers how to find winning variants.

How to measure whether it is working

Judge value-based bidding on the metrics it was built to move.

  1. Track total conversion value and ROAS, not cost per conversion. Cost per conversion often rises, because you are paying more for better clicks.
  2. Compare against a baseline. Use a before-and-after window of equal length, or run the new strategy as a campaign experiment against the old one so seasonality does not fool you.
  3. Use bid simulators. Smart Bidding simulators help you gauge the trade-off between volume and value and find a ROAS target that delivers the volume you want.
  4. Follow value downstream. For lead generation, check how many leads become pipeline and revenue, not just how many forms were filled.
  5. Watch volume. Confirm the drop, if any, is acceptable against the revenue gain.

A fair test runs long enough to cover a full buying cycle and enough conversions for the difference to be more than noise.

Where Loop Optimizations fits

Bidding is one lever. What the platform learns depends on the signals you send, and what a click is worth depends on the ad and page it lands on. Loop Optimizations is an AI platform that continuously analyzes your website, launches and tests multiple ad variants across Google, Meta, TikTok and LinkedIn, rewrites pages to match the promise in each ad, attributes full customer journeys and feeds performance signals back to the ad platforms. You get a closed loop in which creative, landing pages and signals improve together, with no code or engineering setup.

Pricing is published: Website Optimization is $497 per month, each connected ad or email channel is $297 per month per channel, and the Agency plan for multi-client management is $499 per month. See the pricing page for what is included, or read how we run tests on the methodology page.

A simple 30-day starting plan

  • Week 1: audit. Check that every conversion action carries a real value, that currencies match and that refunds are handled. Fix gaps before touching bidding.
  • Week 2: baseline. Record the last 30 to 90 days of conversion value, ROAS and volume. This is what you compare against.
  • Week 3: switch one campaign. Choose a campaign with steady conversions. Move it to Maximize conversion value, or to Target ROAS with a target near your historical return.
  • Week 4: read and adjust. Check value and ROAS against the baseline. If the result is positive, extend to the next campaign. If not, look first at data quality, then at the target, and only then at the strategy.

Keep a log of every change with its date. It is the only way to explain results later.

FAQ

What is the difference between Maximize conversion value and Target ROAS?

Maximize conversion value spends your budget to bring in as much total value as possible, with no specific return goal. Target ROAS aims for a return you define, such as 400%. The first favors growth, and the second favors efficiency control.

Can I use value-based bidding for lead generation?

Yes, as long as you can assign values to leads. You can base them on lead scores, form answers, margins or predicted lifetime value, and ideally import real outcomes from your CRM. Without differing values, the strategy has nothing to optimize toward.

Why did my conversions drop after switching?

That is common. Value-based bidding prioritizes higher-value users, so total conversion volume can fall even while revenue or lead quality rises. Compare total value and ROAS against your baseline before deciding it failed.

How much data do I need before starting?

Enough history for the machine learning to find patterns, and Target ROAS has its own thresholds. Exact minimums vary by platform and change over time, so check the platform’s current documentation. If conversions are scarce, begin with Maximize conversion value.

Sources

Put these ideas on autopilot.

Loop Optimizations writes, launches, and continuously tests your ads and landing pages so results compound every day.