NovaCRM Reduced Cost per Demo by 52% with Offline Conversion Imports

Context

NovaCRM is a B2B SaaS company selling CRM software for mid-market businesses (50-500 employees). Their sales process involves a demo request → discovery call → proposal → close. Average deal size is €8,400/year.

They were running Google Search and LinkedIn Ads in-house, generating roughly 120 leads per month at €65 per lead. The problem: only 12% of those leads were sales-qualified. The sales team was burned out on unqualified demos, and marketing was frustrated that “the leads were fine, sales just doesn’t follow up.”

Funnel type: B2B demo-request lead generation Sales cycle: 45-90 days Target CPA for qualified demo: €150 (they were at €540)

Constraints

• Budget: €12,000/month across Google and LinkedIn
• Tracking: Standard conversion tracking only — Google Ads knew about form fills but nothing about lead quality or pipeline stage
• CRM: HubSpot with clean pipeline data — this was the untapped goldmine
• Sales team bandwidth: Could handle 25-30 demos per month, not 120 unqualified “demos”

Hypotheses & Experiments

Hypothesis 1: Optimizing for form fills attracts the wrong people

Google Ads was set to maximize conversions (form fills). The algorithm learned to find the cheapest form fillers, not the best prospects.

Experiment: We set up offline conversion imports from HubSpot. Every time a lead became a Sales Qualified Lead (SQL), that data was sent back to Google Ads. We then switched the bidding strategy to maximize conversions using the SQL event, not the form fill event.

Hypothesis 2: LinkedIn targeting is too broad

LinkedIn campaigns targeted “Marketing Managers” and “Sales Managers” at companies of 50+. This captured too many people at companies that were too small or in wrong industries.

Experiment: We narrowed targeting to specific job titles (VP of Sales, Head of CRM, Revenue Operations Manager), company sizes (100-500), and industries where NovaCRM had product-market fit (SaaS, professional services, financial services).

Hypothesis 3: Landing pages are generic

One landing page for all ad groups. Whether someone searched “CRM for mid-market” or “Salesforce alternative for growing teams,” they saw the same generic landing page.

Experiment: We built 4 dedicated landing pages: one for competitor comparison searches, one for category (“CRM for mid-market”), one for problem-aware (“sales pipeline management”), one for specific use cases (“CRM for professional services firms”).

Week-by-Week Progress

Weeks 1-3: Set up HubSpot → Google Ads offline conversion import. Mapped pipeline stages to conversion actions. Built new landing pages. Weeks 4-6: Launched new landing pages and restructured Google campaigns by intent tier. Initial data started flowing from offline conversions. LinkedIn audiences narrowed. Weeks 7-10: Google Ads started optimizing toward SQLs instead of form fills. Lead volume dropped from 120 to 65 per month — but SQL rate jumped from 12% to 28%. Weeks 11-16: Algorithm fully calibrated. Lead volume stabilized at 70/month. SQL rate reached 34%. Cost per SQL dropped from €540 to €260. Months 4-5: Refinement phase. Added LinkedIn retargeting via Meta (cheaper platform, same audience). Tested new CTA variants on landing pages. Pipeline influence from paid channels increased 180%.

Results

Metric	Before	After (Month 5)	Change
Monthly Leads	120	70	-42%
SQL Rate	12%	34%	+183%
Monthly SQLs	14.4	23.8	+65%
Cost per Lead	€65	€94	+45%
Cost per SQL	€540	€260	-52%
Pipeline Generated (monthly)	€86,400	€200,000+	+131%
Sales Team Satisfaction	Low	High	∞

What Did Not Work

• Google Display for prospecting: We tried running display campaigns to reach in-market B2B audiences. CPC was low but lead quality was terrible. We pivoted all Display budget to retargeting only.
• LinkedIn Lead Gen Forms: Pre-filled forms generated cheap leads but qualification rate was even worse than landing page leads. People submitted without reading. We switched to landing page traffic only.
• Short-form content offers on LinkedIn: “Top 10 CRM tips” generated downloads but almost zero intent to buy. We switched to comparison content and demo-focused CTAs.

Takeaways

1. Offline conversion imports are the most important thing you can do in B2B lead gen PPC. Stop optimizing for form fills. Start optimizing for qualified outcomes.
2. Fewer, better leads beats more, worse leads every time. The sales team went from drowning in bad leads to closing more deals with less effort.
3. LinkedIn works when targeting is precise. Broad targeting on LinkedIn is expensive waste. Narrow targeting with the right offer is worth the premium CPC.
4. Landing page specificity matters. Matching the landing page to the search intent is not a nice-to-have — it directly impacts both conversion rate and lead quality.