NOTE: I shared my opinion on this process (more like I vented) in the "HEAR ME OUT" section below. Feel free to skip ahead to "How to use this."

HEAR ME OUT

You need to understand how your ideal customer behaves online, how they search, how they engage, and where they show up across platforms throughout the year. But here is the part most teams overlook: as an organization, you already have the data to find them. Your CRM holds the answer. Every closed-won deal, every SQL, every MQL is a signal map of exactly who you are trying to reach more of.

The problem is that this data rarely makes its way back to the platforms that do the targeting. And sometimes it does not provide the best picture due to the data entered or selected. Instead, teams spend cycles managing keyword lists, fielding requests from stakeholders and C-Suite leaders to bid on niche terms that do not reflect how buyers actually search, and debating channel investments based on surface metrics rather than pipeline reality. Keywords matter, but they should not be the primary lever.

The best targeting audiences are built on offline conversion data, fed back into platforms that use AI and ML to find more of the right people at scale. That is the shift this guide is about.

Two things tend to block this in practice. The first is data privacy concerns. When senior leadership hears "share your CRM data with an ad platform," the instinct is to say no. The reality is that data sent via Enhanced Conversions or Customer Match is SHA-256 hashed before it ever leaves your system. Google never sees a name or an email address, only an irreversible string of characters that can be matched against their own hashed data. The risk is not zero in theory, but it is as close to zero as any data transfer gets. Organizations blocking this on privacy grounds are leaving one of their highest-leverage growth levers on the table.

The second blocker is operational. There is nothing more demoralizing than doing the coordination work to get Enhanced Conversions or value-based bidding live, watching performance improve, and then seeing it collapse because a Salesforce Administrator removed a field the platform was mapping to without telling anyone, because he wasn't aware of it, because YOU were too excited to get this effort running and thought Mops would be enough. Clean signal infrastructure requires cross-functional alignment, not just marketing execution. That is a revenue operations problem that will impact the growth function, so make sure to strategize on how to brand it and get buy-in.

This checklist covers both the strategic setup and the operational discipline required to make AI-native Google Ads infrastructure actually work.

How to use this

Each pillar below covers the why, a TL;DR, setup steps, and sourced links. Work through them in order -- each pillar builds on the one before it. Pillars 1 through 3 are prerequisites. Do not attempt value-based bidding or Performance Max until your conversion foundation and offline signal infrastructure are in place.

The interactive checklist at the bottom lets you track your progress across all six pillars.

Pillar 1: Conversion foundation

The why: Optimizing toward form fills does not equate to optimizing toward the pipeline. The algorithm is only as good as the signal you feed it. Most B2B SaaS accounts have too many conversion actions, the wrong ones set as primary, and attribution windows that are too short for a 60 to 90 day sales cycle. Fix this before anything else -- every other pillar depends on it. It doesn't have to be a perfect mirror or exact, but it has to reflect the buyer journey and goals assigned to you or your team.

TL;DR: Before anything else, make sure you are measuring the right events and that your conversion data is clean, deduplicated, and mapped to real revenue outcomes.

Estimated time: 1 to 2 days for setup. Allow 48 hours for conversion data to populate in reports before making any optimization decisions.

Stakeholders: Paid search manager, marketing ops, web or engineering team (Google Tag or GTM access required).

Setup:

Enable Google Ads conversion tracking natively -- do not rely solely on GA4 imports (different attribution models, fewer conversions captured)
Audit your conversion actions: remove duplicates, suppress any action that fires on irrelevant events (page views, PDF downloads without intent)
Set your primary conversion action to the event closest to revenue that has sufficient monthly volume (30+ per month minimum for Smart Bidding to stabilize)
Set conversion windows to match your actual sales cycle -- up to 90 days for click-based B2B conversions
Use data-driven attribution as your model -- it distributes credit across touchpoints rather than assigning it all to last click

Sources:

Pillar 2: Enhanced Conversions

The why: Standard cookie-based tracking is losing signal to Safari ITP, Firefox ETP, iOS privacy changes, and GDPR consent requirements. Every untracked conversion is a missed signal that weakens Smart Bidding's model. Enhanced Conversions closes that gap by sending hashed first-party data alongside your existing tags -- without exposing any raw PII.

TL;DR: Enhanced Conversions sends SHA-256 hashed first-party data (email, phone) alongside your existing conversion tags so Google can match conversions even when cookies fail.

On data privacy: SHA-256 is a one-way hashing function. Google never receives raw PII -- only an irreversible string of characters. Google then compares that hash against hashes it already holds from signed-in Google accounts. There is no key, no reversal, no exposure of raw data. Blocking Enhanced Conversions on privacy grounds is blocking your highest-leverage signal without reducing risk. If you are having to create a proposal for SLT or your manager make sure to include and clearly outline the first 5 sentences of this paragraph!

Estimated time: A few hours using GTM client-side method. Allow 48 to 72 hours for the diagnostics tab to populate after setup.

Stakeholders: Paid search manager, marketing ops, web engineer (only required if GTM access is restricted or a data layer push is needed).

Setup:

Accept Google's Enhanced Conversions customer data terms in your Google Ads account (Goals > Conversions > Settings)
Choose your implementation method: Google Tag Manager (recommended for most teams), Google tag (gtag.js), or Google Ads API for high-volume setups
Configure user-provided data variables in GTM: capture email, phone, first name, last name as separate variables from your conversion page
Enable Enhanced Conversions for Leads if you are a B2B team with offline sales cycles -- this captures lead data at form submission for later offline matching
Verify via the Diagnostics tab in Goals > Conversions -- check match rate; anything below 25% after one week indicates a configuration problem

Sources:

Pillar 3: Offline Conversion Import (OCI)

The why: Your ICP is in your CRM. Every closed-won deal, every SQL, every MQL is an insight into exactly who you are trying to find/capture more of. OCI closes the gap between ad click and revenue by feeding downstream CRM events back to Google -- so Smart Bidding optimizes toward outcomes that connect to the pipeline, not form fills. This is the single highest-leverage move in a B2B paid search program. If you are curious about what might happen if you don't set this up properly, read my little vignette, "HEAR ME OUT."

TL;DR: OCI closes the gap between ad click and revenue by feeding downstream CRM events back to Google so Smart Bidding optimizes toward pipeline, not form fills.

Before you proceed: Google recommends at least 30 offline conversions per month for tCPA to stabilize, and 50+ for tROAS. Below that threshold, use a mid-funnel proxy event (MQL, qualified opportunity) rather than closed-won as your primary signal. Also, make sure your Salesforce admin knows which fields the integration maps to before you go live.

Estimated time: HubSpot native connection takes approximately 15 minutes. Salesforce requires a custom GCLID field and workflow build -- plan for 1 to 3 days with Salesforce admin involvement. Smart Bidding typically needs 2 to 4 weeks to learn from new conversion data, with performance improvements visible in 4 to 8 weeks.

Stakeholders: Paid search manager, marketing ops, Salesforce or HubSpot admin (required), sales ops (for deal stage and field mapping alignment).

Setup:

Enable GCLID auto-tagging in your Google Ads account settings -- this is required for OCI to work and is now on by default for new accounts
Add a hidden GCLID field to every lead form on your site and configure your CRM (Salesforce, HubSpot) to capture and store it on lead creation
Create a conversion action in Google Ads for your target CRM stage: qualified opportunity is the recommended primary event for most B2B SaaS teams
Set your conversion window to 90 days minimum to account for B2B sales cycle length -- extend to 180 days for enterprise ACV tiers (recommended)
Upload conversions via CSV (manual, minimum weekly) or via native CRM integration -- HubSpot and Salesforce both have direct Google Ads connectors
Assign conversion values by pipeline stage to enable value-based bidding: e.g. MQL = $100, SQL = $900, Opportunity = $3,000, Closed-Won = actual ACV

Sources:

Pillar 4: Value-based bidding

The why: tCPA treats a $5K ACV trial and a $500K enterprise opportunity as equally valuable. Value-based bidding tells the algorithm which conversions are worth more -- so it finds more of the right ones, not just more of any kind. This only works if Pillar 3 is in place and your conversion values are populated accurately. Without real values flowing in, tROAS is optimizing against noise.

TL;DR: Value-based bidding tells the algorithm which conversions are worth more so it finds more of the right ones, not just more of any kind. Google's own data shows a median 14% increase in conversion value when moving from tCPA to tROAS.

Estimated time: Bid strategy configuration is same-day once conversion values are assigned. The learning period runs 4 to 8 weeks -- do not evaluate performance or make changes during this window.

Stakeholders: Paid search manager, marketing ops (conversion value configuration), finance or revenue ops (to validate deal value inputs per pipeline stage).

Setup:

Start with Max Conversions (no target) to build conversion data -- do not attempt tCPA or tROAS before hitting 30 conversions per month
Once at 30+ conversions per month, move to tCPA set at or near your historical actual CPA -- adjust in 10 to 15% increments, never make large jumps
Configure tiered conversion values mapped to your pipeline stages before enabling tROAS -- without values, tROAS cannot function correctly
Transition to Maximize Conversion Value, then introduce a tROAS constraint once value data has accumulated for 4+ weeks or 3 conversion cycles
Set tROAS 20% below historical average ROAS to give the algorithm room to learn -- tighten gradually as performance stabilizes (recommended)
Never switch bid strategy and make major creative or landing page changes simultaneously -- isolate variables so you know what moved performance

Sources:

Pillar 5: Performance Max for B2B

The why: PMax is a complement that finds audiences earlier in the funnel using Google's ML across all inventory. The default setup optimizes for cheap form fills. Properly configured with offline conversion data and CRM audience signals, it extends your reach beyond keyword-triggered intent into audiences that match your closed-won profile. But it only works if pillars 1 through 4 are already in place.

TL;DR: PMax is a lever that finds audiences earlier in the funnel. Feed it the right signals or do not run it. Search should remain 60 to 90% of your budget.

Before you proceed: Do not launch PMax before your Search campaigns are fully funded and your conversion foundation (Pillars 1 through 3) is in place. PMax needs your existing conversion data to learn from. Without it, it optimizes for volume over quality -- and median budget waste in unguarded B2B SaaS PMax accounts runs high.

A note on this pillar: I have not personally run Performance Max. The guidance here is grounded in current research and practitioner consensus, not direct execution. I will update this section with firsthand observations once I have had the opportunity to run and measure a PMax program in a B2B context. If you have run PMax for B2B lead generation and want to compare notes, reach out via LinkedIn.

Estimated time: 1 to 2 days to configure campaigns and asset groups. Full optimization cycle typically takes 2 to 3 months for B2B as closed-deal data accumulates and the algorithm refines its targeting.

Stakeholders: Paid search manager, creative or design team (asset groups require headlines, descriptions, images, and video), marketing ops (Customer Match list export), optional: ABM team for account-level audience signals.

Setup:

Create your audience signal stack before launch: closed-won Customer Match list (strongest signal), high-intent custom segments, pricing and demo page visitor lists
Build asset groups by funnel stage and buyer persona -- not by product feature; each group needs distinct messaging, headlines, and creative
Add campaign-level negative keywords immediately (up to 10,000 as of March 2025) -- exclude informational queries, job seekers, free and open source terms
Turn off Final URL Expansion unless your site has strong, relevant landing pages for every possible query path -- for most B2B teams, turn it off
Set brand exclusions to prevent PMax from cannibalizing your existing brand Search campaigns and inflating overall account metrics
Allow 8 to 10 weeks before evaluating steady-state performance -- B2B PMax learning runs longer than ecommerce because offline conversion events take longer to import

Sources:

Pillar 6: Dynamic landing pages and Customer Match

The why: A strong ad followed by a generic landing page is simply a great trailer with a mediocre movie. Message match between ad copy and landing page is one of the highest-leverage CRO moves in paid search. It improves Quality Score, lowers CPC, and increases conversion rate simultaneously. Think of it as you forcing the prospect to make a decision, which they shouldn't! It should be a continuous, smooth process (great reads: Don't Make Me Think and Buyology). Customer Match activates your CRM data as a targeting layer, letting you suppress existing customers, bid up on target accounts, and seed lookalike models from closed-won data. Your exclusion lists should be as good, if not better than, your targeting audiences!

TL;DR: Dynamic pages match headline, copy, and CTA to the query intent and audience segment. Customer Match activates your CRM as a targeting layer -- suppress existing customers, bid up on target accounts, and seed lookalikes from closed-won data.

Estimated time: 1 to 2 weeks, depending on the landing page tool, the number of intent clusters, and whether web engineering support is required. Customer Match list upload is same-day once the list is exported from your CRM.

Stakeholders: Paid search manager, web or engineering team (URL parameter implementation), creative or design team (landing page variants), marketing ops (Customer Match list export and CRM segmentation).

Setup:

Implement URL parameters (?keyword={keyword}&segment={adgroup}) to pass query and audience data from ads to landing pages
Use a dynamic landing page tool (Unbounce, Mutiny, or custom implementation) to swap headline, subhead, and CTA based on URL parameters
Create dedicated landing pages per major intent cluster -- one page per use case or buyer segment outperforms a single generic page every time
Upload your closed-won Customer Match list to Google Ads -- this is your highest-quality seed for lookalike modeling and bid adjustments
Use Customer Match in Observation mode for broad campaigns (lets Smart Bidding adjust bids for matched users without capping reach) and Targeting mode for ABM and retargeting campaigns
Suppress existing customers from new-customer acquisition campaigns via Customer Match exclusion list -- required to protect CAC efficiency

Sources:

Interactive checklist

Use the checklist below to track your progress across all six pillars. Come back as you complete each step.

What to read next

The keyword work sits alongside this infrastructure, not before it. Once your conversion signals are clean and your bidding is anchored to pipeline outcomes, keyword strategy becomes sharper -- you know which queries actually drive revenue and which ones just drive volume.

Build smarter keyword clusters with the B2B Paid Search Keyword Planner

HEAR ME OUT

How to use this

Pillar 1: Conversion foundation

Pillar 2: Enhanced Conversions

Pillar 3: Offline Conversion Import (OCI)

Pillar 4: Value-based bidding

Pillar 5: Performance Max for B2B

Pillar 6: Dynamic landing pages and Customer Match

Interactive checklist

What to read next

Notes from Page 2