Google Ads was built for e-commerce. Same-session conversions. Clear attribution. Fast feedback loops.
But your sales cycle is 6-18 months. By the time a deal closes, Google has long forgotten which click started it. The GCLID expired 90 days ago. Smart Bidding optimized for the wrong signals. And your CFO wants to know why you're still spending.
B2B sales cycles commonly stretch 6-12 months, with complex enterprise deals running even longer. Meanwhile, CPCs are rising nearly 13% year over year—the fifth consecutive year of increases—and only 2% of visitors convert on their first visit.
The platform's constraints work against you. This is how to make it work anyway.
The 90-Day Problem (And Why CRM Still Matters)
Google's attribution window maxes out at 90 days. Your deals close in 12 months. This creates a fundamental mismatch that breaks everything downstream: Smart Bidding can't learn from your actual revenue, in-platform reporting becomes fiction, and optimization feels like guesswork.
So why bother with CRM integration at all?
Because the GCLID expires, but your data doesn't have to.
The value of CRM isn't feeding data back to Google. It's building your own attribution layer that outlasts the platform's limitations.
Here's how it works:
Capture everything at form fill. When a lead converts, store the GCLID, all UTM parameters, the landing page, and the timestamp in your CRM record. HubSpot, Salesforce, and most modern CRMs support this natively or via hidden form fields.
Preserve the connection. That lead record now carries its origin data through your entire pipeline: to SQL, to opportunity, to closed-won. Twelve months later, when the deal closes, you know exactly which campaign, keyword, and ad produced it.
Run your own analysis. Export closed-won deals quarterly. Match back to campaigns. Calculate true cost per acquisition and revenue per campaign. This is your real performance data.
Google won't optimize to your closed deals. But you will. The CRM data informs your budget allocation, campaign decisions, and where to double down. You're building institutional knowledge that compounds over time.
The 90-day window is Google's problem. Don't make it yours.
What to Optimize To
You can't wait 12 months for closed-won data to train Smart Bidding. The algorithm needs feedback now. But if you only close 5 deals a month, there's not enough signal for machine learning to work.
The solution: optimize to a micro-conversion further up the funnel.
Think of your pipeline as a ladder:
Pick the conversion closest to revenue that still has enough volume. The rule of thumb: Smart Bidding needs 30-50 conversions per month per campaign to learn effectively.
If you have 5 closed deals a month, that's not enough. But you might have 25-30 SQLs. Optimize to that.
If SQLs are sparse, move up to SALs or discovery calls completed.
The tradeoff is real: the earlier the conversion, the more you sacrifice quality signal. You're trading precision for volume. But some signal beats no signal.
Enhanced Conversions for Leads lets you import CRM stage changes back to Google within the 90-day window. When a lead becomes an SQL on day 45, that data feeds back to the algorithm. It's not perfect, but it gives Smart Bidding better information than form fills alone.
The practical framework: start with SQLs if you have the volume. If performance degrades or you're starving the algorithm, move one step up. Monitor lead quality as you go.
How to Optimize
Strategy matters more than button-clicking. Here's what actually moves the needle.
Campaign Structure for Long Cycles
Separate campaigns by funnel stage intent, not just product line.
Bottom-funnel keywords like "enterprise software demo," "vendor comparison," and "pricing" deserve their own campaigns. These searchers are close to a decision. CPCs will be higher, but so will close rates. Optimize aggressively.
Top-funnel keywords like "how to improve X" or "what is Y" need patience. Different KPIs apply. You're building awareness and capturing emails, not expecting same-month pipeline. Set expectations accordingly.
Mixing intent stages in one campaign confuses the algorithm and muddles your reporting.
Smart Bidding Settings
Target CPA should be based on your micro-conversion, not your final sale. If you're optimizing to SQLs and an SQL costs $500, set tCPA at $500. Don't try to reverse-engineer closed-won economics into the platform.
Use conversion value rules to weight higher-quality leads. If enterprise leads are worth 3x SMB leads, tell Google. The algorithm can learn to favor them.
Set realistic targets. B2B CPAs will be higher than e-commerce. If your tCPA is unrealistically low, the algorithm will starve your campaigns trying to hit it.
Audience Layering
Export first-party lists from your CRM: current customers, open opportunities, churned accounts, and high-value prospects.
Start with observation mode to see how these segments perform. Then add bid adjustments based on what you learn.
Critical: exclude existing customers from acquisition campaigns. Otherwise, you'll pay to acquire people you already have.
Negative Keyword Discipline
B2B requires aggressive negative keywords from day one. Add negatives for: jobs, careers, free, DIY, student, intern, template, sample.
Review search terms weekly, not monthly. In long sales cycles, every wasted click compounds. A 10% reduction in irrelevant spend compounds into significant savings over a year.
Landing Page Reality
Long sales cycles need nurture paths, not just hard demo CTAs.
Not everyone is ready for a sales conversation. Offer a content download alongside the demo request. Capture at lower commitment, then nurture via email.
A "Download the Guide" conversion is still a conversion. It feeds the algorithm and gets the lead into your system. The demo can come later.
Proving Value to Finance and the Board
Your CFO doesn't care about CTR. CPL is interesting but incomplete. The only metrics that matter at the executive level: pipeline and revenue.
Build a reporting layer outside Google Ads.
Monthly reporting: Spend versus pipeline generated. Pull from your CRM, not Google. Show how much was spent and how much pipeline was created by leads from paid search.
Quarterly reporting: Spend versus closed revenue, with time lag explicitly noted. January's spend won't show up as closed revenue until Q3 or Q4. Present this as a cohort: "Q1 spend has generated $2.1M in pipeline, of which $340K has closed to date. Based on our average close rate, we project $1.2M in eventual revenue."
Annual reporting: Blended customer acquisition cost and payback period. How does paid search compare to other channels? What's the CAC relative to customer lifetime value?
Comparisons That Work
Executives understand relative performance. Show:
What Not to Present
The goal is building credibility over time. When you consistently show the connection between spend and revenue, budget conversations get easier.
When to Take Off Autopilot
Smart Bidding isn't set-and-forget for B2B. The algorithm optimizes for what you tell it to optimize for. If the inputs are wrong, the outputs will be too.
Signals That Automation Is Failing
Lead volume is up but SQL rate is dropping. The algorithm is finding easy conversions that don't turn into deals.
CPA looks good in-platform but pipeline quality is declining. You're hitting your targets with the wrong leads.
You're attracting the wrong companies. Too small, wrong industry, wrong titles. The algorithm doesn't know your ICP unless you tell it.
When to Intervene
New product or market launch. No historical data means the algorithm is guessing. Use manual CPC until you have 30+ conversions to learn from.
Major ICP change. If you're shifting upmarket or into a new vertical, past data will mislead the algorithm. Reset and rebuild.
Seasonal patterns the algorithm doesn't know about. Budget cycles, industry events, and fiscal year timing affect B2B buying. The algorithm learns slowly; you can anticipate.
CRM data contradicts in-platform data. If Google says performance is great but your pipeline says otherwise, trust your pipeline.
What Taking Off Autopilot Looks Like
Switch to manual CPC for new campaigns until you accumulate enough conversion data.
Use portfolio bid strategies to pool conversion data across multiple campaigns. This helps smaller campaigns benefit from larger ones' learning.
Add more conversion actions to give the algorithm more signal. If you're only tracking form fills, add thank-you page views, engaged sessions, or scroll depth as secondary signals.
Tighten targeting to trade volume for quality. A smaller, more qualified audience may convert better even with lower reach.
The Hybrid Approach
You don't have to choose all-automation or all-manual.
Let Smart Bidding run campaigns with proven performance and sufficient data. Keep manual control on experiments, new segments, and anything without enough conversion history.
Review weekly. Adjust monthly. Automation is a tool, not a strategy.