Conversions Matter / Anonymised measurement case study
Analytics was under-reporting online sales by approximately 95% compared with banked revenue, raising a fundamental question: what was the analytics actually measuring?
01 / The challenge
The client compared online sales reported in analytics with banked revenue. Analytics-reported sales were approximately 95% lower, making analytics an incomplete picture of commercial performance.
Before using those numbers to judge online performance, attribution or experiments, the business needed to understand the coverage behind them.
What the figure means: ~95% describes the under-reporting of sales in analytics relative to banked revenue in this engagement. It is not a revenue uplift, a consent rate, or a claim that 95% of sales were recovered.
02 / The finding
I investigated opt-in rates alongside browser and extension blocking, using a combination of Measurement Protocol and server data to assess what the browser-based analytics could observe. The gap reflected incomplete measurement coverage, rather than missing sales.
That distinction changed the interpretation. A missing analytics record did not, by itself, mean a sale had not happened. Equally, the revenue comparison did not provide an online attribution record for every sale.
The useful question became which decisions each source could support, given what it covered.
03 / How I checked
I reviewed the consent opt-in rate to establish how much activity the consented analytics setup could cover. I then examined browser and extension blocking, which could further reduce the activity recorded.
I used Measurement Protocol alongside server data to compare those coverage signals with the activity visible in analytics. Looking across these sources helped me distinguish measurement loss from commercial performance, rather than attributing the full gap to consent alone.
The approximately 95% figure is the overall gap between analytics-reported sales and banked revenue; it is not a consent refusal rate or an estimate of consent’s contribution on its own.
04 / What changed
The client incorporated offline sales data into its measurement approach, bringing a broader commercial view alongside online analytics. This changed which source the team used to understand total sales and which measures it used to assess the observable online journey.
This did not make every offline sale attributable to an online interaction. It provided additional business context alongside the online analytics.
The client refined the KPIs used for online attribution and experimentation in light of the measurement coverage.
The distinction between overall sales performance and what could be observed online became central to interpreting those KPIs.
05 / The decision
The client moved from treating analytics as the full picture of sales to using it within a broader measurement approach. Offline sales data provided commercial context, while attribution and experiment KPIs were refined around the coverage each source could support.
This case illustrates why coverage matters before an analytics number becomes a business conclusion. More data is useful when its scope is understood; it does not automatically establish attribution or prove an experimental effect.
Context and limitations: This case study is anonymised to protect client confidentiality. The discrepancy is approximate. No quantified improvement in revenue, attribution accuracy or experiment performance is claimed.
