Conversion validation
Reviewed application and request-for-information events after the transition.
Higher education institution · Website relaunch analytics · 2026
Separating Relaunch EffectsFrom Channel Performance
I analyzed traffic and conversion behavior following a domain and website transition, separating changes across organic, paid, direct and secondary search channels.
01 · The challenge
A relaunch changed the website and the measurement context at the same time.
Following a domain and website transition, topline traffic could not explain whether important actions were improving or declining. The analysis needed to separate channels and conversion types to identify materially different outcomes.
Application submissions and request-for-information activity did not move in the same direction, making a single success metric misleading.
02 · Scope
Channel-level analysis of the actions that mattered
Channel comparison
Separated organic, paid, direct and secondary search performance.
Interpretation
Distinguished encouraging application growth from weaknesses elsewhere in the journey.
03 · Campaign process
Follow each conversion separately through the transition.
Confirm the measurement
Verify the relevant events and post-launch date range.
Segment by acquisition channel
Compare outcomes without allowing aggregate traffic to hide differences.
Contrast conversion types
Evaluate applications and information requests as distinct behaviors.
Surface follow-up questions
Identify which gains appeared durable and which declines required investigation.
04 · Results
The relaunch produced a mixed—but more understandable—performance picture.
+35%
Google organic application submissions
+39%
Direct application submissions
Percentages reflect the analyzed comparison period. Confirm client approval before publishing even anonymized performance figures.
05 · Reflection
A relaunch can improve one part of the journey while weakening another.
The analysis avoided treating total traffic or a single conversion as a verdict. Segmenting both channel and action exposed a more useful set of questions for the next round of optimization.