Conversion rate optimisation is the practice of increasing the share of visitors who complete a valuable action, using research, hypotheses and controlled experiments. Marketing analytics is the measurement layer underneath it: tracking, attribution and reporting. Avar Systems builds both — GA4 and tag management implementations you can audit, and experiments run against revenue metrics, not surface engagement.
This is the data layer itself — the tracking, the reconciliation, and the variance between what the platform claims and what your CRM recorded.
Figures describe a mid-sized engagement — a single commercial funnel with enough traffic to test on. Illustrative of an engagement at this size, not a named client’s results. These describe the volume of work and how it is governed — not an outcome we are claiming on someone else’s behalf.
The 5 this one is judged on
8 workstreams. Each one is a thing that gets done, not a thing that gets discussed.
GA4 and Google Tag Manager built to a documented measurement plan, not accumulated tag by tag.
An agreed definition of what counts as a conversion, applied consistently across every channel.
Closing the gap between what the ad platforms claim and what your CRM recorded.
Analytics, session recordings and form analysis to locate where intent is being lost.
Hypotheses, test design, and results read honestly — including the tests that found nothing.
The highest-leverage surfaces, because they sit directly under paid spend.
One view across ad spend, site behaviour and CRM outcomes, in Looker Studio, Power BI, Metabase or a custom app.
Tracking QA on a schedule, so measurement does not quietly rot after launch.
5 stages. The first two happen before any spend is committed.
Work happens in accounts and properties you own. Each system is reached through its own MCP server, authenticated with a token scoped to an account you own — so the data is read where it already lives rather than copied over to us.
Analytics & tags
GA4 Google Tag ManagerOther platforms
Google Optimize-class testing tools PostgreSQL / SupabaseCRM & lifecycle
HubSpot / Zoho / SalesforceWarehouse & BI
Looker Studio Power BI MetabaseNamed as the tools we work in, not as partnerships or certifications. Accounts and properties stay in your ownership throughout.
Agreed before work starts, so a later claim of improvement is checkable against a baseline rather than asserted.
Because every optimisation decision after it depends on the measurement being right. Scaling spend on a broken conversion definition multiplies the error rather than the result, so tracking comes first in every engagement.
Some variance is expected — attribution windows and de-duplication differ by design. A large or unexplained gap is a tracking defect, and finding which of the two is wrong is part of the audit.
Enough for a result to mean something. Below that threshold the useful work is qualitative research and fixing obvious friction, and we will say so rather than run tests that cannot reach significance.
You do. Analytics live in your own properties and dashboards are built in your accounts — consistent with how Avar handles data residency on engineering work.
Book a free discovery call, or ask our AI assistant a question in the corner.
Or email us directly at avaris@avarsystems.com
You're all set
Thanks for reaching out.