Measure a homepage promotion by separating the offer, the creative and its position, then checking that views and selections are collected consistently. For an Egyptian ecommerce team, this creates a useful basis for deciding which banner deserves further testing. A click count alone does not establish that the promotion generated additional profitable orders.
Start with the merchandising decision
Before adding tags, write down what the team wants to decide: retain the offer, change its message, move its placement or investigate the destination. These are different interventions. Changing all of them together makes the resulting comparison difficult to interpret.
Keep a small promotion register with the actual offer, approved dates, destination, products, language version and accountable owner. Record any eligibility conditions on the destination itself. Use the same commercial definition in Arabic and English; a translated headline should not silently change the discount or qualifying purchase.
Separate promotion identity from presentation
Google's ecommerce guide documents view_promotion for promotion impressions and select_promotion for selections. Its event reference describes promotion identity, creative and placement fields.
Use those fields deliberately. Keep a stable promotion identifier for the commercial offer, and document how creative revisions and placements are distinguished. Maintain a mapping between Arabic and English creative labels when both refer to the same offer. Do not make analysts reconstruct that relationship from an image filename.
Include the appropriate items data for promotion reporting. Google notes that omitting it from select_promotion excludes that event from some reports and metrics. Check the current reference rather than copying an old implementation unchanged.
Define what a recorded view means
Write an implementation rule for when the website sends an impression. Check the actual behavior of rotating banners, mobile layouts, page revisits and content loaded after navigation. A banner existing in page markup and a shopper having an opportunity to see it are different conditions.
Your rule should be reproducible. Record which condition the implementation detects and what it cannot establish. Do not describe an event count as verified human attention.
Assign one implementation owner so a theme component and a separate tag do not independently track the same action. Keep business identifiers free of customer contact details.
Test the journey before reading the chart
Use an acceptance sheet with expected behavior, observed behavior, evidence and an owner for each exception:
- Load each supported language and device layout.
- Observe the banner's impression behavior.
- Select it and confirm the correct destination and promotion identity.
- Navigate back and repeat the action to investigate duplicate sends.
- Check that disabled or replaced promotions do not retain stale labels.
Google's DebugView guidance explains how to inspect collected events with debug mode enabled. Respect the site's privacy and consent controls while testing; an absent signal requires investigation, not automatic removal of those controls.
Read performance at the right level
Google's Promotions report documentation defines item promotion click-through rate using users who selected a promotion divided by users who viewed it. Do not substitute raw event counts and present the resulting ratio as the same metric.
Compare like periods and record changes in placement, stock availability, pricing, language mix and the offer itself. Where order values are assessed in Egyptian pounds, keep the reporting currency and commercial revenue definition explicit. Do not mix analytics revenue with verified net receipts.
Use the ecommerce measurement guide for Egyptian teams to connect this work to purchase and refund reconciliation. A promotion-associated purchase remains an observation; establishing incremental impact requires an appropriate comparison design.
Give the readout an action
Close the review with a documented decision and its evidence: fix tracking, repair the destination, revise the creative or prepare a controlled test. Identify unresolved gaps before allocating more homepage space to the apparent winner.
Discuss a measurement review with Kitely. Bring the promotion register, current event definitions and unresolved discrepancies so the discussion can focus on a concrete implementation scope.
