N
NexusFlow
B2B SaaS — Workflow Automation
Google Ads + Meta
Singapore
90 Days

They were spending $180K a year bidding against themselves.

Three agencies in 18 months. Each layered new campaigns on old. Nobody audited what was already there.

app.nexusflow.io
DashboardWorkflowsTeam
Overview
Automations
Integrations
Analytics
2.4K
Workflows
98.7%
Uptime
340
Teams
Workflow Executions
!
Acquisition brokenGreat product. Ads fighting each other.
What we walked into
0
Active Campaigns
0
Spend Wasted
0
Cost per Lead
$108K effective
$72K cannibalized
$0$180K annual spend
Diagnosis
Four findings. Two critical.
Critical

Self-cannibalization across 14 campaigns

Same search terms triggering competing ads in 7+ campaigns. The algorithm was bidding against itself.

Critical

Conversion tracking duplicated 3×

Three tags firing on the same event. Reported CPL was $42. Actual was closer to $65.

High

Zero negative keyword architecture

No cross-campaign negatives. Brand terms leaking into generic campaigns at 4× cost.

High

Bidding strategy mismatch

Maximize Clicks on campaigns that needed conversions. Optimizing for the wrong thing.

The client asked for
more budget.
We recommended
less campaigns.

Tear it down. Rebuild from scratch. Four campaigns by buyer intent. 90 days to prove the thesis.

Our Hypothesis
Three bets. Each one testable.
If we consolidate 14 → 4 campaigns
CPL drops below $25 in 8 weeks
Clean signal lets Smart Bidding find the cheapest conversions.
If we fix conversion tracking
Quality improves, reported CPL rises initially
Algorithm optimizes on truth instead of noise.
If we add negative keyword architecture
Brand waste drops 80%+
Traffic routes to the right campaign at the right bid.
Known Risk — Flagged Upfront
Lead volume will drop 20–35% in weeks 1–2 as the algorithm recalibrates. This is expected and necessary.
Pilot
14 campaigns became 4.
Before
A B C D
14
campaigns · 47 ad groups
After
HUB BRAND 3 groups HIGH 4 groups MID 3 groups PROSP 2 groups
4
campaigns · 12 ad groups
Week 1–2
Consolidation
Recalibrating
Week 3–4
Tracking rebuild
Clean data
Week 5–8
Negative keywords
-28% CPL
Week 9–12
Scale proven
+220% vol
Outcome
The hypothesis held. The numbers moved.
$42
$0
Cost per Lead
↓ 57%
1.0×
0
Lead Volume
↑ 220%
14
0
Campaigns
↓ 71%
1.8×
0
ROAS
↑ 128%
Cost per Lead — 12 Week Transformation
Baseline
After rebuild
$10$20$30$40$50 Wk 1Wk 3Wk 5Wk 7Wk 9Wk 12 ConsolidateRebuildScale $42 Expected dip $18
Hypothesis vs Reality
What played out. What surprised us.

CPL dropped below $25 within 8 weeks

Hit $22 by week 6 — two weeks ahead. Reached $18 by week 12.

Confirmed

Lead quality improved beyond expectations

40% higher SQL rates. Sales noticed before we reported it.

Exceeded

Week 1–2 dip was 30%, recovered by week 4

Flagging upfront preserved client confidence. Recovery faster than projected.

Confirmed
~

Brand waste reduced 65%, not 80%

Some brand-adjacent terms kept by design — CPL acceptable, drove incremental volume.

Partial
Key Learnings
Consolidation beats optimization

The rebuild outperformed 18 months of incremental tweaks in 6 weeks.

Flag the dip before it happens

Without the warning, the client would have paused the experiment.

Clean tracking > more tracking

Smart Bidding got smarter when we gave it truth instead of noise.

Perfect is the enemy of done

65% waste reduction was enough to prove the model and reinvest.

"
We thought we needed a bigger budget. Turns out we needed someone willing to tell us the previous agencies had built on a broken foundation.
VP Marketing — NexusFlow, Singapore
What happened next

Scope expanded

Google Ads grew to include Meta lead gen and LinkedIn ABM. Retainer doubled in 6 months.

Capability transferred

Client's team runs day-to-day on the framework we built. We focus on strategy and testing.

Results sustained

12 months on, CPL holds at $16–20. Sales reports 40% higher SQL rate.

Full synopsis

Diagnosis

NexusFlow cycled through three agencies in 18 months. Each inherited the previous account structure and layered new campaigns on top. Nobody questioned whether the foundation was sound.

Kaliber's audit uncovered the root issue: 14 campaigns organized by product line were cannibalizing each other in every auction. Smart Bidding had no clean signal — it was optimizing chaos.

Conversion tracking was duplicated across three tags, inflating reported conversions by ~35%. The client believed $42 CPL. The real number was closer to $65.

Pilot

Kaliber proposed a complete rebuild — consolidating 14 campaigns into 4 by buyer intent (Brand, High Intent, Mid Funnel, Prospecting). All duplicate conversion tags stripped to a single GA4 source. 400+ negative keywords deployed in a cross-campaign matrix.

Week 1–2 saw a predicted 30% dip as the algorithm recalibrated. By week 4, volume recovered. By week 8, every prior benchmark was exceeded.

Outcome

No additional budget. The same $15K/month produced 3.2× lead volume at 57% lower CPL. Waste was redirected into audiences already converting.

12 months later, CPL holds at $16–20. Sales team reports 40% higher SQL rates. The engagement expanded to Meta and LinkedIn ABM, with the retainer doubling in six months.

Sound familiar?
Most performance problems aren't about budget.
They're about architecture.
Book a diagnostic →