Trying to get a segment created seeing how an event and evar are together

[edbudev]

Hi, I have a success event and with that event, I have an eVar that sets, now the eVar also sets in other places but my intended goal is to check ALL of the times that success event fired and when that success event fired, check to make sure that the eVar successfully was present as well. Just want to make sure the success event and the evar match 1 to 1 with no data loss. Appreciate it

[Jennifer_Dungan · Accepted solution]

If this is sort of an investigation.. you can create multiple segments to compare side by side...

So if I understand, you have:

• an eVar which is used in multiple places (I am going to assume it's also a VISIT level expiry)
• an event which is fired on a subset of places, also in correlation with the eVar (or so you hope)

Now, when you are setting the event, I assume you are also setting the eVar (as in you are not relying on the VISIT attribution to set the value)

So to see the event in isolation, you can create a segment like:

HIT

    eventX exists

Now if you need to see explicitly where the eVar is set, and not just by VISIT level attribution, you can create a segment like:

HIT

    eVarY exists

    AND

    eVarY Instances exists

^ The instances of the eVar is a specific metric that is set only when a value is specifically passed to an eVar, but not when the value is carried forward by attribution.

Then, the last part, you can either stack those two segments on top of one another, or create another segment for:

HIT

    eventX exists

    AND

    eVarY exists

    AND

    eVarY Instances exists

Then you can compare where you potentially have mis-matched tracking

[RobertBlakeley]

We here use this approach all the time for ongoing QA. You can create rules to test the dependencies of different variables, like the one above where eventX should always be set at the same time as evarY. Or you can get more complex, e.g.  evarY is set with a value of C when (clickA  AND clickB) are triggered and if true, set eventX.

If such expectations are consistent across all or a part of your site, you can create a kind of segment library that you can use for QA every time you launch or update something. Or you can create a QA monitoring workspace and use alerts to signal when some dependency is not happening as expected. (Or you could buy Observepoint which is a rules based QA tool).

[edbudev]

Thank you for this information as well! We used to use observepoint but we no longer use it.

[Amruthesh_AG]

@edbudev 

observepoint tool is a paid version.

[edbudev]

Thank you so much for this! 

[Jennifer_Dungan]

You're welcome.. Good luck with your investigation!