eCommerce · Hedges & Plants // UK, ongoing

From 5x to 6.5x ROAS.
And £60k more profit
in peak season.

How fixing a 90-day lookback window, building a profit model and deploying real-time spend scripts turned a good account into a great one.

Garden hedges and plants
Industry
eCommerce · Hedges & Plants
Channel
Google Ads only
Monthly spend
~£30,000
Status
Ongoing
// The situation

The account was performing.
The data was lying.

This UK hedges and plants brand was running a healthy Google Ads account with a solid ROAS. The brief was simple: scale it. But before touching budget, something didn't add up.

The purchase conversion tracking had a 90-day lookback window, meaning any customer who clicked an ad and purchased within 90 days was attributed to that click. For most eCommerce businesses that's standard. For a seasonal hedging brand where customers research, decide and buy within 10 to 12 days, it was massively inflating the reported ROAS.

"If your conversion window is 9 times longer than your actual buying cycle, your ROAS is a fiction. You're optimising towards a number that doesn't reflect what's really happening."

The business context
Seasonal business. Profit-first brief.
+
Hedges are a considered, seasonal purchase. Peak demand runs late January through early April. The client's brief wasn't just about ROAS. It was about maximising profit during those peak weeks. That required a different kind of model entirely.
The data problem
90-day window. 10-day buying cycle.
+
The mismatch meant the platform was attributing purchases to ad clicks that had no realistic influence on the decision. The fix was to stop relying on in-platform data and cross-reference everything against GA4, which showed the true picture: a 10 to 12 day purchase window was the real conversion pattern.
// The approach

Build the model.
Then automate it.

Once the data was clean, the work moved fast. The goal was to build a system that could respond to real-time profit signals, not just ROAS targets.

// 01
Data fix
Switched to GA4 as the source of truth
Stopped relying on in-platform conversion data. All optimisation decisions were cross-referenced against GA4 revenue data, which reflected the actual 10 to 12 day buying cycle and gave a true read on which campaigns were genuinely driving purchases.
// 02
Profit modelling
Built a seasonal profit model with real margins
Using the client's actual product margin data, we built a model that calculated the ROAS target needed at each point in the season to hit a specific profit outcome. Not a single ROAS target — a dynamic one that shifted based on where we were in the seasonal curve.
// 03
Custom labels
Categorised every product by profit tier
Products were segmented by profit margin using custom labels in Google Ads. High-margin products got more budget weighting. Lower-margin products were deprioritised. Budget allocation was driven by profit potential, not just search volume or historical ROAS.
// 04
Automation
Deployed real-time profit scripts in Google Ads
Built custom scripts that pulled GA4 revenue data hourly and calculated estimated profit in real time. On high-performing days the scripts automatically increased spend to capitalise on demand. On lower-performing days they pulled back. Profit maximisation running continuously, not checked weekly.
// 05
Learnings loop
Built the playbook for next season
Every insight from peak season — which product tiers outperformed, which days generated the highest profit signals, how the model behaved under different demand conditions — was documented into a seasonal playbook. The next campaign starts smarter than this one ended.
// The results

30% ROAS improvement.
£60k more profit in peak.

Against a £30k monthly budget, with Google as the only active channel, and a seasonal business model that gives you a narrow window to get it right.

6.5x
ROAS achieved
Up from 5x. A 30% improvement on a £30k monthly spend account — meaningful at that scale.
£60k
Additional profit in peak
Late January to early April. The model identified when to push and when to hold — and the scripts executed it in real time.
10–12
Day true buying window
Identified by cross-referencing GA4 against platform data. The insight that unlocked everything else.
Hourly
Profit signal frequency
GA4 revenue pulled hourly. Spend weighted up or down in real time based on profit performance, not gut feel.
1
Channel activated
Google Ads only. No additional channel spend required. The gain came from making existing budget work harder.
Year 2
Playbook ready
A full seasonal model and automation framework built for the next peak. Learnings compounding into the following year.

"The £60k profit uplift came from one realisation: the data we were optimising against wasn't real. Fix the data first. Everything else follows."

// The takeaway

Most Google Ads accounts
are optimising against the wrong number.

A mismatched conversion window doesn't just skew your reporting — it actively trains the platform to target the wrong people. The algorithm is optimising towards a signal that doesn't reflect real purchase behaviour. And the longer it runs, the harder it is to untangle.

Methodology applied
Data fix → Profit model → Custom labels → Automation
Built in-house. No off-the-shelf tools. The scripts, the model and the profit calculation framework were all custom-built for this account's specific seasonal pattern and margin structure.
// PROFIT_MODEL.log
$
// Work with us

Think your ROAS might be
built on wrong data?

A free discovery call is the right place to find out. No pitch deck. No pressure.

Book a Free Discovery Call Back to home
ThinkingMan Creative © 2026 ThinkingMan Creative · London, UK