I recently learned about the T_DEFAULTS setting "General Public Zip Question ID" that allows Analytics to pull in the Survey Question from "0" General Public orders as a zip code for overall Analytics evaluation - and this has been awesome for us.
We have a new quandary that I'm not sure the answer.
In a somewhat archaic way, we have an event where we intend to collect first name, last name, email, and zip code on paper for guests who wait in line for an Adult Night. Then, when they make their Walk-Up purchase, the cashier will write down the Order Number on the piece of paper (this is all about keeping the line moving on this busy night, rather than real-time data entry - I know, this is its own topic :). Later, we intend to have staff look up the order and attach it to a purchaser knowing that most will be new constituent records that we will create.
So, with that set-up, my question is - at the time of purchase, the Cashier will have input a Zip Code into the survey field. When we later go back and add a constituent with a known Zip Code in the Zip Code field...what happens in Analytics? There will be two zip codes (that hopefully match) From an Analytics standpoint, which Zip Code is the recorded Zip Code? Will it pull both and skew data? Does anyone know the logic behind dueling Zip Codes?
Hopefully this makes sense. And, hopefully a Tessitura Staff member may be read this, too!
Thanks in advance,
-Mark
Hi Mark,
In this scenario, the Tessitura Analytics Primary Address information will reflect the newly associated constituent rather than the information from the survey response. However, please keep this in mind from the 15.1.6 / 15.0.13 release notes:
The issue there was, when the Gen Pub Zip Question is enabled, orders for known constituents, who answered that post code survey question were being being updated to look like gen pub orders in Analytics.
Best,Chris
Chris Wallingford Product Owner Tessitura Network office: +1 888.643.5778 x553 chris.wallingford@tessituranetwork.com
Thank you! Good to hear. This was our hopeful answer. Appreciate the quick reply,
I have a related question. Our marketing team is looking into Zip Code data. For most of our sales pre-Covid, we had walkups, and we used survey responses to track zip codes, with no Constituent selected.
We also had online sales, where guests manually entered their zip codes (or staff did that on the back end).
Does Analytics treat them the same? If I pull a pivot together with attended count by survey response, will it be the same as post code?When I pull both, it looks like they line up closely, but there are differences. I'm not sure how they relate.
It's a one-directional relationship. So, for the question you have configured as the 'General Public Zip Question ID' in T_DEFAULTS, those answers are combined with the postal codes from actual constituent records' primary addresses into the Primary Address Postcode. In other words: if specific real constituent A has has a primary address, Primary Address Postcode comes from that constituent's address; if an order belonging to Gen Pub has a postcode question survey response, their answer is in that field.
If you filtered down to only constituent A, you'd still see their Primary Address Postcode populated from their primary address. If they never answered the survey, you won't see a postcode in the Survey Answer (If you did ask them the survey then the Survey Answer would have that. And they may or may not be the same, depending on what was entered)
If your goal is to look at people like A and B at the same time and analyze where they're from, I'd recommend using the Primary Address Postcode and leaving the Survey Answer out of it. In fact, in many cases I'd almost recommend not setting the post code question as one of the 10 survey question/answers in Analytics. If you want to analyze those postcodes as postcodes, setting the applicable question as General Public Zip Question ID will allow you do that by looking at the Primary Address Postcode.
The main case for using both and having the survey answer for this question in there is if you sometimes get (purposeful, intentionally) different answers from non-Gen Pub constituents and wanted to compare the two - or only consider the answers rather than real addresses. Or if you have addresses as not required, so some real constituents have no primary postal code, but do have a survey answer.
I ask because we will have to compare data from years with mostly survey responses for zip codes, and years with mostly postal codes for zip codes. (How) can we do that, cleanly?
Theoretically, "more survey responses for zip codes" should mean "more General Public orders" and thus should not make one approach more "right" than the other. Remember the survey answer is only funneled into the postal code in Analytics for General Public orders. The purpose of that defaults setting is to create a method of supplanting that information about people for whom there is nowhere else to store it (because the orders are all lumped on the General Public account). I'd argue that even asking the survey question of real constituents for whom you can and have collected actual address data is redundant and introduces ambiguity where there needn't have been any. So I'd generally recommend using the Postal Code as the starting point.
If the scenario is there were a switch in business practice from real constituents have addresses and General Public has surveys into nobody needs an address but everyone gets a survey that muddies it a bit. However if these are mostly web transactions, I'd think you'd likely have a billing address. Because of that I'd likely still lean toward using postal code. You could also potentially do some sort of bucketing/formula where you take the postal code, but if that's "(none)" display the survey answer instead. Those are really the people you'd need to account for. People with one postal code on their primary address but potentially other postal codes in other orders' surveys aren't really different than people with additional non-primary addresses. You're picking one, either way.
As I understand it- the Survey Response is brought in as Postal Code for General Public orders. If there was no constituent, Postal Code is the same as Survey Response.
If there are both connected to an Order, both will be presented.
Are those statements correct?