Helpland Google Ads Case Study: 99% ICP Conversions | Jam 7

Helpland’s Google Ads account had a strong foundation, but performance was being dragged down by a persistent problem: tenant enquiries.

Helpland is a landlord-first business, yet tenant-intent searches were slipping through and generating phone calls that were a poor fit. Those calls weren’t just a waste of spend. They diluted conversion learning, reduced call quality, and created operational noise.

Jam 7’s goal was simple: protect call quality without reducing legitimate landlord demand. The work focused on clarifying landlord-only positioning at the moment of the click or call, and tightening targeting hygiene so the account learned from the right audience.

The TL;DR

• Who: Helpland, landlord-focused eviction support.

• What: Reduced cost per conversion and improved call quality by filtering tenant-intent traffic.

• Why: Tenant calls were undermining efficiency and polluting conversion learning.

• When: Within the first month.

• How: Search-term forensics + intent-cluster negative architecture + match-type tightening + RSA and asset alignment to landlord-only positioning + call-quality measurement feedback loop.

Situation and stakes

Helpland serves landlords and letting agents who need specialist support fast. In search, that means high-urgency buyers with low patience and little room for ambiguity.

But despite a solid starting point, Helpland’s Google Ads performance was being undermined by tenant leakage.

The stakes were commercial and operational:

• Wasted spend on the wrong audience
• Lower call quality, with more time spent handling irrelevant calls
• Distorted Smart Bidding learning (bad calls are bad learning)

The strategic bet: make “landlord-only” unavoidable (and speed up learning)

We focused on reducing tenant leakage at two points:

1. Pre-impression targeting hygiene: systematically block tenant-intent themes with negatives and a tighter match-type posture.
2. Pre-click clarity: make landlord-only positioning obvious in RSAs and assets so tenants self-filter.

This preserved legitimate landlord demand while reducing irrelevant calls and accelerating learning.

Implementation timeline (what happened when)

Phase 1: Diagnosis (days 1–3)

• Search term review to identify tenant-intent patterns slipping through
• Review of existing negatives (400+) to spot gaps
• Campaign and match-type audit to locate where looser matching pulled in the wrong audience

Phase 2: Restructure and hygiene (week 1)

• Built an intent-cluster negative framework (blocking themes, not chasing one-off terms)
• Tightened match types and rebuilt keyword posture where leakage was occurring
• Refocused targeting around landlord legal intent themes (e.g., Section 21, Section 8, possession and eviction)

Phase 3: Message alignment and measurement (week 2–4)

• Reworked ad copy and assets to make landlord-only intent explicit
• Introduced call-quality as a first-class KPI
• Fed learnings back into exclusions and messaging for rapid iteration

The operating model Jam 7 deployed (human-led, AI-powered)

Ownership & cadence

Jam 7 owned diagnosis, architecture, copy alignment, and iteration cycles. Helpland provided domain context and feedback on call quality.

“Marketing brain” approach to paid search

Instead of treating optimisation as isolated tweaks, we treated it as a system:

• Memory: an exclusion structure that persists and scales (intent clusters)
• Consistency: copy and assets that repeat the same qualifier (“landlord-only”) everywhere
• Learning: faster feedback loops so the account improves week by week

Human-in-the-loop (HITL) governance

This is the core principle behind our Agentic Marketing Platform®: automation is only as good as the signals you give it.

We used a human-led review loop to ensure:

• exclusions were blocking the right themes
• RSAs signalled the right buyer and self-filtered the wrong one
• the account learned from qualified calls, not noise

Results

Feb 2026 delivered material improvements vs baseline:

• Cost per conversion: 20% reduction
• 60s+ calls: 17% increase
• Client feedback: 99% high-quality calls (from a very low previous benchmark)

The qualitative result mattered most:

“We hardly get any tenants now. When an enquiry comes in, 99 times out of 100, it’s a landlord. That’s everything we wanted.”

— Lee Daniels, Helpland

Why it worked (the non-obvious bit)

The win wasn’t “more budget” or “more keywords”. It was conversion hygiene.

Tenant calls are uniquely damaging in a call-led account because they:

• waste spend and team time
• distort conversion learning
• make automation optimise towards the wrong people

By systematically reducing tenant leakage and making landlord-only intent explicit in ads and assets, Helpland improved the quality of what the algorithm learned from and efficiency followed.

Transferability

This approach is a strong fit for any service business where:

• there is audience confusion risk (consumer vs professional, free advice seekers vs paid service buyers)
• lead quality matters as much as CPA

In those contexts, speed comes from structure: a clear intent taxonomy, fast iteration, and consistent messaging that helps the wrong audience self-select out.

Replicable playbook

1. Define what “qualified” means (call quality rules).
2. Make the qualifier unavoidable in ad copy and assets.
3. Build negatives as intent clusters (themes, not one-offs).
4. Tighten match types until quality stabilises.
5. Review search terms and ship exclusions immediately.

Conclusion

Helpland’s performance improvement came from focusing on the right outcome: qualified landlord calls.

By tightening audience fit and making landlord-only messaging unavoidable, Jam 7 reduced cost per conversion and improved call quality — while setting the account up for the next performance phase, once prioritised.