Feature Idea: NRQL: who is dropping off in a funnel()

Hi All,

I have a funnel on a dashboard to monitor our registration flow, this is using a mix of screensViews and buttonClicks.

FROM MobileBreadcrumb SELECT funnel(sessionId,
where screenName = ‘app.verify number’ AS ‘Start Registration’,
where screenName = ‘app.create password’ AS ‘Create Password’,
where buttonName = ‘app.create password.create’ AS ‘Create Password Button’,
where screenName = ‘app.verify email email’ AS ‘Email Verify Page’,
where buttonName = ‘app.verify email.enter email’ AS ‘Send Email Button’,
where screenName = ‘app.sent email popup’ AS ‘email sent’)
SINCE 1 week ago

image.png1033×478 33.6 KB

This is showing quite a big drop off at a very unexpected point ( Just after where buttonName = ‘app.create password.create’ AS ‘Create Password Button’, ) in our flow which is a good thing to spot and one I want to dig into.

But I now need to get some information about who those people are?

Thanks

Kev.

New Relic Edit

We take feature ideas seriously and our product managers review every one when plotting their roadmaps. However, there is no guarantee this feature will be implemented. This post ensures the idea is put on the table and discussed though. So please vote and share your extra details with our team.

Hi @kthompson

Are you sending any identifiable information as custom attributes? by default there is not much that can be used to identify specific users, but if you are sending that as a custom attribute you could then query on something like:

SELECT count(sessionId) FROM MobileBreadcrumb WHERE buttonName = 'app.create password.create' AS 'Create Password Button' AND screenName != 'app.verify email email' AS 'Email Verify Page' FACET customUserId

Im not overly interested in identifying specific users per se, more a users journey through the app.

so I can get the sessionID and track that session so I can see what the user was doing around that point.

this is a great starting point. Thanks

Understood! I think that the query above would be helpful, alternatively rather than faceting and counting, just pulling back a list of Session IDs would be helpful.

I will note that the numbers returned from your funnel query will be different to any other queries you may run.

The funnel will return data where the session IDs have taken all of the actions you are filtering for.

The query I shared above factors in people who have taken a smaller subset of actions. So the funnel counts will likely always be smaller, since that is a more specific filtering mechanism.

Thanks, That did close to what I needed so I marked this as the solution.

Another quick question though, is there a way to enforce order in a funnel? I just noticed that insights says funnels do not enforce order.

No, not right now. Funnels will pull sessions that match all of the specified conditions, but not the order you list them in.

I can get that added as a feature idea for you though.

Hi Ryan, That would be a good feature I think, at the moment we’re trying to detect drop off points in our registration flow and without knowing the order its a bit fuzzy, its currently helpfull but full order would be more accurate.

Another one for funnel that maybe a nice feature ( if its not possible eslewhere and i’ve not seen it ) is to be able to split a journey.

For example. 1 of our pages has 3 main exit points. currently that requires 3 separate funnels to be made. If there was a way to split this into steps in a single funnel that would be amazing.

Absolutely, @kthompson

I can get both of these sent as Feature Ideas, for enforced order, and for split journey funnels.

Just as a side note - if you are using funnels more and want to get some additional data in your funnel visualisations, you could try the Customer Journey nerdpack.

I don’t believe it allows for enforced order, but you can see in the example screenshots that it can give you additional performance stats for all steps, and there are 3 funnels in one, so you can do separate journeys.

Excellent, I’ll take a look at that for sure

Thanks

No worries @kthompson - Hope it’s helpful