Skip to content

On - AP/AR Data Collection - 2025-02-26

Metadata

  • Date: 2025-02-26
  • Company: On
  • External Participants: Rodrigo Cabrera (Treasury)
  • Palm Participants: Emma Sjöström, Gurjit Pannu
  • Type: Customer Call
  • Domain Areas: Cash Forecasting
  • Recording: https://tldv.io/app/meetings/67bf13ed2adeaf0013cefe72/

Summary

Context

Operational call to collect AP/AR invoice data from On's systems (Dynamics, Looker, BigQuery) for Palm's forecasting model development. This is a technical data collection session, not a discovery call.

Key Discussion Points

  • Data sources: On uses Dynamics as ERP, with data replicated to BigQuery and visualized in Looker (BI tool)
  • Dynamics limitations: Must export entity-by-entity (no group-level export); historical invoice data includes deleted/duplicated records
  • Target entities: AG and UK selected as good mix of activity for initial modeling
  • Data fields needed: Vendor name, amount, due date, processing date, close date - everything except contact info
  • AR data cleaner than AP: AR data is more accurate; AP has some data quality issues (e.g., due dates showing 1900)
  • Intercompany included: Vendor data includes intercompany transactions
  • Date range: Targeting 1 year of data (6 months minimum if too large)

Data Insights

  • Close date = 1900 means invoice not yet closed/paid
  • Salaries not in invoice data - would need separate data source (phase 2)
  • Data available back to 2020-2021

Action Items

  • [ ] Rodrigo to pull AP data export for AG and UK entities (1 year if possible)
  • [ ] Rodrigo to pull AR data export (cleaner/ready reports available)
  • [ ] Include field documentation with exports explaining data interpretation
  • [ ] Palm team to start analysis next week

Notable Quotes

"AP data vendor data is a little bit work in progress because it's not really a hundred percent accurate. If we compare it to Dynamics, especially in the due date, close date - there's some weird stuff going on." - Rodrigo Cabrera

Full Transcript

00:00 Gurjit Pannu: Good good, that's fine. Emma and I were both just complaining actually right before you joined, so it's good. That one of us is like we're talking about our lack of sleeve last night, for different reasons. But it really thinks Thanks again, man, for setting up the time as we chatted last week, right? What we're trying to do or actually all Emma can explain, we're trying to do, but in this session we're just kind of trying to tap into some data sources to help us kind of build out some forecasting for you, but I think Emma's best.

00:31 Gurjit Pannu: If you drive this one that you have the better context, you share,

00:35 Emma Sjöström: Sure.

01:54 Rodrigo Cabrera: And no, not a super on board and aligned. Yeah. happy to provide any any any data that you guys might need in that, of course will be helpful for For what we're trying to achieve. A.

02:13 Emma Sjöström: Super.

02:14 Gurjit Pannu: Here. Yeah. Yeah.

03:09 Rodrigo Cabrera: H. Yeah, of course, of course.

03:13 Gurjit Pannu: Cool. Emma, do you. Okay, I mean, I think Do you have an idea on Where do you want to start? Is it better for Rodrigo to share or maybe Rodrigo? If you know where some of the data is? So I'm not familiar with with dynamics. I've never been in it to be able to say Hey you need to go here for the data.

03:29 Gurjit Pannu: But how do you guys want to tackle this?

03:31 Emma Sjöström: I'm not that familiar either, to be honest.

03:34 Gurjit Pannu: Yeah.

03:48 Rodrigo Cabrera: Yeah, no, I think most if not all of the Important data to put it some way. It's that it's in dynamics. It's also in Bigquery and Hexin looker at the end of the day, looker. It's our BI tool. So we try to keep most of the data there. a So I think we can I can I can share my screen and we can go over some of Some reports in all of that.

04:23 Rodrigo Cabrera: Just to give you an idea of what will be possible. in in Dynamics, I'm not super familiar either with this, a module of A Like I can't payable bend or something invoices and all of that. But there's a screen, I don't know which I was before, but you're not, there's a screen where you can, when where we have all open vendor invoices, right? There's another screen of all historical opening voices.

04:57 Rodrigo Cabrera: But in that one, I, I actually discussed this with the AP team when we first started working on this request that, that could be a good thing to share, but it's not extremely accurate because if If invoice was deleted or it was duplicated for some reason or, or whatever it will still be there.

05:23 Rodrigo Cabrera: So you it's not really super accurate. Another downside of dynamics, if I wanted to share the data is that I need to go entity by entity. I cannot share the data for like like for the group. A. So I think if we look into looker that will be the same data that we have in the in the hex like in Bigquery, I think that that will be a The best.

05:56 Rodrigo Cabrera: Way to picture everything that we have available.

06:00 Emma Sjöström: Okay. because,

06:02 Rodrigo Cabrera: In.

08:36 Gurjit Pannu: Yeah.

08:54 Emma Sjöström: Yeah.

08:58 Rodrigo Cabrera: request.

09:45 Emma Sjöström: All. but ideally, for the same time period that we've got, you know, transactions for Cash flows. Some backwards.

10:05 Rodrigo Cabrera: I think we,

10:49 Emma Sjöström: that's,

10:49 Gurjit Pannu: I'm that should give us everything we need, right? Like maybe, like I don't know, I don't know how much data is how long it takes to pull like in one go but like I don't know. Two, three weeks worth of. We pick a couple of accounts maybe because if you do every account that might be too much but Your call but I would imagine that Rodrigo all these fields here.

11:16 Gurjit Pannu: We can technically select all of them if you wanted to, right?

11:19 Rodrigo Cabrera: Yeah. Yeah.

11:21 Gurjit Pannu: The. Yeah, so we have actual data all the way back to what 2020? like,

11:40 Rodrigo Cabrera: 220, I think, or 2021

11:44 Emma Sjöström: Do we have all that data in home?

11:47 Gurjit Pannu: Yeah.

11:55 Rodrigo Cabrera: Maybe it's at the big one, but maybe on UK that it a decent size. We have a few accounts, A

12:03 Gurjit Pannu: Yeah.

12:04 Rodrigo Cabrera: and we can, we can put it together a report based on on this we filter The based on the, on the due date. From? I don't know today and the all 2024. So 2024 and in January, February 25, or we can go as far back as as you want, right? A.

12:29 Rodrigo Cabrera: and here we we have not only like actual vendors, but also intercompany

12:36 Gurjit Pannu: oh, That's really nice.

12:44 Rodrigo Cabrera: On a G. Yeah.

12:48 Gurjit Pannu: Intercom. That's good. No that would be. Yeah I think I mean if you get a year that I did a great date range to have

12:57 Emma Sjöström: Good. Yeah. And I mean,

13:20 Rodrigo Cabrera: Yeah. We?

13:43 Emma Sjöström: No.

13:46 Rodrigo Cabrera: but if we if we go to the I don't know voucher or invoice code level, then it They, it will be a little more tricky or they will, there will be more data.

13:59 Gurjit Pannu: Yeah, I think in terms of priority correct before wrong Emma, it's data first, then it's the entities and accounts and then it's the date range, right? So if we have to have a shorter six months day range, instead of a year on two accounts would be as long, I think the date is the main, the main driver of what we're looking for.

14:16 Gurjit Pannu: Because that's gonna allow us to kind of run run the modeling or run the AI testing in like

14:21 Emma Sjöström: Yes. For. What? I, I'm happy to have like you tell me also, right? If there's anything that you feel like, this is actually important information on an invoice. Or this is typically things that we care about. We don't want to miss that.

15:05 Rodrigo Cabrera: A yeah. I have to think, but I don't think so. They just keep me out of their rooms are

15:13 Gurjit Pannu: That's okay.

15:13 Rodrigo Cabrera: E.

15:26 Gurjit Pannu: Yeah.

15:36 Emma Sjöström: Here, that's a good point.

15:37 Gurjit Pannu: Um, but in terms of vendor information, I think we want all the vendor information in regards to name and amount and due date and processing date, and all that, because that'll that'll be the data that drives the cash flows.

15:48 Rodrigo Cabrera: Yeah.

15:49 Gurjit Pannu: and so,

16:03 Emma Sjöström: Probably not.

16:04 Gurjit Pannu: Yeah.

16:05 Rodrigo Cabrera: To.

16:07 Gurjit Pannu: That's I think that's how I would probably tackle it.

16:11 Emma Sjöström: Make sense.

16:12 Gurjit Pannu: oh,

16:22 Emma Sjöström: Country could be.

16:23 Rodrigo Cabrera: E.

16:41 Emma Sjöström: Yeah.

16:42 Rodrigo Cabrera: A.

16:42 Gurjit Pannu: Yeah.

16:47 Emma Sjöström: In.

16:53 Gurjit Pannu: Yep. Okay, cool. So ideally then Rodrigo would be everything outside of contact information. From a data perspective, The fields perspective on AG on UK. I think are two great entities that have enough of the the mix of activity for us to to explore. And then if you could do a year, that'd be great if that's too much six months.

17:16 Gurjit Pannu: Um and just one pull of that is gonna should give us enough to to start cracking away at this.

17:23 Rodrigo Cabrera: Okay, makes sense makes sense. Hey yeah.

17:27 Gurjit Pannu: I had one more question. What about on the AR side of things? So the invoices you're sending to your, to your customers.

17:36 Rodrigo Cabrera: Yeah, I think that it I think the AR it's a little bit more cleaner. This one AP data vendor data is a little bit working progress because it's not really A hundred percent accurate. If we compare it to To dynamics, especially in the in the due date. Close date.

18:00 Rodrigo Cabrera: There's some like weird stuff going on here. Like I don't know, maybe you like that, you date and it's the 1900 which doesn't make more much sense. A data for AR is much more accurate and the We have some reports ready that I can and download and share the also historical data.

00:00 :

18:25 Gurjit Pannu: Yeah.

18:25 Emma Sjöström: So how do we interpret that? Because in the close date 1900 men, it was it's not closed yet, right?

18:33 Rodrigo Cabrera: Yeah.

18:33 Emma Sjöström: but, Yeah.

18:48 Rodrigo Cabrera: It. But if if not I'll I'll share the data and I'll also tell you how how we look at it.

18:55 Gurjit Pannu: Yeah.

18:57 Emma Sjöström: very,

20:33 Rodrigo Cabrera: Yeah, I don't think we

20:35 Gurjit Pannu: Yeah.

20:35 Rodrigo Cabrera: We have that I just had to go to through an audit request. That was exactly that, they confirming, which

20:45 Gurjit Pannu: Yeah.

20:45 Rodrigo Cabrera: In.

21:01 Gurjit Pannu: I figured Yeah.

21:05 Rodrigo Cabrera: Yeah, but yeah, let me let me think if there's a way and I'll let you know,

21:09 Gurjit Pannu: that's,

21:24 Rodrigo Cabrera: Not yet. Not yet. And then we here we include most of at least for vendors. Most of the Outgoing Payments. Of course, we don't have a salaries in it, like an invoice level.

21:43 Emma Sjöström: Yeah.

21:48 Rodrigo Cabrera: don't have With. But I guess that's for for a second phase.

22:07 Gurjit Pannu: Yeah, exactly. That's another part there. And we're familiar with some of those flows. So this, this is this is great here, like the, for the, for the data part that we're talking about the near term stuff. I'm seeing on this screen share, you're taking out the Intercool a group code.

00:00 :

22:23 Rodrigo Cabrera: I, I just wanted to look at if

22:25 Gurjit Pannu: Okay.

22:34 Rodrigo Cabrera: That. Okay guys, so that too, but we'll see.

22:39 Gurjit Pannu: Yeah. Oh, you got some time to pay that one, but

22:54 Rodrigo Cabrera: Yeah. Let me talk with a team. Let me see if there's already a better report that we can share. If not, I'll put one together quickly and I'll I'll send it.

23:11 Gurjit Pannu: Perfect.

23:12 Emma Sjöström: Really appreciated. Thank you, Rodrigo.

23:15 Rodrigo Cabrera: yeah, and then Yeah, not sure if you have anything else.

23:24 Emma Sjöström: Not really, I think that's what I mean, getting this would be amazing.

23:29 Rodrigo Cabrera: Yeah, perfect. Then

23:31 Emma Sjöström: It's not.

23:33 Rodrigo Cabrera: A, I'll try to bury the same calendar. Maybe not today but they I'll try if not, they tomorrow or Friday, I'll send it.

23:45 Gurjit Pannu: Yeah.

23:51 Emma Sjöström: Yes, well yeah, we're planning to start looking at it next week. If we.

23:56 Rodrigo Cabrera: Okay, perfect.

23:57 Emma Sjöström: So, no stress. Or hopefully that relieves a little bit of.

24:04 Gurjit Pannu: Then your boss is leaving for a while, so I'm sure there's a lot to do before that as well. So,

24:09 Emma Sjöström: Okay.

24:10 Gurjit Pannu: And okay cool. Rodrigo, thank you so much man. This has been this is gonna be super helpful. I think we're excited to kind of come back to you with the results that we find and we'll share those as we as we get those, I get this for you.

24:22 Rodrigo Cabrera: Sounds.

24:25 Emma Sjöström: Well.

24:27 Rodrigo Cabrera: You too.