The dashboard looked fine at first glance. Google was generating 70+ leads per month. LinkedIn was adding another 25–30. Cost per lead was "within range."
But pipeline was flat for three consecutive quarters. And when I asked "Which of these leads actually became customers?" — silence.
Under the hood, the problems were compounding. Google campaigns were set to optimize toward four different conversion actions simultaneously — form fills, chatbot opens, page scrolls, and PDF downloads — all weighted equally. The algorithm had no idea what a "good" conversion looked like.
58% of total Google spend was going to broad match keywords. The actual search terms: "invoicing app free", "free invoice templates", "fintech internship." Opportunity rate on broad match: less than 1%.
One Performance Max campaign was generating the cheapest leads in the account at $14 CPL. In CRM: zero of those leads had progressed past initial contact.
LinkedIn was running cold demo ads to poorly targeted audiences. Job titles like "Financial Analyst" and "Accounting Assistant" with no seniority filters. Junior bookkeepers at 10-person firms were seeing the same ads as CFOs at scaling FinTech companies. No remarketing, no company size filters. Leads included interns, students, and people at companies well below ICP.
No offline conversions on either platform. The algorithms were completely blind to what happened after a form fill.
The founder's exact words: "I think Google is working, but honestly, I can't prove it."
Data infrastructure was broken at the foundation. GTM container had duplicate conversion tags inflating Google's reported conversions by ~25%. The "demo_request" event in GA4 was triggering on page load of the thank-you page, catching bot traffic and accidental navigations. Pipedrive deal stages existed but weren't mapped to any ad platform.
Budget was feeding the wrong campaigns. 58% of spend on broad match with near-zero pipeline. Performance Max looked like the winner on paper — cheapest CPL, zero opportunities. The best-performing ad in the LinkedIn account was getting less than 10% of the budget.
No awareness layer existed. Every dollar across both platforms was going to bottom-funnel conversion ads to cold audiences. No remarketing pool, no content promotion. Cold "Book a Demo" to strangers.
Based on these findings, I structured a 90-day roadmap in three tiers — prioritized by immediate revenue impact.
🔴 Critical fixes (Week 1–2): Fix duplicate GTM tags. Consolidate Google to one primary conversion action. Pause all broad match and Performance Max. Set up offline conversion imports from Pipedrive to both platforms. Immediate impact: $7K/mo in wasted spend recovered, clean signal for the algorithms.
🟡 Quick wins (Week 3–4): Build negative keyword lists seeded from 6 months of search term data and share across all campaigns. Tighten LinkedIn targeting — exclude irrelevant titles, add company size and seniority filters, restrict to target industries. Rewrite the top-performing LinkedIn ad (the one with the specific pain-point hook that was getting less than 10% of the budget) and shift allocation toward it.
🟢 Scale opportunities (Month 2–3): Restructure Google into intent-based campaign groups (branded / high-intent exact+phrase / competitor). Launch a LinkedIn awareness layer — customer proof content and product walkthroughs — to build a remarketing pool before running conversion ads. Once the pool hits 1K+, layer in demo CTAs to warm audiences. Test landing pages with outcome-specific headlines vs. the generic "Book a Demo" page.
The client executed the critical fixes and quick wins internally within 3 weeks, using the roadmap as a step-by-step guide. I stayed available for two months after delivery — clarifying specifics as questions came up and sharing references where needed. Implementation was fully on their side.
Metric
Delta
Timeframe
−28%
60 days
4% → 7% (+78%)
60 days
Opportunities
+58%
60 days
MQL volume
−13% (by design)
60 days
$24K → $17K (waste cut)
Immediate
A full audit of tracking, account structure, and messaging-to-funnel resonance, with a 90-day execution roadmap. If your dashboard says "on track" but your pipeline says otherwise
— start here: