searchdiscovery / Apollo-Documentation-Connected-Vehicle

An Apollo generated Event Driven Data Layer
0 stars 0 forks source link

Page Load Started #2

Open msawlorSDI opened 1 year ago

msawlorSDI commented 1 year ago

Page Load Started

QA: https://github.com/searchdiscovery/Apollo-Documentation-Connected-Vehicle/issues/2#issuecomment-1894681281 QA Jan 26: https://github.com/searchdiscovery/Apollo-Documentation-Connected-Vehicle/issues/2#issuecomment-1912758719

jeffkush-sdi commented 11 months ago

PAGE LOAD STARTED STATUS: Event order/bracketing may need further discussion (#2). QA was only for datalayer structure/order @ this point in time, ie no values reviewed.

https://cwp.sit.siriusxmguardian.com/manage Page Load Started event should be fired as the first event on any page/virtual page load.

  1. We have noticed that the page load started event is not always/necessarily the first event fired on each page/virtual page load, but, page load completed may be the first event in some page load sequences.
  2. DEV DISCUSSION: In previous meetings there was discussion around 'when to fire a page load completed event', several thoughts were discussed and while firing page load completed as the first event (in some cases) on pages (essentially closing out the previous page hit) can be accommodated for within Adobe, it will overall, generate more hits/costs from Adobe in the long term @msawlorSDI
  3. RECOMMENDATION is to ensure page load complete always fires in the same page context as a page load started event. (ie a start and finish point)
  4. ex Page Context: [Page Load Started > ...bracketed event(s), if/where applicable... > Page Load Complete ]

QA:

  1. AFTER Authentication user is redirected > [https://cwp.sit.siriusxmguardian.com/manage]
  2. Image
  3. Page Load Start event does fire and some parameter are populated. (...review still in-progress, pending further page sequences)

(old) @ Team Initial quick spot test, noticed Page Load Started firing (x2). Should only fire (1x) Image

jeffkush-sdi commented 9 months ago

@Shocopop VIN_MASK parameter: We want to ensure that when providing a vin_mask parameter in our datalayer events that we are HASHING this value on the SXM side. In reviewing the above screenshot we see a value for vin (3F4...) but are not clear as to whether this is an actual 'real' VIN or a masked version of a 'real' VIN. Obviously using the same HASH is important as this this may become a 'unique identifier' in reporting.

jeffkush-sdi commented 9 months ago

@Shocopop

'Page Load Started' DataLayer QA: (updated results 1-17-24)

Use Case: https://cwp.sit.siriusxmguardian.com/manage; other use cases to follow, where applicable

Datalayer event schema looks correct and the event does appear to be bracketed with a closing Page Load Completed datalayer event. In this use case the expected events Product Listing Displayed + Checkout Step Encountered are bracketed properly.

Image

ISSUES: Below are some parameters that need discussion/fix. Specifically in this use case we are not seeing the 'vin_mask' parameter, whereas this has been seen @ other times in the dataLayer.

Image

ISSUE #2: Multiple datalayer pushes after login, but only sometimes(?). To Reproduce: simply login to the site after using email access code. landing page is 'manage'. I'm wondering if the difference is because in one UX I used the 'login' button in masthead > maybe this is causing an issue with multiple datalayer pushes...tbc (see screenie below). Normally I've been using this link to initiate a login...https://cwp.sit.siriusxmguardian.com/manage?u=64c17925-1ca4-4593-a2a5-261d6d975b08&v=bef0957f-673b-48bf-b1d6-53ae157ff65a&lang=en-US

Image

[FWW] To Review Server Calls/Rules... cc @msawlorSDI

(disregard below comments!) Image Image

Shocopop commented 9 months ago

@jeffkush-sdi this event is now populated with proper data

jeffkush-sdi commented 9 months ago

@Shocopop Parameters look good.

Image

jeffkush-sdi commented 8 months ago

@Shocopop Thank you. Looks good will review w/ @msawlorSDI on Monday.

QA .SIT

NO ISSUE (use-case below to be reviewed in later steps...)

Will verify context issue between pageName in some Checkout Step Encountered steps. We saw some issues regarding out of sync parameter values between these (2) events. UPDATE: verified that for .SIT user the pageName + checkoutStep were in-sync.

Image

jeffkush-sdi commented 8 months ago

@Shocopop @triegelman

Issues seen in PROD https://owner.siriusxmguardian.com/manage PROD QA: https://owner.siriusx mguardian.com/manage?u=2e8d5d4e-7136-4477-9fec-d275042af013&v=1-HOOZ-125 email: stl_prod_us11@yopmail.com

(.SIT) - could not log in recently https://cwp.sit.siriusxmguardian.com/manage?u=64c17925-1ca4-4593-a2a5-261d6d975b08&v=bef0957f-673b-48bf-b1d6-53ae157ff65a&lang=en-US Email: 3f4nkcac8jt630290@yopmail.com

  1. vin_mask: We are seeing an unmasked version of this value in reporting but unclear as to where this may be happening in the OTP flow appDataLayer event logic. (data from 2/22-2/23)
  2. vehicleId + user_Id: We are seeing a conflation (contextually interchanged) of these values in some UX; in some cases we are seeing vehicleId as the user_id (custKey) and visa versa...would apply to Page Load Started, User Detected...other(custKey) user related events?
  3. Page Load Started: customer_flow: missing value from applicable event. (see screenie for 'assumed/expected' value...not sure if there was a determination on naming convention?)

Image

cc @msawlorSDI Please review/add as needed for further clarification.