ON - Scenario Planning & Stress Testing - 2026-03-04¶

Metadata¶

Date: 2026-03-04
Company: ON
External Participants: Lucia Galan Caceres (Treasury), Amanda Mitt (Treasury), Yulia Ershova (Treasury - IC & Investments), Rodrigo Cabrera (Treasury)
Palm Participants: Emma, Sarah, Simon
Type: Customer Call
Domain Areas: Scenario Modelling, Cash Forecasting, Bridging, IC Activity, FX, Investments & Debt, Cash Visibility
Recording: https://tldv.io/app/meetings/69a833e9ff9b0400135d2b6d/

Summary¶

Context¶

Deep-dive session with ON's treasury team to understand their strategic needs for scenario planning and stress testing. Palm showed the scenarios prototype. ON has growing maturity needs — moving from "we have plenty of cash" to needing rigorous minimum cash thresholds, working capital steering, and CFO-level guidance anchored in data.

Key Discussion Points¶

ON uses Palm at three levels: Lucia (group/global), Yulia (entity-level), Amanda (transaction/categorization level)
Strategic importance: ON expects more questions from CFO about minimum cash, M&A capacity, and safe cash thresholds — needs data-driven stress testing to feel confident providing guidance
Controlling team provides 5-year P&L/budget plans; treasury wants to sense-check and stress test these using Palm's direct cash data
Intercompany payment terms and reliability are a major pain point — want to model impact of late/changed IC payments on entity and group cash
New store openings (e.g., Sweden in April) are a concrete near-term use case with zero historical data
FX interest is operational (missed IC payments creating currency exposure), not rate forecasting
Working capital: simple 2026 objectives — collect AR 100% on time, pay AP on time not early — but hard to measure
BigQuery/ERP data connection is a prerequisite for long-term scenario planning (indirect-to-direct bridging)
Regional treasury teams could use scenario tool in live meetings to negotiate entity cash buffers

Pain Points¶

Can't bridge controlling's indirect P&L plan to direct cash forecast without manual work
Intercompany payment terms not enforced/automated — Yulia spends significant time chasing IC positions
Regions overestimate cash buffer needs; no data-driven tool to challenge their assumptions
Can't yet show Palm forecasts internally because they're not based on same assumptions as controlling
No visibility on new store cash flows (no historical data to train on)
Tariffs and ad-hoc external shocks are hard to model

Feature Requests & Needs¶

Timing shift assumptions (delay/accelerate inflows or outflows by N days) — not just percentage adjustments
Dynamic minimum cash buffer line — a forecast-derived threshold (e.g., "2 weeks of upcoming payables") overlaid on the forecast chart. Must be calculated from the forecast, not hardcoded — moves as the forecast changes and when scenarios are applied. Used for: CFO M&A capacity guidance, regional buffer negotiations, stress testing under scenarios
Entity-level scenario modeling (bottom-up approach for validation)
Multiple LEGO-style assumptions combined into a scenario
Before/after breakdown showing category-level impact of assumptions
Natural language / prompt-based interface for scenario input (future)
FP&A data ingestion to enable long-term scenario planning
Proactive alerts when expected IC/FX flows don't materialize
Goal-seeking: "What needs to happen to reach X cash target?"

Jobs & Desired Outcomes¶

Job: Stress test the liquidity forecast under various business assumptions

Desired Outcomes: - Minimize the time required to validate controlling's budget assumptions against direct cash data - Reduce the uncertainty when advising CFO on safe minimum cash levels for M&A decisions - Minimize the manual effort required to model the cash impact of intercompany payment term changes - Increase the ability to quantify the impact of business growth (new stores, inventory expansion) on entity-level cash - Reduce the time required to respond to ad-hoc management requests for cash impact analysis (tariffs, FX, delays)

Job: Determine and defend minimum cash thresholds per entity

Desired Outcomes: - Minimize the time required to calculate data-driven cash buffer recommendations per entity - Reduce the frequency of disputes with regional teams about required cash levels - Increase the confidence in minimum cash guidance provided to CFO and management

Job: Validate indirect P&L budget assumptions using direct cash forecast as independent sense check

Desired Outcomes: - Minimize the time required to sense-check controlling's 5-year budget against direct cash data - Increase confidence that FP&A assumptions are realistic by cross-referencing with cash flow patterns - Reduce reliance on mental calculations when assessing whether budget assumptions hold - Increase ability to feed budget-level assumptions (e.g., +40% inventory) into the direct forecast and see the cash impact

Domain Insights¶

ON's treasury team operates at three distinct levels: group (Lucia), entity (Yulia), transaction (Amanda) — scenario tool must serve all three
"We have a lot of cash" has historically masked urgency for precision, but growing maturity and potential M&A are changing this
Kiruba (Kyriba) is used alongside Palm but lacks visual/analytical depth for category-level analysis
Regional teams own entity-level cash calculations today — Palm could automate this and shift conversations to data-driven
Controlling plans P&L for 5 years; treasury wants to challenge this with scenario planning
Internal P&L structures exist at some companies (Decathlon example) where FX variance from wrong volume forecasts is charged back to regions

Action Items¶

[ ] ON team to identify 1-2 most valuable entity-level scenario use cases (e.g., store opening)
[ ] Schedule deep-dive interview on timing shift workflows (how they model today)
[ ] Palm to explore timing shift support in scenario architecture
[ ] Palm to explore minimum cash threshold overlay on forecasts
[ ] Consider before/after category breakdown in scenario preview UI

Notable Quotes¶

"I think in the next years, we're becoming a much more mature company. And I think we will need to do more stuff with our cash. And so I expect more questions to come from our CFO" - Lucia

"I can see that we're going to need to provide anchors to the management on figures. And I want to feel safe that we've stress tested this enough to really feel this is a secure spot." - Lucia

"With the stress testing, it feels like something that we dreamed about, you know, like one year ago." - Amanda

"Even in the worst case scenario of plus 40% cash outs, you're still fine. With this, you know. So that's a nice use case." - Lucia

"Sometimes it's difficult to understand how good we are at [collecting on time]. So we rely on the regional team's report." - Lucia

"This idea of minimum cash... we never want to have less than two weeks worth of payables... especially when it's not hard coded, but rather calculated, it really helps in the conversations with the region to say, hey, you keep telling me that you need two hundred, but what we see is that you would be okay with 50." - Lucia

"Our controlling team plans P&L and budget for the next five years... it's very difficult for me to tell if it's realistic or not because it's just the consequence of a lot of other planning. So this is where I use Palm a lot just to do a sense check that is more data driven than my own mental calculations." - Lucia

Full Transcript¶

Introduction¶

Emma: Hi. I'm. Just try and get the tea to work. Hello.
ON Team: Long time no see.
Emma: Very long time. Now say okay. Start at TLDB recording as well. I think that's nice. It's a bit slow. All right. Sima. Have you met everyone on the team, by the way?
ON Team: A few of them.
Emma: Sarah. No. Let's do a quick intro round.
ON Team: Hello, taker. I'm so funny today.
Emma: It's Glad we still have some social skills left, considering all the remote working.
ON Team: I don't think I have. Esther. A little bit of a broken brain and sleep deprivation. It's a win, win. Let's see. I'm sending some skills to Nils. Do you guys want my cloud skills?
Emma: Well done.
ON Team: No, I said I already got them. So yorish. Thank you. Very good.
Emma: I was just curious what the skills are.
ON Team: It's basically an entire programming workflow. Let me take them step by step. Any skills. We have a skill to add skill. We have a skill to continue work. After starting a new work tree, we have forecasting, debugging, We have finishing work, we have getting customer, we have getting environment, we have reading PR comments, answering and fixing them, we have review with various roles and starting bug fix and then copying environments and terraform.
Emma: Very cool. Hey, julia.
ON Team: Hello. Hi, everyone.
Emma: No. Worries.
ON Team: I'm just going to grab a glass of water.
Emma: Yeah. Dude.
ON Team: Now I'm with you.
Emma: Hey.
ON Team: How are you doing?
Emma: Good. Thank you.
ON Team: Perfect.
Emma: Lots of stuff going on. How are you?
ON Team: Yes, I'm here. So I'm here. I was one week off last week, so coming back with nice. It's always nice to be relaxed and energized during the holidays. And have a bit of more energy. And the sun is everywhere now, so it's perfect. That gives even more energy. We like the sun. Yes. How is it your place, Is it similar?
Emma: Yeah.
ON Team: It's not warm, but it's sunny at least. Exactly. Well, that's already a first step. Oh, we cannot hear you.
Emma: No, you're muted.
ON Team: Sorry. Sorry. I had to say. Before we start, Simon, I hope Janni's passed on the great feedback we got from the presentation from Paul and Christoph. So it was very much appreciated. Thank you so much for being there. Appreciate you appreciating me. Happy to help.
Emma: That was difficult.
ON Team: Thank you.
Emma: All right. I'm guessing you just came out of the call with Jen. Yeah. So why are we here today? We're very curious about your. You mentioned, Amanda, that you have a focus on scenarios and stress testing that you would like to. Maybe go deeper on what it means for you, why you have that focus now.
ON Team: Y.
Emma: But before that, maybe let's do a super quick intros. If anyone hasn't met anyone here yet, I think. Sarah, would you like to begin?
ON Team: Eah, yeah. My name is Sarah. I work as a software engineer at Palm and I've been working on various things like the investments and cash POWs and other things that we have and super excited. To hear your thoughts about scenarios as well. So nice. Welcome. Have you been at pub for a long time? I've been here for a year, and it's quite a long time. Very nice. Are you also based in Stockholm, Sweden or embas in Stockholm. Okay, nice. To meet you. Nice to meet you guys.
Emma: Yeah, sarah.
ON Team: Sarita.
Emma: Sarah's doing a lot of work behind the scenes, so I think it's really fun.
ON Team: I can sell.
Emma: That she's with us today and gets to learn from you firsthand.
ON Team: That's so cool. So we'll do a quick round of introductions. I'm Lucia in the treasury team. And maybe I give you the introduction of where I use palm. I use Palm to check the cash positions and to do my first. So this is my favorite use of palm right now, just checking where the cash is at and how it's developed over time. So I really appreciate the reporting side of things. I'm also looking more and more into the forecast and the variances. And yes, I think it's really cool that we speak about scenario planning, because yesterday I was telling Amanda how key it is and how amazing it is to have Palm to challenge what we see from controlling in terms of planning. So I know that you only have certain set of data and confrunning has another set of data. But we really love to challenge what we see on their side with what we see on your side. So being able to do scenario allows us to tweak a bit what we see and stress test what they take for granted. So super, super curious to hear more. Nice to meet you, Sarah. So, Amanda, I joined on almost three years ago and yeah, it's been really nice because I work a lot with I think, I think after. I don't know. Three months after I joined, we started working together. So it was cool to see the company evolve. Also with my journey at on, so it feels like it's was the same. And on. Also, like, sometimes colleagues, I think. I think it's a really nice collaboration and I've learned so much about technology. I think I see your name popping up sometimes in the group, like, oh, let me select this Sarah. Let me. Sarah from Emma. So it's nice to put a face to the like. Yeah, I know. And yeah, I think for me it's been a mix. I use a lot. I used a lot poem as well for some of cash management like Lucia mentioned. But I think now it's really trying to understand what the team needs, what we need, how can we push and move forward with it feels. Like we're always ahead, I feel already. But it's nice that we worked a lot with the categorization and now we're in a good place to start checking more the forecast and the scenarios. I think there's a lot to improve. In the categories. You know, you've seen the cash flows and the intercompany calls that are so full of volume. But, yeah, I think it's nice that we're, like, reaching that next stage together. With the stress testing, it feels like something that we dreamed about, you know, like one year ago. Yeah.
Emma: Nice.
ON Team: Perfect. Then I'll go next. So I'm with on for almost one year and my main, let's say responsibility or within. Within the team. Is related to inter company and to to invest. So this is also where I use £ quite a lot, potentially more on a granular or entity by entity level. So I'm the PK one who might be coming with small things, but which are quite crucial for me because what I would do is I need to to do certain planning on the financing of our subsidiaries. So I need to understand what is their budget for the year, what are, what were they spendings? And I love the visibility in Parliament on the categories and basically how easily I can differentiate what is. Yeah, what is Capex, for instance, or what is, what are any other payments what are incoming. So quite often I come up across different also categories which we need to change or amand or how somehow improve and also check it with Kiriba. So Cariba gives a certain advantages, but unfortunately not to that visual level, which can allow me straight away to either extract certain things based on the description or category or whatever. So this is something that, that I use quite a lot. And also it helps even us to bring an extra layer for capex. Right? We have some, some limitations how this segregated even in financials or anywhere. So seeing the cash flows and being able to see that in that category, it's also quite became quite crucial, which we wouldn't expect. And yeah, with investments as well. This is also part of. The tasks which we have. I would say don't use it that much palm yet, but I'm looking forward because I see now it is, it is changing. So this is, this is the next step. Where I would like to. To get more into the details.
Emma: Super nice. I think you've all met Simon, then. Yes.
ON Team: Have you met him, Yulia? I don't remember. Introduce myself briefly. So you see, the. The forecast line. That is what I'm doing.
Emma: Okay?
ON Team: At Tom and various other, like, data stuff, I guess. And a little bit of packet. Nice. Nice to meet you both. Likewise. Nice to meet you. Too.
Emma: Super nice. I would actually like to just dive straight to it. We have an hour. Or 50 minutes. Left. And I'm really keen on understanding the strategic importance of this for you. Where it's coming from, why now, all of that and what do you expect? To drive with it. Like, what are the outcomes you're looking for? Using scenarios, stress testing. Call it whatever you want. Right? Like, what are the. Could we even discuss. Like, we would be super keen if we could even land in two to three specific areas. Like, is it? You've mentioned a few already, but is it? I've heard so many different ones. It's effect, it can be working capital, it can be. So, like, what are different types of outcomes you're looking? To to drive with a tool like scenario analysis.
ON Team: I discussed a bit with Jen. I don't know. You can also throw some ideas, but I think a nice use case would be on the minimum cash, for example. And then we could. I think either play with some categories that we know for sure are growing based on. Also something that like now we really have, for example, more surety about Echo Zone inventory, for example. And then we have from the liquidity planning that we're growing 40%. Maybe it would be cool to try to kind of do the group cash outs and then estimate how would that look like? What is it like, for example, what is the minimum cash now? And then add that 40% growth on inventory to see if that will change. You know the forecast.
Emma: Gotcha.
ON Team: In the next quarter. I don't know something like this, but, I mean, it's not based on systematic. For example, I don't know if. Yeah. That's something we need to understand better before creating the use cases. I think that's also could be a good one, yeah.
Emma: I. Think. 100%. Like, let's not be restricted. I don't think we have a great way to proxying inventory right now. Correct me if wrong, but I think that would be needed. For example, but I think, like, let me concern too much with the current restrictions or data gaps.
ON Team: Okay?
Emma: I'd be more keen to, like. Let's maybe back up a little bit, like, the strategic reason. Why are you focusing on this right now? Let's start there, and let's then try to see. If we can arrive, if it's working capital, if it's fx.
ON Team: So I think something that is quite interesting just here in the how Amanda, Yulia and I speak about palm. I think we all use palm at different levels. So, for example, I look at the global, Yulia looks at the entity, and Amanda looks at the transactions because she's trying to help create a better forecasting result. Right. So I think we use different levels. So I think that's for me when I speak about the strategic importance of scenario planning and rather thinking of the global. And then happy to hear what Yulia and I'm having mine. So for me, Let me describe a scenario. So our controlling team plans P and L and budget for the next five years. And then, as a consequence, they have some sort of cash planning. Which I've seen, and it's very difficult for me to tell if I, if I think it's realistic or not because it's just the consequence of a lot of other planning. So this is where I use Palm a lot just to do a sense check that is more data driven than my own. Mental calculations. And so if for me, scenario planning is something that I think of in the long term, like if. Pal has a, let's say a forecast. And I see in the controlling plan that we plan to invest and build up inventory by 40%. Can I type that in and see how that graph would change and see if it matches what I see with controlling, you know, ballpark figure. So for me, it's about being able to input or to add a layer of. Information that I don't think Pan has because it's not even existing yet. It's just predictions from the controlling team and seeing where we land, you know, so this is one use case for me.
Emma: Correct me if wrong, but. But would this be also living some If I say bridging a direct and indirect budget? Is that towards like it would require at least some way for you to, I guess, correct me if wrong, but get that budget data in.
ON Team: Yeah. Y. Eah. So the thing is it's sometimes it's about, hey, can I. Exactly as I said, based on indirect planning, can I get a direct forecast? But that would be me trusting that what controlling is planning is accurate. And then I want to have the bridge. But sometimes I just want to stress test that, so. I say, okay, they are assuming that we get to 2 billion because of this change in inventory. Okay, let me plan that change in inventory in Parliament, see where I get right. So I don't necessarily want to trust it from the get go, but rather say to have another way of checking how our prediction changes. So this is one scenario more on the longer term, more on the validation of. Just the drivers and the understanding of what changes in here makes on the cache. I think on the FX it's actually. I will let me think about it. Why Yulia and Amanda think Because I think there are interesting use cases there. But this is the first one that comes to mind. I don't know. What do you think when you think scenario planning? I think. Yeah, some cases when I think, like you're taking, maybe you can add, like when we see new, like new stores, openings as well. I think it's a good use case for scenario planning. It's something that we have. No, Even if I add, like, either I do it manually, adding the cash outs that I get from the controlling team for, like, for example, new stores, that will be a lot of manual work or, and it's not correct in the end or, you know, or I do scenario planning. I think for the cadence. So sorry. One thing I forgot to mention before I ask you, Yulia. When it comes to strategic importance, I hope. I think until now, I mean, you see, our cash. I'm not disclosing anything confidential. We have a lot of it. So the question has not been do we need more, but. Rather, what do we do? We put it in investments that roll because we have a lot of it. So I think in the next years, we're becoming a much more mature company. And I think we will need to do more stuff with our cash. And so I expect more questions to come from our CFO as well, like what's our minimum cash? Right. So if I want to do an M and A, how much of the cash can I, can I use? I also personally speaking, fear as well. Yes, we have a lot of cash, but a lot of it is in the regions, so we assume certain payment terms. But what if we manage to reduce the payment terms within the company to 10 days? What effect would it have on the cash at onage? So just modeling this kind of intercompany movements as well. Would help me feel safer giving an Amber to suif on saying, hey, anything on top of 300, you're free to use. We are fine with it. Right. So just to keep that reassurance. So that's why I think it's strategically important now, Emma, because I can. I can see that we're going to have. We're going to need to provide, like, anchors to the management on figures. And I want to feel safe that we've stress tested this enough to really feel this is a secure spot.
Emma: Super fair. And I'm hearing correctly, then, that it's actually for you a lot around working capital.
ON Team: There is an element of working capital. Yes. I mean. So we don't really right now, we don't guide on working capital externally. I also think this might change in the future. The working capital guidance still comes from controlling, but again, it's about stress testing this, right? So.
Emma: Exactly. So, like, how would you. Let's say the. The days. Like you have invoices outstanding before. How would you proxy that today, for example?
ON Team: Yeah y. Eah. So we calculate this based on balance sheet, right? Now, I guess the interesting thing would be on the other direction. So if we do something on the working capital, if we say, hey, we want to extend the payment terms by 10 days, what would be the impact? So the impact on the balance sheet controlling us. But what's the impact for us on cash, then that's something we could do with the scenario planning. But honestly. Working capital is not the person that comes to mind right now in terms of. Scenario planning. It's more the in the company side of things. So what if we change the payment terms for all of the subsidiaries and they have to pay now, immediately? And then what does it do for the cash that Yulia has to invest, for example, or that we can kind of clear for investments?
Emma: Right now.
ON Team: Something like this. But again, these are some of my use cases. Yulia, what do you have in mind when you think of scenario planning? I think majority, at least the main ones you and Amanda mentioned. So of course like with a new entities as well. That's. That's quite important because there is no historical data. So this is the only thing what we can input more kind of like as how the company will be developing cost estimation as well, so. Let's say assuming the best case scenario, worst case scenario, potentially like, what will be the cash outflows if something changes? I don't know. Like for instance, even tariffs, right? Like when I was at certain amount might have been added and then want to see what what will be our net position at the end. So kind of like, I would say they are relatively. Ad hoc or cases or dependent on the just situation and needs of the either management or us, adapting so that at the end of the day we want to see what is our realistic cash balance.
Emma: Cool. So even if they're, like, ad hoc, would. Would it. Would there need to be some deliverable? Here that you use internally to discuss around or discuss with your CFO or board even, would you. Like, how would the process look like to get to that? Like, more. Okay. This forecast is now supported by assumptions on top of the baseline that we support, like how what does that process look like getting from modeling that and adding.
ON Team: I would say that if. If we have the model in. In place, right, which we can trust, the main result is that we pick the target amounts at the end. Right. So this, this is kind of. I don't know if your question is related more to the, like, how the report look like or something, but for us, just the basis. So I was speaking about this with Christoph after the presentation the other day because we spoke about how that Slack connection and the reports from Palm and I was saying as of today, because of the data that Palm has and the predictions being based on historical data, I only would want Christoph and us to be able to speak to Pandia Slack because I don't want someone in the, I don't know, in the Stockholm store to think that they can call Pam and this is going to give them a prediction because it's not linked to the controlling prediction. So for me, If you say what would you show internally? First we need to speak about the BigQuery connection. So the rip. And if we want to start looking long term all we would also have to use the FP&A data because that would be the starting point. Right? And then on top of this we can stress test so. But as you said, FP enables indirect. We would plan direct. So it's already. Palm would already help us brix that gap, right, by providing that data. Plus you would have much fancier models than what we have internally to arrive to that conclusion. And then the scenarios come on top. Right. But first, if we want to be able to use this information internally, it needs to be based on the same. The same assumptions. Right? So this plan. So I think the scenario is just a cherry on top that we can say, hey, on top of. So, for example, what if we predict we want to stress test the liquidity and we say, hey, let's assume that the sales caches are going to slow down by 10 days, so every customer is going to pay later. What happens to our cash then? Are we still above the minimum cash or what happens if the customs go up by 10%? Do we still feel safe. Like we that we are above certain level. Like so. Just to answer your question of how this data is used internally right now, I would be able to show it as you listen. I would only take the big number and say I feel safe with 400 million. But if they ask how I get there, I wouldn't want to show the data. Because it's. We cannot breach it to what they have, you know. So I can only say we've done our own predictions and we feel confident with 400, but I cannot use that data internally yet. If you know what I mean.
Emma: Okay? That's to be fair.
ON Team: Yeah. We also heard a lot in the presentation about, like, how do we influence the forecast? Will that be fed from the scenario planning to the like? I don't know, Simon. I think you explained it, I guess. What we want you to. What we want to enable you to do. Is say okay. We think that everything will go down by 10% by this state. Right. Want to feed that into the forecast. And then we want to use the forecast to use, you know, different correlations between different categories. For example, let's say that taxes are related to how much you pay in payroll. And then that change would be systematic throughout with all the correlatives and. And the correlatives. Would be nice. It's a work in progress in a tricky one. Yeah, but that's. It's fantastic. It's tricky, but it's much more advanced than what we can do, because what we can do is do at times 0, comma, 9 to 1 category. That's relatively easy to do. In Excel, but as you said, it ignores the correlation between the different categories. But we're working actively on getting this into good shape. We're talking about forecasting V3, right? So soon? Trademark. But hopefully not too soon.
Emma: Yeah. Yes. I'd still say we need ways to figure out how to do, like, so timing adjustments of inflows, for example. So that's one kind of scenario. Time adjustments. I think that's something we should really take with us and think deeper and how we might be able to model that out throughout our system. Because what we are currently, if we're now going briefly into solutioning space, I would love to go back into discussing the problems and the outcomes you're looking for. But what we're currently looking to support very near term, is more like what you're doing in Excel, but hopefully a nicer workflow. In terms of adding percentage based adjustments across entities categories. So to start you off somewhere.
ON Team: Yeah. Thank you.
Emma: But that's why it's so important. Also, now understand if it is these timing adjustments that are really, really interesting to you, I would love for us to already now start, like, exploring internally how to best support that.
ON Team: Yeah, I think it's a. It's a volume and it's a timing, right? But there is. So I was also telling Rujit Emma because he mentioned effects might be something in your roadmap and.
Emma: Yes.
ON Team: So for us, when it comes to effects, we don't currently have open position, so we don't have forwards, but we might in the future. And I think a very interesting use case for polling is for assist. You're forecasting on a currency level. And you also know the entity level, so to warn us, like, hey, you have an open position in onn game next week. And you were supposed to receive CNY from the Chinese entity, but that payment never came in.
Emma: Yeah.
ON Team: You're short. You have a problem there. Right. So I think that's a really beautiful overlap of cash management forecasting and FX trading. So not really doing the trades itself or the rates forecast, which you spoke about. Simon, let's not go there. But more on the cash forecasting impact. So you assume this, but it's not there. Do you. Do you want to do something about it? And there the scenario can also be interesting. So again, not really doing scenario on what happens if the FX rate goes up and down, which I don't know, I just don't. I don't. I. I refuse to play into this. I've been asked. To provide this. I refuse.
Emma: All right.
ON Team: Yeah, but more on the operational impact on FX flows not materializing. Outside flows, but also insights. You see, I refer a lot to intercompany. And I know you're already working on a lot of intercompany, Sarah, and, and teaming because it's, it makes such a. Has such an impact on our day to day. So also this scenario modeling in that sense, I think would be really cool. Also, Yulia, thinking of the conversations we have with the regions, so just being able to show them like this is what happens if you're late by 30 days. This is the impact globally, right? To be able to quantify this. It's really interesting. Yeah.
Emma: Just explain it to me, like, in five. I really understand the use case proactively. Hey, guys. You assume this inflow is going to help cover this position here? It didn't happen. Do you want, like, the proactive part? Super rare in my mind, but what kind of. What would trigger the need to kind of. Do scenarios and stress testing on these, on these things. Is it because in general, they're unreliable, these flows?
ON Team: Y. Eah. On the intercompany? Yes, totally. Because we don't have. We don't have right now. The, the process is to pay on time, are not really enforced and not automated. And sometimes it's hard to. Convince them of how important this is. And we go again to the situation that we have so much cash that it's been okay so far. But I think if we do bigger things, we need much more. We need to have reliable intercompany payments, and I feel like having the tools to show them the impact. On the total cash position helps a lot and also for Yulia. So Yulia is right now doing a lot of manual work of chasing intercompany positions, getting those payments done. So if we could have that work, those alerts popping up. I think this would also speed up your work, Yulia, the manual part, so you can really dedicate to the strategy and the fixing and so on. There's an interesting thing that I heard today. So I had a chat with the head of treasury of Decathlon and it was so interesting because he was saying that. They have internal PNLs for different functions. And when it comes to effects, They secure a rate to the other organizations and the rate is secure so they can planning the P and L and there will be no impact from effects in that rate. But if the forecast volumes they gave the treasury were wrong, then they get that FX sent to their P and L. So they need to quantify how much of the FX tradings were wrong because of. Because of volumes. And then they give this back to the region. So I was just thinking this would be a really nice way to model this. If the forecast is off for the FX volumes, how short would I be right? On my hedges or how long. And then I can give that back to. To the organizations and say, hey, this. This was your fault. So this is yours. So interesting. Different governance.
Emma: Yeah, 100%. I think. We need to just. There's a lot of things that we can do. And I wonder what's smarter at this point. Is it for me to show you the prototype we have and take it from there? Or is it to have, like, one more round of just, like, thinking if you could? Because I'm really keen on, like, if we could arrive at the one most important use case for scenarios. Like, what would be the most valuable one? Use case. If you had to just choose one. And then go deeper on that.
ON Team: What? Yeah, I don't know if it's scenario planning, but if you ask me, could I upload the PNA plan and see what comes out? But I don't know if that's pure scenario planning or it's rather feeding what we spoke about, feeding FPNA data and then letting influence the long term plan. I don't know. Yulia, like quick, maybe wins with just being able to play a bit. With the data and the categories by entity. For example. Maybe that's. I don't know. That could be a quick. Use case with your prosopagnate. But I don't know if the priority. I think it's something that I think sometimes ergo and you describing a bit on a day to day and I think it could be helpful. Yeah if they could just get an entity as like add some quick estimates instead of Excel, use actual data. Plus some very fine logic other than percentages. I think that could be cool. I think for the testing in general or for the first appearance. I also prefer bottom up approach because it is easier to see or can validate if it is working or not working. So let's say from, from entity perspective that, I mean the ultimate, the most important would be, I would say group. That's definitely but then there are lots of scenarios inside, so that's why. Sorry.
Emma: When you're saying. When you're saying bottoms up, from what level do you begin and group? What do you mean?
ON Team: So for me, bottom up is entity level. So what Amanda mentioned. So I think this is the easiest one to validate if in general the forecast assumptions or stress testing is working. But the ultimate target is to have the total consolidated cash through the group in either all the currencies or just converted in different currencies to see what is like to stress test that amount. But I think.
Emma: To toggle, you can model out the entity level. At Saber, you can toggle your view the impact like different levels.
ON Team: So, Yulia, would this be a then a use case that you would use straight away? If we were able to do some scenario modeling at an entity level with, I don't know, volume assumptions and dates assumptions, would this be something that would be helpful for your day to day managing the entities? I think. I think, yes. Because like this I can at least. Let's say, question or confirm if intercompany is even possible. So, for instance, we have, of course, we have more intercompany payments than usually the cash at the level of the entity. So sometimes it's dependent on the payment terms for some entities. For others, it depends on the how quick they are collecting cash. So for instance, for apoc, it could be also a powerful information to tell them like, you know, what you are planning to pay us within the next three months is actually too low because the expectation of you having higher balances. So again, this will be on the short term entity forecasting because for the long term we still rely on the FP&A for the intercompany lines and so on. So this will be for the short term steering. That's right. Yeah. Okay?
Emma: Simon, did you have a question?
ON Team: Was a little bit of a statement question. But I think a few of these things we can just show the prototype and then. Okay, you can simulate this, but maybe it's not on the right level. And also it sounds like you guys will have use of just having okay, copy this entity and then play around with the values a little bit as well. Maybe that's misunderstanding something.
Emma: Good. It's all good.
ON Team: It's a good question. Statement. I agree. Why are you wary of showing the prototype? Emma, you don't want to bias what we think or.
Emma: I figured, let's do that. It's always a little bit hesitant because I don't want us to lock into this. It's a prototype, right?
ON Team: Yeah, yeah.
Emma: But I think it can serve as a nice sort of point of discussion to just.
ON Team: Yeah. Yeah.
Emma: Share something that Sarah has been putting together quickly. I won't be seeing you guys right now. So just let me know if there's anything. So we have this, like, scenario study studio. Sorry, that currently looks a lot like the forecast overview page. The idea is you still have these filters here. To control sort of what you're seeing. So let's say, Julia, you just want to look at on the G entity, for example.
ON Team: Interesting.
Emma: You try and add. So we have different ways of playing around with the concepts, like a scenario or assumptions. FYI, right now it's not in the prototype, but we're leaning towards scenarios, a set of assumptions. What do we mean by that? We mean that for now, we can just add one assumption here, unfortunately. But imagine you could add, like, a little. You could use assumptions like Lego and combine different ones into an overall scenario. For example, that. The store opening scenario, for example. So please feel free to just go, oh, that's a shit idea. Or.
ON Team: Emma, when it comes to the inputs, I know that this is just a prototype, but what's your end game here? So do you want to have, like, this LEGO bricks already predefined and say, this is what you can play with, or do you think rather from style?
Emma: We definitely want to explore the PRN style or the natural language interface. But for now, just to make sure it's actually supported across the systems, backends everything, we deliberately chose a more like traditional ui.
ON Team: Yeah.
Emma: So that's what we're doing right now.
ON Team: Yeah. I think.
Emma: But there's no reason this couldn't, at a later stage be.
ON Team: Yeah. I get it. And I think it creates trust as well, because when you change one brick of the Lego, as I said, I think it's easier for us to judge if the change that we see makes sense or not. So if you change just the week, then you expect to see one week. You know, I mean, there are inter correlations. On the back end, right? That will not help us. Like, we will not be able to arrive the same result, but we should be able to sense Jack. So I think that's good to build the trust. I. I'm just thinking, for example, for the store. So what happens very often is that we. We start. Preparing for the store, but then the store opening is delayed. For whatever reason. So just thinking of how would we model that assumption, right? So we would say cache installed by two months or something like this. I don't know. And in this kind of scenarios, it's easier to prompt it and say, hey, the store is going to be delayed. I don't know. But anyways.
Emma: 100%. 100%. Let me put it like this. Let's say even if you interact with a system via natural language, Or you interact with the system in this way, we still need to support the same sort of fundamental capabilities underneath. So let's say this one. What I'm already understood, right? It's missing some pretty vital things, such as modeling timing shifts.
ON Team: Yeah.
Emma: For now, all this does is it allows you to set a scope for the assumption it defaults to the same scope you were looking at here. But you can say, hey, it's a global scope. It's all entities. These two entities, and so on and so forth. So some way of defining the scope of your scenario assumption. Being able to say, well, is it like a very near term thing? Is it a five year thing? Is it a 13 week thing? And for now, this prototype only. Only does very basic. Hey. Category level assumptions, percentage based. So let's do it like this. If we're looking at Onnji, Is there any category that would be interesting for you?
ON Team: Y. No. Yeah. Inventory could be a good one.
Emma: What if. This one. Would you like to see if it increases or decreases?
ON Team: Yeah. Let's see the 40% in class. 40%?
Emma: Plus.
ON Team: 40. Yeah, yeah, 35.
Emma: Is this okay?
ON Team: Yeah.
Emma: Okay? And since this is actually calling our backend and doing very like exact calculations, we have this step right now that you click preview. And then it does a lot of calculations, and then it returns something. We've already understood that this UI is not the best preview ui. Because, you see, there is. It's not like you can look at the balances, but you don't really see a huge difference.
ON Team: I think it did. It decreased somehow. Maybe it's minus 30% in this case.
Emma: I think something is probably off. Because. But yeah.
ON Team: It should be even more negative. Yeah, more negative. So if it's more negative, maybe it's because the inventory is negative. So you have to say minus 30. Yeah, minus 30.
Emma: Let's try. So now.
ON Team: Small, but.
Emma: Yeah, you can see it, or like, you can see it with a tooltip, right? But maybe this is not. Top notch visualization and we can.
ON Team: I'm more surprised about the impact. It's really small. Just three minutes. I think we're confusing it. So this is the group balance. So it was right before.
Emma: This is for all my j.
ON Team: I'm confused. So now it's showing. Wait. Scenario.
Emma: It's the blue line.
ON Team: I think. Okay. Sorry. It's lower. Is it? Plus 2 million. So that's. Also, Pam was already forecasting. Or they knew the data for inventory for the last three years now. So maybe part of the forecast they already expected some. Increase in cash outs. Yeah. Without us to manually add 40% or 30% on top. Yeah. I think we need to check with designs and the impact, but it seems to be quite small because I think we pay around okay longer term. No, I think weekly. Let me check. One second. Inventory. Yeah. How much? Weekly cash out. We have, like. Like 30 million. 30? Yeah. Is this for a week? Yeah.
Emma: Okay?
ON Team: Then. Yeah, now it's too low.
Emma: So I guess something that would be helpful here is seeing a breakdown by category two. Maybe if you're modeling on category, like seeing the actual numbers somewhere.
ON Team: For that category that you mollied. Yeah.
Emma: Before and after assumption. Was. Just a thought. So again, this is. This is a prototype. We can change a lot, but we just felt like this is such a. Such a big topic and. Different things to different customers, right?
ON Team: Yeah.
Emma: We just said. Okay, let's start with the simplest possible scope. First.
ON Team: Y.
Emma: But we still want to land it somewhere where it's usable, right? Where it adds value.
ON Team: Eah.
Emma: And I would prefer it if you're very, very honest. You don't have to, you know, be worried about my feelings, so.
ON Team: Y. No, no, I think it's more. I don't know. I think in terms of priorities. I would really focus on. So once we have the BigQuery connection and we have the Arab land data as well in, then this becomes even more helpful because then we can model longer term. Right? So I think that's why I don't think we should be too demanding on what we expect to see here in the in the first version. You know? So we rather give you our examples and then let you decide what makes more sense to develop first. Because I think we're going to use it even more in six months from now when we have all of that data, and then we can give you even better use case feedback.
Emma: Of course. Yeah, for sure.
ON Team: Do you know what I mean?
Emma: And at that point it would be or even before, if you already have a vague idea how you would go about, let's say I know you don't want to do it in Excel, but let's say you used Excel or another process. How would you go about these use cases? Because we are quite keen on learning your true workflows and then see where we can fill the gaps more if that makes.
ON Team: Exactly. Exactly. I think this is what's very interesting. So especially Yulia. Rodrigo, whenever you have to do something manually. That you write this in the palm chat and say, hey, today I had to build this. This. This would have been nice to do in palm. And then you have a definition of a real scenario. That we did.
Emma: Yes, and like. But I'm talking like, even deeper. Like the interview. I know JN is interviewed, right? Rodrigo, you on your liquidity analysis process, was it?
ON Team: Yeah, that was great.
Emma: Yeah. And I love that the output of that interview because what it gives us is we're very transparent with you guys, but it gives us basically a blueprint. Hey, can we support, like, correct output and data in all of these steps?
ON Team: Yeah, yeah.
Emma: Be it, let's say a natural language interface, be it something else. But can we arrive at really correct, trustworthy, allowed data output across all of these apps?
ON Team: Y.
Emma: Right. So it's an easier way for us to operate than sort of guessing across all our customers. Sort of what? What do they want to see and what do they want to say?
ON Team: Eah, yeah, totally. Gu.
Emma: That would be super cool to conduct. Maybe one or two more like deep interviews. But for that, we need to feel like. This is like your top two use cases for scenarios. For example. So that we could go.
ON Team: Ys. That's what I mean. When I think of the scenario, I'm always thinking of the longer term. And since we don't have this yet in palm. It's difficult for me to get there, so. Because I would love to walk you through Emma, one of the scenarios on my side.
Emma: Yeah.
ON Team: Like, okay, I'm being asked to perform this. This. This. This is the data that I'm getting. These are my concerns. And, you know, spoke to how I would do this. But I believe there are a big a few puzzle pieces missing from Palm's data, which I know we're working on. For us to be able to be really able to do it in pub. Right. And show this.
Emma: That would be amazing. I think that that would help a lot and expedite our internal development process of this as well.
ON Team: Yeah. Are you working with other clients on bridging that FP&A data to liquidity? Plan.
Emma: There are other clients that are. We're seeing it especially more in US Based companies. They seem to have a very strong culture of like bridging their FP and A or budget with their treasury. And they have like these recurring things between the teams and trying to. But. But it's more. Not necessarily scenario, but more why was there variance? Why doesn't this happen as we thought like to sort of help steer the business and a lot of working capital location. We need to really start collecting sooner. We need to so like optimizing business processes.
ON Team: Yeah.
Emma: More. Than us as a way to sort of hedge internal risk or if this happens or that happens.
ON Team: Y. Eah. So on the working capital, I can tell you there are two things that we want to achieve this year. On the accounts receivable side, we want to collect 100% on time. And the AP side, we want to pay on time and not earlier. So they're very, very simple objectives. We're not speaking about fancy stuff, but sometimes it's difficult to understand how good we are at this. So we rely on the regional team's report. So they say, hey, we collect 100% on time. We're like, okay, fine, but we don't really know, so this is information. I'm wondering, when we integrate the erap, whether we couldn't also see this in palm. So this is our working capital. Steering this year, which is very, very simple. But it's difficult to get to in terms of data.
Emma: Yeah, I would love to start exploring. Let's say stuff like, hey. Your top 10 customers or whatever, Right? And then see, can we model the impact of those? Like, what if this happens and they are late?
ON Team: Y. Ep. Yep.
Emma: Yes. But would do something like this tool still be useful for you? Let's say we add timing shifts that say we add an easier way to understand what the impact of this assumption was. It like Amanda, maybe. Yeah.
ON Team: I. I think timing shift is actually quite, quite good and quite useful in lots of different cases. So this at this level, right, the entity level. I mean, potentially, potentially should be at an entity level because it is unlikely that there is a shift in timing for all at once. And normally when you check, it's because. I don't know, Rodrigo. Like the case we had last week with the Swedish Corona being short. Right. That was also a. Hey, why. Why are we. Before we predicted to have X, we have Y. Y is this. No. Exactly. And I think then modeling like remodeling based on the scenario. Okay, let's assume that we will still have this delay for a few more weeks. What does it do to our plan? Do we need to. You know, what do we need to do with the S.E.C. so. Maybe something like this could be supported by the scenario.
Emma: Yeah.
ON Team: Yeah, I think we mentioned something similar to Janice last week, right, Amanda? It was mainly around this topic of the new British, for example. We know retail stories coming. I joined late the meeting. I don't know if this was also discord before. No, we did, but it's nice. To get her, in fact, as well. Yeah, I think for some. Like, for this. Yeah, I. Here. Of course. You'll add more categories, I guess, right? Possibility to change more than one category. Emma.
Emma: I think so. What we're playing around with right now is again, like the set of assumptions, one assumption being. Yeah. One set of these settings, like, oh, for this entity, this category, this shift. And then you can have, like, one for cash out inventory. Minus 30%. Then maybe you want to add something. Like if you know, to create something, an event or happening or a scenario. But I would. I think we need to actually maybe do a deep interview with you in terms of the timing shifts. And how would you, like, how would you currently model this out? And go quite deep. Just. Just to make sure we can support it in our data models and in our overall architecture to begin with.
ON Team: That's going to be. Yeah. About the big query, for example, for the new retail stores, do we see already the invoices in Cabrera? For Sweden, for example. Or. Are you also struggling? We don't, we don't have any disability in Cabrera Right of invoices for. I mean, if they are received in their posted in erp, then they should be there, right? But if nothing. So basically everything goes through Cooper for instance, and then from erp. Let's say, if there is an invoice which has never been, let's say, uploaded, then, yeah, we won't see it. But then it comes to the reliability of. And when are they planning to open the store? I think Sweden, like, in April? I think so, yeah. I think this could be a perfect case. The retail stores. So it's super short. Like, it's opening in one month and we still don't have. Even if we connected it with BigQuery, we wouldn't have any visibility anyways. You mean for. For collection of sales? Both. I don't know. I understood we didn't have anything. For sales. Of course not. Right, Because. I mean for short term plan, yes, there should be basically nothing in the sales anyway. Not there. But for payables, I would assume that PayPal, regardless if if store is there or not, they are. They're already contracted. Right. So actually majority of payables are happening even in advance and then just stable ideally. Stabilized when the store is. Is opened with no extra. I don't know, maybe Capex Investments or whatever. I would be curious, actually, to know. So this 30%, this is the assumption on the moment where we are right now, right, saying that it is one off 30% drop.
Emma: It's actually. Yeah, it's actually across the whole 13 weeks and from today, and it's like a new baseline. When I was assuming, today everything just drops 30%. How would the effect be like across 13 weeks? Assuming nothing will change during those setting weeks. So it's.
ON Team: But you can.
Emma: Fair. Sorry.
ON Team: No, I was thinking you can play with the 13 weeks. So you can shorten it to two, three weeks if you want to.
Emma: Exactly. So you could have, like, one in this case, if it becomes more LEGO style, you could have one assumption, oh, for the next two weeks is this. And then. Yeah, that's also an aspect or dimension of timing. Right. From when to when. So that's something we've been thinking about, but we. We want to keep this simple because there's so many things we can add. But if we add everything into the prototype right now, I think it again narrows and limits the discussion on, like, what. What the opportunities are. But yeah. I would love to dive deeper. I know we just have five more minutes left of this session, but. Do you want to have, like, a thing on your end? One or two, like store opening. Love to do that. I just want to make sure you feel like this is valuable for us because we're likely ship something more along these lines. First. And then we are, you know, we really want to go deeper in the AI and natural language as an interface track as well. We'll be more flexible in many ways, but we still need to support the underlying concepts, if that makes any sense.
ON Team: Again, we can discuss a bit.
Emma: Yeah.
ON Team: But for the group, I feel like it's too early. Yeah, I did that. Yeah. So let's focus more on the entity level.
Emma: Yeah.
ON Team: Yeah, if it's your turn. And the data that we have. Yeah.
Emma: Is that fair for you to have a think and.
ON Team: That's good.
Emma: I don't like use up too much of your time. But what I'm asking for is one more opportunity to do more like of a deep dive into.
ON Team: No, it makes sense. Let us think of that scenario and we do it. We walk you through what we do.
Emma: Yeah, that would be fantastic. But thank you so much. I think this was really useful. And it gave me at least a lot to think about.
ON Team: Thank you.
Emma: Sarah. We are sarah.
ON Team: The three of you for your time.
Emma: I just wanted to make sure if. If anyone had any remaining questions.
ON Team: Yes, yes. This. Emma will hate me for this. But, like, do you want different patterns? So let's say you have a slope and you just push the whole thing upwards. For example, the AG was like, so. But maybe you want to say it's going to be sold. Like using the stuff to change. I mean, it's not very data driven. It's more like my gut feeling says this is the slope. No. Yeah, Maybe it's counter. Counterproductive to have even as an option, right? But for me, I like to squeeze the data into where I want it to be. But then it's not scientific, so. It's. It's not as wild as you would say. So I think there. Is also a case where you say, hey, What needs to happen for me to reach this target in cash. Right? So it's not so wild, to be honest, to think, hey, I want this entity to have, especially when you want it down, like, I want this entity to have this amount of cash by the end of the year. What needs to happen so that we reach it's not super out of what we are asked to check sometimes.
Emma: That's very cool.
ON Team: I think, especially for the intercompany part. Right, Yulia? So when we tell the company, the entities like, hey. We agree that your maximum cash is 100. And then they get back and they're like, no, but this is too low. With the expected cash out that we have, we really want to keep more. So I think it's nice to have tools to show to them. No, even in the worst case scenario of plus 40% cash outs, you're still fine. With this, you know, so. Yeah. That could be a good case indeed. So maybe. Maybe Emma won't be mad at me. Maybe she won't be mad.
Emma: I would never be mad at you. No, but there's, like. There's so many nitty gritties. We can really immerse ourselves then. But I still feel like I really need to understand.
ON Team: Yeah. I really. Yeah, I know, I know. I think one thing that I would love to see overall across Palm's predictions, not only in the scenario, it's this idea of minimum cash, which we define as a company. So we say, hey, we never want to have less than two weeks worth of payables. I don't. Know, that's the most simple version of it.
Emma: Sure, but yeah.
ON Team: And then you say, okay, then I'm going to, like, put that line across all forecasts so that you can always check how you stand against that.
Emma: Yeah.
ON Team: Minimum threshold, minimum security. So that's one thing I think that helps a lot. And especially in this, especially when it's not hard coded, but rather calculated, it really helps in the conversations with the region to say, hey, you keep telling me that you need to hundred, but what we see is that you would be okay. With 50. So what are we getting wrong?
Emma: 100%. You like just a derived line based off of the forecast.
ON Team: And these conversations do happen a lot where they say, yeah, but with apac, for example, they say, yeah, but last year it was like, we're growing. They're doubling the business every year. So it's like we're doubling. Plus, there are some assumptions in the plan but weren't reflected. We actually expect even more here. So it would be fantastic to have a tool and say, okay, plug your assumptions, let's see what we get to. And then say, okay, still it's 70, it's not 200, you know. So I think this is a, this is a good case of collaboration. And I think the tool helps have that conversation in a more data driven way. So yeah, I know that we haven't really opened the Palm tool to our regions, so we haven't really let them play with it yet. But we could.
Emma: I think that's a really fun idea. Also, it's substantially overlaid. You can also say that there are assumptions, only they are driven by the forecast. And you can say, hey, show me that overlaid. And you know how.
ON Team: Definitely save save time. Like, even if we don't extend the access to palm to regions even within the meeting, we just run through assumptions together. Boom. And then see the result instead of running manual adjustments. Yes, because this is a process that they own, Emma. The calculation of the cash and their equity level. They own it right now because we also don't have the power, like the capacity to really calculate this. That's why we want to automate that calculation, which Amanda will lead this year. But they're always going to debate it, right? So if we say it's 50 for your entity and they think it's 200, they're going to throw all sorts of arguments at us. And it's just nice to be able to have a tool to say, okay, let's model them. In. And then. So that's a nice use case.
Emma: Love it.
ON Team: And I heard there's no user limit. With film, we can share access before everyone. Yeah, million people.
Emma: Yes.
ON Team: Now we want to be. Want to be careful because we want it to be ready to before we release it. But I think this is a nice way to introduce a tool to them and explain how we forecast, especially for the short term right now, how it happens.
Emma: Of. Course.
ON Team: And then we can check with them.
Emma: Yeah. It's a very cool idea. I like it. Looking forward to onboarding more on users. It's so cool. All right. Thank you so much for your time today. We will get back to you. On this.
ON Team: Thank you, emma.
Emma: Have a think. 12, top valuable. Like what would make your lives easier? Right now. Given something like this. Given what? Like the entity level category. Maybe some timing chips. Like what would be really helpful. I'll take without. Like, we'll take with us also the overlaying of lines that are calculated based on our forecasts. And all of these things, like the. The buffers. I think that's a really cool idea. But yeah.
ON Team: Perfect. Thank you very much. Thank you. Thank you for your time. By. Thanks.