Paid Media // VBB & tROAS // Reference v1 // 2026

Value-Based Bidding & Target ROAS

The math, the process, and the conversion setup behind value-based bidding, plus a calculator that turns your margins and your actuals into a target you can actually hit.

The one idea

A Target ROAS is a constraint, not a wish.

It tells Google the lowest return you will accept, then the algorithm skips the auctions that would not clear it. Base the target on your actual margins and returns. Everything starts from one line: break-even ROAS = 1 / your gross margin. Price the target above that for profit, near your actuals for volume.

00

The model

Smart Bidding optimizes toward whatever you tell it to want. Count-based strategies want more conversions. Value-based strategies want more conversion value. The whole game is teaching the machine which conversions are worth more, then deciding how hard to push for return.

VBB

Value-based bidding

An umbrella for strategies that optimize to conversion value, not conversion count. The two members are Maximize Conversion Value and Target ROAS.

Max Conv Value

Spend the budget, chase value

Get the most total conversion value within a fixed budget. No return target. This is the entry point and the data-building phase for VBB.

tROAS

Target ROAS

The same engine with a return floor bolted on. You name the return you need, Google bids to hold it. It is a setting on Max Conversion Value, not a separate thing.

The tradeoff

Return versus volume

Raise the target and Google walks away from cheaper, lower-value auctions. You get a higher return on less spend. Lower it and volume returns. There is no free lunch.

Count or value Use count-based bidding (Maximize Conversions, Target CPA) when every conversion is worth about the same and you just want more of them at a price. Switch to value-based bidding when conversions carry different worth: an enterprise deal versus a trial, a closed lease versus a tire-kicker, a $400 cart versus a $40 one. If the values vary, count-based bidding is leaving money on the table.
01

The math

Six formulas carry the entire topic. ROAS is usually shown as a percent (5:1 is 500%) or a ratio (5:1). The calculator below runs all of these for you, but know them cold.

ROAS
ROAS = Conversion value / Ad spend

Revenue (or assigned value) returned per dollar spent. $5,000 value on $1,000 spend is 5:1, or 500%.

Break-even ROAS
Break-even = 1 / Gross margin

The floor. At a 40% margin you need 2.5:1 (250%) just to cover product cost and ad cost. Below it, every sale loses money.

Target ROAS for profit
Target = 1 / (Gross margin - Profit goal)

Set the net profit margin you want to keep after ad spend. 40% margin, keep 10%: target = 1 / 0.30 = 3.33:1 (333%).

Cost of sale
Cost of sale = Ad spend / Revenue = 1 / ROAS

The mirror image of ROAS. If ads can be at most 25% of revenue, that is a 4:1 (400%) target. Useful when you think in spend ratios.

Lead value
Lead value = Deal value × Close rate

The proxy value of a lead action. A $5,000 deal closing 6% of the time makes that lead worth $300 to the bidding engine.

POAS, the advanced move
POAS = Profit / Ad spend

Bid to profit, not revenue, by passing margin-adjusted value (send profit as the value). Stops the algorithm chasing high-revenue, low-margin junk.

02

The target calculator

Plug in your numbers. Everything updates as you type. Start with break-even, set a profit-aware target, value your leads, then check whether you have the volume to run a target at all.

Read this first Your math target and your achievable target are not the same number. The formulas tell you what return you need for profit. Your account tells you what return it can deliver at the volume you want. Set the initial Target ROAS near recent actuals (10 to 20% above), then step toward the math target. Jumping straight to the dream number starves volume and the data dries up.
A · Break-even

Your ROAS floor

The lowest ROAS that does not lose money, from your gross margin.

%
B · Profit target

The target you need

The ROAS that leaves the profit you want after ad spend.

%
%
C · Spend ceiling

Cost of sale to ROAS

If you think in "ads can be at most X% of revenue."

%
D · Lead values

Value your lead actions

Proxy value to assign each conversion action for VBB.

$
ActionClose %Value
%
%
%
E · Readiness and stepping

Are you ready, and how do you get there?

tROAS needs data. This checks your volume against Google's thresholds and maps a stepped path from where you are to where you want to be.

%
%
03

Setting the target, step by step

The order matters. Skip the data step and you are bidding to value you never measured. Skip the actuals check and you choke volume on day one.

Get clean value flowing first

You cannot bid to value you do not measure. Confirm your primary conversion sends a value greater than zero. Google only counts conversions with a value above zero as eligible for tROAS. No value, no value-based bidding.

Find break-even

Break-even ROAS is 1 / gross margin. This is the floor. Any target below it is a decision to lose money on the marginal sale. Calculator A.

Price the profit goal

Decide the net margin you want to keep after ad spend. Target ROAS is 1 / (gross margin - profit goal). Or pick a cost-of-sale ceiling and invert it. Calculators B and C.

Pull your actuals

Look at achieved ROAS over the last 30 to 90 days. This is the reality your account delivers today. If your math target sits far above it, you have a gap to close gradually, not a number to flip a switch on.

Set the initial target near actuals

Start about 10 to 20% above current achieved ROAS, not at the aspirational number. The algorithm needs room to operate. An aggressive target out of the gate forces restricted bidding and kills volume.

Leave it alone through learning

Expect a one to two week learning period after you set or significantly change the target. Performance wobbles, then settles. Every change resets the clock, so resist the urge to tinker.

Step toward the goal

Move in roughly 10 to 20% increments, one move at a time, watching what happens to both volume and value. Calculator E maps the steps. Climbing in stages keeps the machine learning instead of stalling.

Match the window to your sales cycle

Read ROAS over a window that excludes the recent conversion-delay period, or recent days look artificially weak because the conversions have not landed yet. If you just started reporting value, wait about four weeks or three conversion cycles before adopting tROAS.

04

The maturity ladder

Which strategy belongs at which stage, and what has to be true before you climb. You do not start at Target ROAS. You earn your way there with data.

0
Count first / no value data yet
Strategy: Maximize Conversions, then Target CPA. Use when conversions are roughly equal value or value tracking is not built. The job here is to gather signal and clean the data.
Graduate when: value greater than zero is firing reliably on your primary action.
1
Assign worth / static values
Strategy: Maximize Conversion Value, no target. Give each conversion action a fixed value that reflects relative worth. This is the front door to VBB. You are teaching the machine that a demo beats a newsletter signup.
Graduate when: value is stable and you are clearing volume (15 minimum, 50 ideal per 30 days).
2
Differentiate / value rules and weights
Strategy: Maximize Conversion Value with value rules layered on. Adjust value by device, location, audience, or new versus returning customer. In Floodlight platforms, weight the activities. Sharper signal, same engine.
Graduate when: differentiated values are trusted and volume holds at 50 plus.
3
Constrain to a return / Target ROAS
Strategy: Target ROAS. Once value-bearing conversions are stable and volume clears, add the return floor. Best in class: import real closed-won values so the target chases what actually closed.
Maintain: review monthly, step the target, match the window to the sales cycle.
05

Conversions and Floodlight, set up right

VBB performs to the quality of the value you feed it. This is where most accounts quietly sabotage themselves: counting the wrong actions, carrying no value, or double counting across platforms.

Eligibility

Value greater than zero

A conversion only feeds value bidding if it carries a value above zero. A primary action reporting no value is invisible to tROAS. Confirm this before anything else.

Primary vs secondary

Only primary bids

In Google Ads, only Primary conversion actions feed bidding. Keep soft proxies like page views and form starts as Secondary so they inform you without distorting value.

Counting

One per click for leads

Lead gen counts one conversion per click so a single person filling three forms is not three leads. Ecommerce counts every conversion and passes the real transaction value.

Attribution

Data-driven, right window

Use data-driven attribution, not last click. Set the conversion window to match your sales cycle so credit lands where it belongs and the target reads true.

Floodlight in SA360 and DV360 Floodlight activities come in two flavors. Counter activities count actions (pages viewed, leads). Sales activities count transactions and carry revenue. For value-based bidding you need value in the signal, so either use Sales Floodlights carrying real revenue, or in the new SA360 assign a static monetary value to one or more counter conversion actions. Formula columns let you build custom weighted metrics from Floodlight activities (for example, upweight revenue from new customers), which you can then bid on. In DV360, weighted values must be whole numbers, the option exists for both Sales and Counter tags, and you pick conversion weights or sales values, not both. When weighting, attach two or more Floodlight activities so the algorithm has something to optimize between. One more sequencing rule in DV360: start a line item on Target CPA before moving it to Target ROAS, and on Maximize Conversions before Maximize Conversion Value.
Calculate correctly Do not double count. Floodlight deduplicates within a shared configuration, but values pulled into two strategies or mixed Google Ads and Floodlight setups can inflate ROAS and trick you into a target that is too high. Exclude the recent conversion-delay window when reading results. And if you just turned on value reporting or changed how value is calculated, include the new values in the Conversions column and wait about four weeks or three conversion cycles before adopting a target, so the rate stabilizes first.
06

Value mapping for lead gen

Ecommerce hands you the value: it is the cart total. Lead gen does not. A form fill has no revenue attached, so you assign a proxy value that reflects what that lead is worth downstream. This is the step that makes VBB work for services, B2B, home, auto, and finance.

The method Proxy value of a lead action equals average closed-deal value times the rate at which that action closes. A deeper action closes more often, so it earns a higher value. Build a ladder, not a flat value. Google's Conversion Value Calculator does exactly this arithmetic if you would rather not. Calculator D above builds the ladder for you.
A worked value ladder ($5,000 average deal)
Conversion action Closes to a deal Proxy value Why
Newsletter signup 1% $50 Barely qualified. Tiny value keeps it in the signal without distorting bidding.
Qualified lead 6% $300 Real intent. The workhorse action for most lead-gen accounts.
Tour or demo booked 18% $900 High intent. Worth bidding up for, three times a raw lead.
Application started 45% $2,250 Bottom of funnel. The algorithm should chase these hardest.
Static is the start, offline is the destination Static proxy values get you bidding to value today. The bigger step is offline conversion import: send the actual closed-won values back to Google so Smart Bidding learns which clicks became revenue, not which clicks became forms. For considered-purchase categories, teaching the system which leads actually closed and at what value is the one change that helps most. Enhanced Conversions for Leads makes the match durable as cookies fade.
07

Decision tables

The settings people get wrong most often, in one place.

Strategy by model and stage
Business model Objective Start with Then move to
Ecommerce Revenue, ROAS Maximize Conversion Value Target ROAS once value data is stable and volume clears
Lead-gen B2C
(home, auto, finance, services)
Qualified leads at a return Maximize Conversions Maximize Conversion Value with proxy values, then Target ROAS plus offline import
B2B / SaaS Pipeline value Target CPA on qualified leads Target ROAS on imported deal values once the cycle closes enough volume
New account, no data Gather signal Maximize Conversions with a sane budget A target only after the learning phase clears and value is firing
Volume you need before setting a target
Campaign type Documented minimum For stability Note
Search and Display ~15 value conversions / 30 days 50+ / 30 days Per campaign, not per account. Below 15, stay on Max Conversion Value.
Shopping ~20 conversions / 45 days 50+ / 30 days Threshold is lower, but richer data still wins.
Video Action ~30 conversions / 30 days Higher Floodlight optimization must be enabled in DV360.
Learning period 1 to 2 weeks after any target change Hold steady Each change resets the clock. Do not tinker mid-learning.
Three ways to land on the number
Method You provide Formula Example
Break-even Gross margin 1 / margin 40% margin to 250%
Profit target Margin + profit goal 1 / (margin - goal) 40% margin, keep 10% to 333%
Cost of sale Max ad spend share 1 / cost of sale 25% ceiling to 400%
08

Money wasters

Every one of these is common, and every one quietly drains spend or freezes volume. None of them announce themselves.

Setting the dream number on day oneAn aggressive target out of the gate forces restricted bidding. Volume collapses, data thins, performance gets worse. Start near actuals, step up.
Bidding to value you never cleanedIf value greater than zero is not actually flowing, tROAS has nothing to optimize. Verify the signal before you trust the strategy.
Letting soft conversions carry valuePage views and form starts as primary, value-bearing actions teach the machine to chase noise. Keep them secondary.
Changing the target mid-learningEvery adjustment resets the one to two week clock. Constant tweaking means the algorithm never finishes calibrating.
One static lead value foreverA fixed proxy is only meant to be temporary. Without offline import, the system never learns which leads actually closed.
Window shorter than the sales cycleRead ROAS before conversions land and recent days look broken. Match the evaluation window to how long deals take.
Same value on every actionIf a newsletter and a demo carry the same value, you have told Google they are equally good. They are not.
Bidding to revenue with thin marginsHigh-revenue, low-margin products can hit a revenue target and still lose money. Bid to profit (POAS) when margins vary by product.
09

Resources

Official documentation first. Thresholds and feature names change, so confirm against release notes before you rely on a number.

Google Ads

  • About Target ROAS biddingEligibility, value greater than zero rule, per-campaign-type minimums.
  • About value-based bidding and value rulesAdjusting value by device, location, audience, customer type.
  • Conversion Value CalculatorTurns deal value and close rate into a proxy lead value.
  • Enhanced Conversions for LeadsDurable lead matching as cookies fade.
  • Import offline conversions (OCI)Send closed-won values back so bidding chases real outcomes.
  • About data-driven attributionThe default model. Set the right conversion window.

SA360 and DV360

  • SA360: How Floodlight counts conversionsCounter versus Sales activities, counting methods.
  • SA360: Formula columnsCustom weighted values from Floodlight activities for bidding.
  • DV360: Value based bidding strategiesWhole-number weights, Sales or Counter, the start-on-tCPA rule.
  • DV360: Custom biddingScripted, weighted Floodlight goals for advanced control.

Sanity checks

  • Calculate break-even before any target1 / gross margin. The floor under everything.
  • Compare math target to achieved ROASStep the gap gradually. Day one stays near current ROAS.
  • Confirm value greater than zero is firingNo value, no value-based bidding. Check it first.
  • Cross-check platform ROAS against GA4 and blendedPlatform-reported return runs optimistic. Verify.