Upkeystore

Campaign Architecture

Upkeystore ran multiple campaign types simultaneously: Performance Max (PMax) for both UK and USA markets, Shopping campaigns, and Search campaigns. The UK campaigns had been running longer and had accumulated conversion data, while USA campaigns were newer and operating without established brand recognition in that market.

The core question from the client was straightforward: should they continue spending in the USA, or would that budget be more effective concentrated in the UK? To answer this properly required more than surface-level metrics.

GA4 Insights

GA4 data revealed significant behavioural differences between UK and US visitors. UK users showed higher engagement rates, longer session durations, and significantly higher conversion rates. US traffic was clicking but not converting, suggesting either a mismatch between search intent and product offering, or a trust gap with an unfamiliar brand.

The product page vs catalogue comparison was revealing: users who landed on dedicated product pages (like /autodesk/autocad) spent 49 seconds on average with 388 events, while catalogue browsers spent only 8 seconds with 44 events. This confirmed that specific, intent-matched landing pages dramatically outperformed general browsing experiences.

Device Analysis

A striking disconnect emerged in device data: 65% of sessions came from mobile, but 43% of actual purchases happened on desktop. This pointed to significant mobile checkout friction. Users were discovering products on mobile but switching to desktop to complete purchases, or simply dropping off entirely. This insight fed directly into the CRO simulation scenarios.

Funnel Analysis

The biggest leak in the purchase funnel was at the Add to Basket stage, with only a 1.7% cart add rate. This meant the vast majority of product page visitors were not taking the first step toward purchase. For US visitors, this rate was even lower. The funnel data supported the argument that improving conversion rate (CRO) on the existing UK traffic would yield better returns than acquiring new traffic in the USA.

Simulation Methodology

Four scenarios were modelled using actual campaign performance data. Each scenario used the real CPC, conversion rates, and spend figures from the existing campaigns as inputs. The baseline (Scenario 1) reflected current performance. Scenario 2 modelled the impact of reallocating all USA spend to UK campaigns at their existing conversion rate. Scenario 3 added a conservative 20% CRO improvement to the reallocation. Scenario 4 explored what a new EU-focused brand might achieve, using three different conversion rate assumptions.

The modelling showed that pausing the US campaigns and reinvesting in UK with CRO improvements (Scenario 3) represented the strongest immediate path — projecting ~107 conversions at £19 CPA, a 34% improvement. For expansion, a new EU brand with a German-localised website (Scenario 4) offered the highest-probability growth path, supported by Germany's exceptional per-user engagement in the test data.