Från mänsklig byrå till AI: Hur Cogny halverade STTs kostnad per lead
Garden Services · Stockholm
The Challenge
Stockholm Trädgårdstjänst (STT) was paying a traditional human agency to manage their Google Ads account. Winter months meant burning budget on poorly converting campaigns — with a January CPA of 567 SEK that the agency had accepted as the seasonal norm.
The Solution
Cogny took over the Google Ads account in January and applied continuous AI-driven optimization — monitoring in real time, restructuring underperforming campaigns within weeks, and adapting bidding strategy as seasonal demand shifted through winter and into spring.
Från mänsklig byrå till AI: Hur Cogny halverade STTs kostnad per lead
Background
Stockholm Trädgårdstjänst (STT) is a Stockholm-based garden services company with strongly seasonal demand — spring and summer are peak, while winter months historically meant low intent, poor conversion, and budget that disappeared without results.
Before Cogny, the account was managed by a traditional human agency. Weekly check-ins, monthly reporting, and reactive adjustments were the norm. For a seasonal business where every winter week of poor performance is money that can't be recovered, this pace of optimization was costing STT real money.
Cogny took over in January 2025. What followed was the strongest three-month winter performance in the account's history.
Where Cogny Made the Biggest Difference: Jan–Mar
The most unambiguous improvements came in the winter and early spring months — historically the hardest period for garden services advertising.
January — The Turnaround Month
- CPA dropped from 567 SEK to 250 SEK (−56%)
- Lead volume increased 75% (12 → 21 leads)
- Achieved at lower total spend than the previous year
A human agency had accepted a 567 SEK CPA as normal for January. Cogny's AI identified the structural inefficiencies — wasted spend on broad keywords, poor bid adjustments for time of day and device, and campaign settings that hadn't been revisited since the summer — and restructured the campaigns within the first weeks.
The result: more leads, lower cost, less spend. Not a gradual drift, but a measurable step change.
February — Efficiency Under Budget Pressure
- CPA fell from 204 SEK to 140 SEK (−31%)
- Total spend was cut by 36% while maintaining lead flow
- The AI continued adapting as seasonal demand began shifting
February is typically a transitional month for garden services — interest starts picking up but conversion intent is still low. Cogny's continuous optimization kept adjusting the bid strategy in response to real-time signals, avoiding the over-spend that typically characterizes this period.
March — Peak Efficiency (Volume Constraint)
- CPA improved from 107 SEK to 77 SEK (−28%)
- March 2025 performance was so efficient that STT couldn't handle the incoming lead volume
- The bottleneck shifted from marketing to operations
This is the problem every business wants to have. Cogny had optimized the account to the point where the limiting factor was no longer budget or campaign performance — it was STT's capacity to service new customers. The marketing engine was outrunning operations.
Monthly Results at a Glance
| Month | Previous CPA | Cogny CPA | Change | Lead Volume |
|---|---|---|---|---|
| January | 567 SEK | 250 SEK | −56% | +75% (12 → 21 leads) |
| February | 204 SEK | 140 SEK | −31% | Spend −36% |
| March | 107 SEK | 77 SEK | −28% | At operational capacity |
The AI vs. Human Agency Difference
Traditional agencies review accounts weekly or monthly. Campaigns run on whatever settings were last touched — until the next scheduled review, the next client call, or the next crisis.
Cogny's AI monitors continuously. It identifies wasted spend in real time and acts on it. For a seasonal business like STT — where every winter week of poor performance is money lost that can't be recovered — this speed advantage compounds into significant savings.
What continuous optimization looks like in practice:
- Bid adjustments updated based on hourly and daily conversion patterns, not last month's data
- Keyword-level performance reviewed constantly, not at the next agency check-in
- Campaign structure evaluated against current intent signals, not last quarter's setup
- Budget allocation shifted dynamically as seasonal demand evolves
A human agency couldn't match this pace. Not because the people aren't skilled — but because the model of scheduled reviews and manual adjustments has a structural ceiling that AI doesn't have.
Key Takeaways
For seasonal businesses managing paid search:
- A high winter CPA isn't inevitable — it's often a sign that the account hasn't been optimized for the seasonal context
- Continuous optimization compounds: each week of better performance builds on the last
- When marketing efficiency outpaces operational capacity, that's a signal to invest in scaling both
- The goal isn't just lower CPA — it's more business at sustainable unit economics
What Happened Next
With the winter period validated, STT entered spring — historically their strongest season — with an account structure, bidding strategy, and optimization baseline that a human agency had never been able to build during the off-season. The low-cost, high-efficiency winter gave them a stronger foundation for spring scaling than any previous year.
The account's history didn't lie: Q1 2025 was the best winter performance STT had ever seen on paid search. And it happened because the AI didn't take breaks, didn't wait for monthly reviews, and didn't accept 567 SEK as a normal January CPA.
Transform Your Marketing Performance
Schedule a demo to discover how Cogny's AI can deliver measurable improvements for your team.
Schedule Demo