Solved

Question about Behavioral Properties


Badge +1

Hi,

 

For the following filters:

 

Users who have not done filters will check in the last 30 days from the current date and time or from when the candidate was found in the graph?

For example, if the candidate is displayed in the graph on 1 March, these filters will check if the events exist before 1 March or before Today.

 

Thank you!

icon

Best answer by jonathan 27 May 2022, 18:35

View original

5 replies

Userlevel 5
Badge +3

Hey @vladiacob — good question.

  • Filters with relative date ranges are based on the current date and time. 
  • Queries with relative date ranges are based on the x-axis buckets. 

So in this case, you’re looking at users who completed a funnel each day, amongst users who have not done either of the specified events in the past month from today. 

There’s a more detailed explanation of how the query lookbacks work here. (Our docs don’t appear to explain the behavior of relative date ranges in filters but my answer is accurate.)

 

While it’s not exactly the same report, you may be able to use Path Comparisons in Funnels to answer a version of what seems to be an underlying question here, which is how many people completed this funnel without a touch from Product or Talent. That feature currently supports looking at an optional step between funnel steps i.e. Touch by Marketing > (Optional Touched by Product or Talent) > Candidate Created.

It doesn’t currently support optional steps before the funnel but we are working on enhancements to journey mapping so should be able to do this at some point. That said, if you have a common top of funnel event that everyone must do—say, visiting the home page—then you could build a funnel Visit Homepage > (Optional Touched by Product or Talent) > Touched  by Marketing > Candidate Created. That would tell you the number of users who converted via Marketing split by whether they had a Product or Talent touch before a Marketing touch

Hope that helps!

Badge +1

Hi @jonathan,

 

Thank you for your response.

I don’t see anywhere in the documentation explaining that loopback for filters is from the current date not from the x-axis buckets. I can find only the explanation for Query. 

 

What we want to calculate is how many users created an account from a marketing touch only (in the last 30 days). We want to exclude the product and talent touches from the last 30 days. 

Userlevel 5
Badge +3

@vladiacob Yes, I can’t find any mention in the docs of that fact either and I’ve mentioned it to our docs team. But trust me, that’s how it works. I edited my original answer so that’s clearer if you choose to mark it as the answer.

As long as the date range in your graph is the past 30 days, then your report does what probably you want it to do. Note that it will exclude anyone who had a product or talent touch in the last 30 days, period. That includes any time after completing the sequence as well as before. I’d suggest creating a funnel Touched by Marketing > Candidate Created > Touched by Product/Talent to verify how many users would be excluded by that filter.

Badge +1

Hi @jonathan,

 

Thank you again for your response!

I don’t think that funnel Touched by Marketing > Candidate Created > Touched by Product/Talent can help us because we want to understand how many candidates were touched only by marketing before the candidate is created.

I think we will need to create 2 funnels with optional steps in order to cover all the cases:

  1. Homepage → Optional Touched by Talent/Product → Touched by Marketing → Candidate Created
  2. Touched by Marked → Optional Touched by Talent/Product → Candidate Created

Am I right?

 

Thank you again.

Userlevel 5
Badge +3

@vladiacob Excluding Touched by Talent/Product in your original query already excludes anyone that had such a touch before the candidate was created. But it also excludes users who had such a touch after the candidate was created. So the maximum “error rate” in that original report is the number of users who did Touched by Marketing > Candidate Created > Touched by Talent/Product. 

So if that number if 100 and the total candidates created is 1000 then you can decide if it’s acceptable to exclude them. 

Those funnels are correct for looking at which touches are better at driving candidate creation but I don’t think you can safely sum the totals as there is overlap. There’s also still a couple small holes users could fall through (we are in beta right now on a much more complex journey builder which should allow us to close those up). My recommendation would be to find out how many users have a Talent/Product after candidate creation and consider that error bars on the conversion rate. My hunch is that it’s probably small enough not to skew the analysis. 

Reply