Measurement and optimisation
Tracking that survives an audit
Operloom specifies, implements and validates tracking across the website, campaigns and CRM, then builds the reporting layer the definitions sit behind.
01Business problems
Reporting people do not trust does not get used
The usual cause is not the tool. It is that no definition was agreed, so every dashboard answers a slightly different question.
- Marketing and finance count a conversion differently
- Attribution changes when nobody changed anything
- Consent state ignored by half the tags
- Events named inconsistently across the site
- No record of what was implemented or why
02Measurement planning
Start from the decision, not the tool
The plan records which decisions the measurement has to support, and only then the events, parameters and dimensions needed to support them.
- 01Decisions and the questions behind them
- 02Metric definitions, written down and agreed
- 03Event and parameter naming convention
- 04Consent requirements per category
- 05Data retention and access policy
03Implementation scope
Specification, build, test
Implementation follows the specification, and nothing is signed off until it has been verified against real traffic.
- Tag manager container and governance
- Analytics configuration and data streams
- Ecommerce or lead conversion tracking
- Server-side collection where it is warranted
- CRM and offline conversion connection
- Quality assurance against the specification
04Consent
Measurement that respects the choice
Consent is a build requirement, not a banner. Tags fire according to the recorded state, and the state is auditable.
- 01Consent categories mapped to tags
- 02Default-deny before a choice is made
- 03Consent state stored and evidenced
- 04Behaviour verified for each category
- 05Documented approach for a data request
05Reporting
One set of definitions behind every view
Dashboards reference the same documented definitions, so a number can always be traced back to how it was calculated.
- Executive, channel and operational views
- Attribution model chosen and explained
- Data quality monitoring and alerts
- A written definition per metric
- Change log when a definition moves
Common questions
Question: Which attribution model should we use?
The one whose assumptions your team can explain and act on. The model matters far less than everyone agreeing what a conversion is and where the data comes from.
Question: Can tracking work properly with consent enforced?
Yes, within limits that should be stated openly. Consent-respecting measurement gives a smaller but honest dataset, and modelling gaps is a decision to make deliberately rather than by default.
Related capability
Where this connects
Next step
Build a systemyour team can actually run
Start with the process as it works today, the platforms already in play and the outcome the organisation needs. We will tell you what we would change and in what order.