Skip to content

Sonder - Reporting Discovery - 2024-10-03

Metadata

  • Date: 2024-10-03
  • Company: Sonder
  • External Participants: David Watt (Treasury)
  • Palm Participants: Emma Sjöström
  • Type: Customer Call
  • Domain Areas: Reporting, Variance Analysis, Cash Forecasting, Cash Visibility
  • Recording: https://tldv.io/app/meetings/66feb8a69ff9c800136e3442/

Summary

Context

Customer call with David Watt from Sonder Treasury to discuss their daily reporting workflows, variance analysis processes, and how they manage cash forecasting. Sonder is a relatively small Treasury team (David + Eric, both 4-5 year tenure) using Kyriba as their TMS and Google Sheets for forecasting.

Key Discussion Points

  • Two tracks of Treasury work: active daily positioning (~30 min) and passive reporting (20-30 automated reports daily)
  • Variance analysis is "the most critical" part of their reporting work
  • Importance of tracking forecast versions separately from actuals to measure accuracy over time
  • Reports need to "tell a story" with comments explaining the "why" - not just "see attached"
  • CFO and executives won't log into dashboards - reports must be delivered via email
  • Team uses distribution lists so stakeholders can self-manage who receives reports
  • Kyriba provides actuals; forecasting done in Excel/Google Sheets with quarterly versions saved

Pain Points

  • Reconciling between Treasury/Accounting/FP&A takes time (cash vs GAAP basis)
  • Different category definitions across departments (Treasury calls it all "payroll" vs accounting splits into taxes/benefits/etc.)
  • Understanding the "why" behind variances requires manual investigation and asking other teams
  • Holiday schedules across different countries affect cash timing unpredictably
  • Google searches needed to understand why collections were zero (checking for local holidays)
  • As a younger company, lack of historical data for seasonality patterns

Feature Requests & Needs

  • Automated variance explanations with possible root causes
  • Reports that come out of the system "100% done" leaving more time for analysis
  • AI to identify patterns (holidays, local events) that explain variances
  • Better connection between data sources for faster variance identification

Jobs & Desired Outcomes

Job: Build and deliver daily cash position to stakeholders before 8 AM

Desired Outcomes: - Minimize the time required to compile daily cash position (current target: 30 min) - Reduce the effort required to distribute reports to appropriate teams - Increase the automation of categorization during transaction ingestion

Job: Perform variance analysis to understand forecast vs actual differences

Desired Outcomes: - Minimize the time spent identifying variance root causes - Increase the accuracy of future forecasts by learning from variances - Reduce the frequency of unexplained variances requiring manual investigation - Increase the ability to provide "why" explanations alongside variance data

Job: Maintain coherent forecast across the organization

Desired Outcomes: - Reduce the misalignment between Treasury, FP&A, and Accounting categorizations - Increase the consistency of forecasting approaches across regional teams - Minimize the effort to preserve forecast history for accuracy tracking

Domain Insights

  • Treasury Timing: Daily positioning in 30 minutes, actuals available day after close (vs accounting 2-3 weeks)
  • Report Distribution: Uses email distribution lists that stakeholders control themselves
  • Forecast Cadence: Weekly forecasts, quarterly version saves, comparing original to actuals over time
  • TMS vs Forecasting: Kyriba is source of truth for actuals, but no forecasting done in Kyriba - all in Excel/Sheets
  • Category Philosophy: Treasury thinks in operational categories (payroll = all payroll-related), accounting in accounting categories (taxes, benefits, salaries)
  • Cash Pools: Uses cash pools in US and Europe to simplify regional visibility

Action Items

  • [ ] Follow up on forecasting topic in future call

Notable Quotes

"I think it is the variance analysis. That's the most critical, right? Like just taking the actuals and overwriting your forecast is not smart." - David Watt

"You should have an opinion about it, right? You should tell... Give me two bullets or something. It's a pet peeve of mine, actually, when people send reports and say 'See attached.'" - David Watt

"Every forecast I build, I assume I'm gonna be wrong, right? So it's just a matter of what did I get wrong? And, like, let's learn from that." - David Watt

"If the report came out of the system 100% done... you just thought about what does it say and just spent all your time doing the variance analysis... that's a lot of time valuable to go spend doing other things." - David Watt

Full Transcript

Date: 03/10/2024, 17:30

00:06 David Watt: Yeah. We really have kind of two tracks. I would say the first would be like, for core Treasury stuff.

00:13 Emma Sjöström: Okay.

00:13 David Watt: You know, we're positioning the cash. So every day we look at where we finish the prior day because there can be transactions late in the day that, you know, or the bank will process even after business hours and post to the account, So, in the morning, we always check where we finish the prior day and because we get the feeds from the bank in our TMS.

00:33 David Watt: And then we try to, we have to categorize those. Those transactions to make sure we know, The nature of them, right?

00:40 Emma Sjöström: Yeah.

00:40 David Watt: Because that impacts how they recur in the future, if they will recur. If So there's there's a process there and that pretty quick because we're 90%, you know, kind of things happen as expected. There's just really 10% that maybe have to be adjusted. Um so we try to do that in like 30 minutes or less for my Treasury function each day.

01:00 David Watt: And of course, if we were a bigger organization that went take longer, right, like higher volumes and such but for us, it's less than 30 minutes. So, that can give us our Updated position to start the day and then we have a few calls, you know, different groups to make sure we understand what's going to happen.

01:17 David Watt: So we can Make sure we have enough. if we need to make any movements during the day to cover all the needs, That's probably the. The kind of active piece, there's more. There's a second track for reporting. That's more kind of passed. I would call it.

01:32 Emma Sjöström: Yeah.

01:32 David Watt: Where again, we ingest all the data from the from the prior day and we've built numerous reports to suit various team members, so different groups and accounting. Right there, be one group looks at revenue and one group that looks at payroll or so forth. or so. And we've built various reports to slice and dice all the bank transactions that we receive and distribute them to different groups.

01:57 David Watt: Mostly by 8 am, right? We haven't run on a schedule, so based on the criteria we've set. We probably have 20 30 reports that go out every day to different group of people. We use distribution lists, sometimes, or sometimes it's to a specific person, but that's just kind of But that's just disseminating the information, you know about what happened because

02:18 Emma Sjöström: All right.

02:19 David Watt: everything it's a bank account has to get recorded in the financial statements eventually. So we're just trying to give people

02:24 Emma Sjöström: A nice. Yeah. Like, snapshot of what's?

02:26 David Watt: Yeah, and most of, it's kind of a daily, but some of it runs, you know, once a month, if there's not a lot of activity, we might just run a monthly report.

02:34 Emma Sjöström: All right.

02:34 David Watt: scheduled. So some things run daily, some weekly, some monthly Just to kind of push the data out there in. As useful time frame as possible, right? If they're only reconcile in the account, once a month, Then they only need to get one report once a month, right? They don't need to get 30 individual reports.

00:00 :

02:51 Emma Sjöström: Yeah, of course.

02:52 David Watt: So, we worked with all the teams in advance to set these reports up. And as things change, we can adjust them. We can add certain fields or remove them or we create a brand new report for different things if needed. If need

03:04 Emma Sjöström: Very nice. Do you also work on like a more typical like management reports or Yeah, reporting sense.

03:13 David Watt: Yeah. we do that's more in the active side of things, so our cash position that we build up, you know, we're dropping in the actuals every day so that becomes

03:23 Emma Sjöström: Okay, so

03:23 David Watt: You know.

03:23 Emma Sjöström: come collaborational dialogue with management or like yours, like whomever you are reporting to that's more like on a day-to-day rather than like and a month or

03:35 David Watt: Right. Yeah.

03:36 Emma Sjöström: Yeah.

03:36 David Watt: We're we don't generally report daily balances some upper main, you know, up because that's kind of

03:44 Emma Sjöström: No.

03:45 David Watt: That's my job, right? Like they don't need to know if I'm new on my job, they don't need to know about it. But certainly, we take that information. We compare it to accounting right on a Because we have the advantage in Treasury of knowing how September finished October 1st.

04:01 David Watt: Right.

04:01 Emma Sjöström: Right.

04:01 David Watt: There's no more paint transactions but it'll take accounting two or three weeks to close the books.

04:05 Emma Sjöström: Yeah.

04:06 David Watt: So there's oftentimes people ask us, you know how did it go? We can give them a quick answer but then we have to work with accounting because you know gap is different than cash and so there's adjustments you have to make. So then we have to go back to, you know, Together, we have to go back and say Well The cash flow is X but the reported cash flow is going to be why and here's the differences, right? The reconcile items kind of thing there cash and transit that kind of stuff.

00:00 :

04:32 Emma Sjöström: Yeah.

04:32 David Watt: So, yeah, and then same thing, And then same That's the actual and then we'd also work with the FP and 18 quite a bit, right? Because they're forecasting. Again, the basis of any good forecast is starting with actuals and so they're They're looking at since we can get the data quicker and they are forecasting cash to some degree too.

04:52 David Watt: So we work with them to kind of keep in sync.

04:54 Emma Sjöström: Right.

04:56 David Watt: Yep.

04:58 Emma Sjöström: And what I like. Are there any steps here that you feel like take too long or require too much effort?

05:07 David Watt: Honestly no, the passive stuff. The Is pretty efficient, right? We've built it. We're tweaking it, but we probably Spend a couple hours a month on it right like mostly it runs automatically we've actually trained the people it's on they're pretty well. I think to use distribution lists that they can control who receives them.

05:25 David Watt: Rather than coming to us and having to update. Hey, Joseph left Susan's joined the team. Please update if they just use a distribution of accounting at Sonder, they control who receives that email. We don't have to, they don't have to bother us.

05:37 Emma Sjöström: Right, you don't even need to, okay. You don't even need to

05:39 David Watt: So we put that back on them for as much as we can say, Hey, why don't you create a distribution? Then you can make these changes as you need see fit. Some, there are some exceptions.

05:48 Emma Sjöström: that's,

05:49 David Watt: So, that piece is pretty easy. The active piece again. We can do the daily stuff kind of in a 30 minute window each day so it's not super complicated, we've built it very efficient. I think the more time consuming piece is the reconciling between us and accounting or snf p&a but that's not really a Treasury thing, right? That's we know, very accurately.

06:15 David Watt: What the cash flows were. Sometimes the categories are wrong. We, you know whether it's a payroll tax or a corporate tax art for us to know. So sometimes we have to adjust things like that. But it's more. Trying to convert cash to gap and, you know, or whatever method they're using.

06:36 David Watt: Because we don't you know, we pay, we'll have interest expense but we we don't necessarily pay the interests, pay it we capitalize our interest on our debt for instance. So those are the kind of things. It's the reconcile items to take the most time, but most For the most part reporting is very efficient for us.

06:51 David Watt: We've set it up quite quickly and Saunders, pretty simple. And that's, you know, only have like, 10 countries. One thing I didn't mention is we use cash pools in our, major locations in Europe and in the US.

07:04 Emma Sjöström: Right.

07:05 David Watt: So we don't need to know the details as closely in those because it kind of gets consolidated and we can go at a higher level. So that helps us to

07:18 Emma Sjöström: That makes sense. But what would you? So will help me understand if you do if you find some like with Brian's for example, then between expected to

07:26 David Watt: Yeah.

07:28 Emma Sjöström: happen and what ended up happening? How you go about like trying to

07:31 David Watt: Yeah.

07:33 Emma Sjöström: investigate, what went wrong, or

07:35 David Watt: Yeah, so that one example for this morning, right? I already did. The position is payroll in Canada, came in lower than forecasted. I can just see like the amount we

07:43 Emma Sjöström: Okay.

07:45 David Watt: paid ADP, which is the payroll processor. I I can't see individual salaries or and that's not. I don't need, I shouldn't know that, right? So the question though is Okay. Well great from a cash standpoint, we have more money left over than we thought because the payroll amount was lower

08:02 Emma Sjöström: Yeah.

08:03 David Watt: But the question I have to go figure out now is, will the next payroll be lower as well, right? Is this a sustained change or did something happen where we didn't make certain payment that will make tomorrow for instance or I don't know, you know, usually goes the other way where Payrolls will come in higher and you have to ask, we work with the payroll team and say Well was there.

08:25 David Watt: Like a local bonus or something, right? Like sometimes sales, people get bonuses and it's kind of lumpy So that's the kind of thing where the variance we have to reach out to the team that actually manages that function to understand. We can see the end result. We don't know why and we need the Y to figure out how the cash forecast should be adjusted in the future.

08:44 David Watt: So that's the kind of variance

08:45 Emma Sjöström: Make.

08:46 David Watt: analysis, we do, right? Is to figure out. You know or like collections come in better this week? Sometimes we'll have to. talk with Fp&a like Well, did you Did we expect that like, because again

09:00 Emma Sjöström: Right.

09:01 David Watt: One thing we see is when there's a holiday, our bookings, go up, right? If people are at home with their family, they tend to book more vacations in the future. And so the week after a holiday we'll see better bookings. And so, you know, in the US we're familiar with the holidays but it takes us more time.

09:18 David Watt: To maybe realize there was a UK holiday, right? Or something else, right? That changed the euros or the pounds of the Canadian dollars. On the flip side holidays, we won't

09:27 Emma Sjöström: Yeah.

09:28 David Watt: We won't receive any collections that day. So you'll have like, zero coming in and you're like, Oh, what happened? But it's just coming in tomorrow or he gets delayed a day or something, right? So those are kind of things we sometimes we just Google it, right? If there's no activity, we check to see.

09:44 David Watt: Was it a holiday? Okay.

09:45 Emma Sjöström: Right.

09:45 David Watt: But if there's significant variances to what we thought was gonna happen, we have to go ask people that are the experts there.

09:56 Emma Sjöström: That makes a lot of sense.

09:58 David Watt: Yeah.

10:00 Emma Sjöström: So I'm gonna keep like, asking a little bit similar stuff, just to really, like, get to the core of it. And it's like, Oh, then certain step.

10:06 David Watt: Yeah.

10:07 Emma Sjöström: in this reporting process that you feel are more critical than others.

10:12 David Watt: I think it is the variance analysis. That's the most critical, right? Like just taking the actuals and overwriting your forecast is not Your Smart. So you have to and you.

10:21 Emma Sjöström: More. Yeah.

10:22 David Watt: able track. What you thought for a long time? Some people reset the forecast is as often as every day, right?

10:31 Emma Sjöström: Right.

10:31 David Watt: But I'm a big proponent of at least you know, holding keep track of your forecast, right? You have forecasting actuals, you don't put them on top, you keep them separate and you can track them. so, we can go back and say, well, How accurate were we forecasting? The end of October's balance at the start of October and how accurate were we on the 15th? Hopefully, we're getting more accurate.

10:54 David Watt: As we get closer instead of getting further away or staying the same, We should be learning something. right? So that's the piece I think for us is. We keep both sets of data and we compare them and we can go backwards in time and say Well what do we think at the start of the second quarter or the start of the first quarter? How was this year going to play out? What have we learned? So So, Because that's where, you know, again.

11:17 David Watt: in the end, we have to go back and provide feedback to the other teams and say, well, If they keep telling us collections are going to be 10 and we keep modeling that and it's always coming in at 11. We need to be, I'll give them some feedback and say Look your sandbagging, right? Every time you tell me 10 It comes in at 11.

11:33 David Watt: I'm gonna put 11 in the forecast. And so, but you kind of have to prove that right? You have to save the data. Or it could be the other way, right? If it keeps coming at nine, that's a different problem, but same idea.

11:47 Emma Sjöström: Yeah, that's very true. So, but like how? How do you currently do that today? Then like is this supported by caribou? Like Yeah, do you save the data externally? Like How do you pull these like

12:03 David Watt: Yeah.

12:05 Emma Sjöström: comparisons and how you did like and how you improved and How also sorry not a lot, but how do you also share those learnings and make sure like you keep

12:12 David Watt: Yeah.

12:12 Emma Sjöström: you keep the mind sort of like, does that make any?

12:15 David Watt: Right.

12:16 Emma Sjöström: Yeah.

12:16 David Watt: No, it does. It's the key. So creepa has all the actuals, right? We don't do any forecasting in cariba really. So creep is the source of truth when it comes to actuals, but we say versions of the forecast, right? We set a queue for forecast and we save it.

00:00 :

12:32 Emma Sjöström: Yep.

12:32 David Watt: You know, and then again, we save it here and then we drop in the actuals, but we're building up a Q4 actuals file, right? We don't just overlay them.

12:41 Emma Sjöström: No.

12:41 David Watt: So we can always go back and compare. And then, you know, when we get to And then, you know, when we get to into Q4, we can we create a new, you know, just that's the beauty of Excel, right? Save as so you create another version of it for Q1.

12:53 David Watt: It's got the actual now through Q4 but you're, you know, looking at it. So start a year, obviously, we create a whole whole year so then we're, you know, and we do it.

13:00 Emma Sjöström: Yeah.

13:03 David Watt: do it by week. So, you lay out the weeks and then

13:05 Emma Sjöström: Hey.

13:08 David Watt: you're dropping in the actuals over the top, but you can see how you're progressing versus, you know. Well, the start of the year, we thought Q1 would be 10 and Q2 would be 10 and Q3 would be 15 or whatever. And then, You can see well, we did 11 and Q one and nine in Q2.

00:00 :

13:23 Emma Sjöström: Yeah.

13:23 David Watt: So maybe there's I mean difference between, you know, something happened in the last week, Q1 versus the first week Q2. That's really bang on forecast, right? We're not worried about a week variant.

13:32 Emma Sjöström: Yeah.

13:33 David Watt: but if it's 11 and then 12, and then well, It's Q3 gonna be 15 or no. It's probably be higher because we've been missing, you know, consistently so that we're trying to adjust for

13:42 Emma Sjöström: I was gonna.

13:42 David Watt: that.

13:43 Emma Sjöström: How you track like then first. Yeah, exactly. It could be just the delayed payment of some form or like, it forecasted.

13:49 David Watt: Yeah.

13:50 Emma Sjöström: forecasted correctly. But then reality, unfortunately,

13:53 David Watt: Just one day, literally one day can change, you know, like the US is kind of the easy one, for example you know New Year's Day as a holiday Fourth of July is a holiday.

14:04 Emma Sjöström: Yeah.

14:04 David Watt: holidays on the start of a quarter basically, right? So it can throw things off quite a bit. And a lot of country, you know, most countries have holidays on Mondays and Fridays, right? So things can move between the weeks if Monday's a holiday, usually, you'll pay in advance, right? You don't usually get to wait.

14:24 David Watt: Like, if payroll, for instance, is gonna hit on Monday, if that's not a holiday, then they pay the Friday. So yeah, that's the variance analysis, right? That's the piece, that's most valuable in the whole thing, is every forecast. I build. I assume I'm gonna be wrong, right? So it's just a matter of What did I get wrong? And, like, Let's learn from that.

14:43 David Watt: So you have to kind of build it with the variance analysis in mind, right? Okay, here's how my actuals come in things, right? things, right?

14:51 Emma Sjöström: Right.

14:52 David Watt: categories, that's why talking, the departments can Can take longer because we call things. payroll, whether it's payment to me at my bank account or payroll taxes or payment You know, to a 401k or something, we

15:04 Emma Sjöström: Yeah.

15:06 David Watt: call it all payroll.

15:07 Emma Sjöström: Yeah.

15:07 David Watt: Because we're talking to the payroll department about that, but Fp&a or county would categorize, it is taxes or benefits or whatever,

15:14 Emma Sjöström: Of course.

15:15 David Watt: right? So they're gonna have different categories.

15:18 Emma Sjöström: Of course, yeah, yeah, that makes Yeah. sense. That And how does these learnings kind of persist within the team today? How do you like Yeah. Basically, how do you persist the learnings most so that it's not repeated, right?

15:32 David Watt: yeah, well I mean we have a very small team but I think we use some of the same approaches like Aduber where I was previously we had people in three different locations right US Amsterdam Singapore covering the

15:45 Emma Sjöström: Right.

15:46 David Watt: whole globe. Each person each region focused on their region. But We still share the information with each other, right? That's again, we're distributionless help a lot, right? Like, It can fill up your inbox a bit, but I always try to, you know, copy the distribution list when I have a question, right? So that the other department other groups can see the questions we're asking and answering of each other

16:08 Emma Sjöström: Okay.

16:08 David Watt: Because we want to have a coherent forecast for the whole company for the globe, within our department, we want to think about things the same way. So, That's kind of an intentional just management approach, right is? Yeah, I'm going to copy six people and only one person's gonna answer me, but the five others, at least see the question and see the answer and they can at least if they have a different opinion.

16:28 David Watt: They can speak up and say Hello, I don't do it that way in my region, right?

16:32 Emma Sjöström: Yeah.

16:33 David Watt: Oh okay. Well, let's make sure we're doing the same way. Or they can you know, personality wise not everyone wants to speak up but they see it and they see the answer and then they can think Oh I don't do it that way, I better change what I'm doing right? Like there's younger people and more experienced and so different personality types.

16:51 David Watt: But and then, you know, there's weekly calls usually too, right? Where you're talking about, what it, you know, what's going on this week, What do we expect?

16:57 Emma Sjöström: Right.

16:57 David Watt: It depends on the flows but it can be weekly or could just be monthly or whatever it is. But there's definitely a forecast review call.

17:07 Emma Sjöström: Super nice. Thank you, let's move on. Let's see. I have some I have That that. So again, it might be a bit repeating, but Okay, like this then. When you review a completed report. How do you currently judge if it's a success or if it needs improvement of some sort? Does that question makes sense?

17:32 David Watt: when we're talking about forecasting, or just any report,

17:35 Emma Sjöström: Let's we can do, maybe. Okay, like this. Sorry it's the time is stressing me but what I've heard are like the various analysis and focusing parts are important so it could be dark part.

17:47 David Watt: Yeah.

17:48 Emma Sjöström: Maybe I'm not sure.

17:48 David Watt: Yeah.

17:49 Emma Sjöström: for you guys but maybe also the data validation part. but if that's if that's not, you

17:55 David Watt: Yeah.

17:56 Emma Sjöström: know, important to you then let's just ignore that and focus on a report that you like, yeah.

18:03 David Watt: I think in general reporting.

18:03 Emma Sjöström: Really.

18:05 David Watt: yeah, like I want it to tell a story in the whole point of reporting is to make is to enable decision-making, right? So it it's a pet peeve of mine, actually. When people send reports, say See attached, right? Like, there's a report that's attached fine, but You should have an opinion about it, right? You should tell you updated report attached.

18:30 David Watt: Give me two bullets or something, right? Like it, you know, collections were better than expected by X percent

18:35 Emma Sjöström: All right.

18:35 David Watt: disbursements were lower or whatever. Same thing on a variance report, it's not just columns of numbers. There's got to be a column of comments, right? Where you explain the variances in 10 words or less. Like, I don't need a paragraph. You need to succinctly. Explain Was this a timing difference, You know, collections or Monday

18:54 Emma Sjöström: Yeah.

18:56 David Watt: Collection expected to be higher. All right, so that's a temporary variant or is it a, You know, a fundamental change. So like my beginning example here, payroll is lower in Canada. You know, due to head count, you know, attrition or something, right? Like is it gonna stay lower or is it just an odd thing? So, I think that's always when it comes to reporting.

19:16 David Watt: Like It's not just numbers or graphs but it's the. So what about it, right? Like, What does this tell me? I think No matter what level you're at, you should do that, right? Like Otherwise, we're gonna get. Eliminated by AI, right? Like eventually we all anyway but you want to delay that time, as long as possible you got to put, put your

19:37 Emma Sjöström: Yeah.

19:39 David Watt: opinion behind it and that's I think the other piece Is the forecast is an opinion about the future.

19:45 Emma Sjöström: Of course.

19:46 David Watt: the forecast, it better be your opinion. Like that's one of the mistakes. I see other people make is Well, someone so told me it was gonna be this. I'm like Well, do you believe that? No, they're always wrong. Well then don't send the forecast out with something you don't believe, right? So, It's got to be.

20:02 David Watt: You know for reporting in general, you have to believe the numbers in there, if you don't believe them, don't send it. Like you need to go do more research to make sure you're comfortable with it. Because, yeah, I will check right like It's easy to check Do the balances foot, you know, like that's kind of basic stuff.

20:18 David Watt: Once I get comfortable with whoever sending reports, I don't always add up the math but to start with, that's an easy check on. All this financial stuff is just Making sure the balance is tie out.

20:27 Emma Sjöström: Yeah. But that makes a lot of sense. but so, like when we're on this topic ish, are there any Metrics or feedback that you wish you had so that you could kind of track that your reports are meeting expectations from you or your stakeholders.

20:49 David Watt: Mmm.

20:50 Emma Sjöström: Or do you already have some? I know you're 30 minute limit is one right? Maybe like but

20:57 David Watt: Yeah, that's more of just a informal thing, right? Like,

21:01 Emma Sjöström: Yeah.

21:01 David Watt: If it takes you longer than your, we need to rebuild the spreadsheet or something. Right? Something we got to make the system better so we don't spend a ton of time like I want again.

21:10 Emma Sjöström: Yeah.

21:10 David Watt: Same idea, get the data quickly and then analyze it and spend the majority of your time thinking about it, not just copying and pasting things together, right? Updating formulas. so yeah, that's I mean again with With improvements in AI and just formulas into, you know, million more elegant formulas and stuff like that.

21:31 David Watt: It'd be best if the report came out of the system, 100% done. And then you just thought about what does it say and just spent all your time doing the variance analysis, right? Like 30 minutes to put the report together. It's kind of the the maximum we could save 30 minutes a day.

21:45 David Watt: That's two hour, two and a half hours a week, right? If you can get it, that's a lot of time. That's You know, valuable to go, spend doing other things. so,

21:55 Emma Sjöström: Yeah.

21:56 David Watt: I think again, we're pretty small and we've been doing it for a while, so Eric, you've, I don't know if you spoke to Eric, but he's my Colleague.

22:04 Emma Sjöström: Yeah.

22:04 David Watt: He's been here, four years. So I've been here five, like we, we understand a lot of this. It's harder when you're brand new to understand. Well, who do I go ask about this and stuff, but we both been here quite a while. So In our particular example, it's pretty efficient.

22:22 David Watt: and again, with That's the thing with with younger companies. We don't really spend a lot of time doing kind of year over year comparisons and things like that. We're more established companies would to say like well what's the seasonality and stuff like Our business is still too young probably to have much seasonality, right? If we add a new building, it's gonna have a bigger impact than then, you know, Then how last year went versus this year.

22:50 David Watt: So as we get more established than you can lay in the seasonality, and try to figure out, A better way to forecast, but we're We're not quite there yet.

23:03 Emma Sjöström: Oh, that makes a lot of sense. So sorry for jumping just around a little bit now, but in an ideal world, when it comes to delivering reports to stakeholders.

23:17 David Watt: Yeah.

23:19 Emma Sjöström: How would that happen like building terms of the format of the report? How it's delivered like anything really is there? Do you have any thoughts there or wishes and hopes and dreams?

23:33 David Watt: Yeah, I mean again I think for me like I don't think we're gonna have success. Having people above me log into a system to get a report. I see, you know, a lot of like dashboards and things like that, get built into these systems, but that's great for the Treasury operators, but you're not gonna get your CFO to log in and look at a report, right? They want it delivered to them whether and it's then it gets to personal preference, right? Do they want it, you know, 10 years ago maybe printed and put on their desk?

24:01 Emma Sjöström: Right.

24:02 David Watt: Email or now you know we use a sauna a lot so it depends on how they want to consume the information. consume the So I think that's the format, you know, and how we're getting to them. I think again I already mentioned it but To me, it's a huge thing.

24:18 David Watt: Is, put some comments in there. Tell them what they should think. Is interesting in here. Don't just say, see attached. We definitely still have some reports that come to us with the see attached or they just re Restate, you know, 10% variance in collections last week or something, right? It's like, Well, I can see in the report, it said, 10% That doesn't tell me anything, right.

24:40 David Watt: But why is it 10%, What do you think happened here? That's what I truly I get my team to do is like, put the Like why should they care about this, right? Do we expect 10% collections to continue forward? Or was that a pull forward from this week or just a one-off holiday that doesn't repeat or like try to find the why and the person you know you're sending this so they can make a decision right with cash.

25:04 David Watt: Often, we're trying to figure how much should we spend, and how much you spend starts with much? Do you have how much do you expect to have? So, those are the kind of questions I try to. Think of step forward, right? Try to get the team to like Why is someone using? Why are you saying this report in the first place? Is that person needs to make decisions so try to help them make a decision.

25:22 David Watt: They may not agree with you and they should know more than you, right? Because they're higher up the chain so you could say It's X, you know?

25:27 Emma Sjöström: Yeah.

25:29 David Watt: hey 10% because of something and they go That's actually not right, you know I We purposely did this and it made it but I think at least you're attempting to explain. You've thought about it, you've done your research, they can have they have more information. And

25:43 Emma Sjöström: Right.

25:43 David Watt: hopefully are smarter than you if they're higher up the chain but at

25:47 Emma Sjöström: It's always.

25:47 David Watt: least try. Yeah, you hope, right?

25:51 Emma Sjöström: Yeah, exactly. Sometimes, let's leave that.

25:55 David Watt: Yeah.

25:56 Emma Sjöström: And that's I was right. So they wouldn't maybe login to some dashboard or whatever, but I thought it was interesting. Just the collaboration around getting to the Y with other teams and with

26:07 David Watt: Yeah.

26:08 Emma Sjöström: other colleagues,

26:12 David Watt: Well, when we receive lottery,

26:12 Emma Sjöström: Yeah.

26:13 David Watt: from Creeper right now, the Treasury team does, right? We do just get data sent to us because again, it's easier to take the Excel and paste it into our Google sheet. Then have to go log into the system

26:22 Emma Sjöström: Right.

26:24 David Watt: and download and then paste it, right? It's just sit in the inbox right away when you get in. So, that's again, the system just produces data, but then then the humans take over and build, you know, we could make it more efficient to shrink that 30 minutes down to 10 minutes, whatever to connect the data pieces.

26:41 David Watt: And then, that's where. Yeah, honestly, that's where AI kicks in. Is the next piece is The variance explanations and the adjustments to the future right now, that's humans. but, yeah, with the payroll example, I don't know, you know, like how would AI We can see the balance drop but it's it is kind of a critical thinking, you have to go.

27:03 David Watt: Ask a question to find out. Is this, you know Why did the band? Why did the amount drop? So,

27:09 Emma Sjöström: Where do you see, like,

27:10 David Watt: I guess you can connect into the payroll system eventually and start to see how many you know, Canadian employees are left or whatever but

27:16 Emma Sjöström: Yeah.

27:16 David Watt: we're nowhere near.

27:17 Emma Sjöström: what do you see Aisa first, like a first Use case that it could help you like Well what would that like a concrete? Just like I know we just have two minutes but minutes.

27:27 David Watt: Yeah, no, I mean I think it should all right. Replace much of what we do, right? Like again, you could save 30 minutes a day on our reporting stuff, right? You can identify the variances more quickly because you have limited

27:39 Emma Sjöström: Right. So identifying

27:41 David Watt: number of categories, basically? Right? Like we have to summarize into Cate.

27:43 Emma Sjöström: Areas.

27:44 David Watt: humans, but humans, but you know, a machine can have a million categories so it should be easier to do the variance analysis, when you can individually categorize every single thing, In my mind. Anyway,

27:57 Emma Sjöström: You know.

27:58 David Watt: So then it gets to that variance quicker. It can understand the holiday schedules. It can understand even, you know, that's where Probably AI can figure out. Is there a World Cup game happening in that city or whatever? Right, that's another one. That's hard for us to figure out is local events, but that's all knowable we Google it.

28:18 David Watt: When we don't understand what happens. So, yeah, I could do that too, right? So Yeah, I think. It can. continue to streamline, like In seasonality and stuff like that gets a lot easier, right? If it's just, it's laying on historical data and stuff like that. So, It, it can tap into more data sources than we can as humans, right? So, in more quickly and start to give us like, You know, hey collections are up 10% You had eight percent more available units.

28:47 David Watt: Your reppar was 6% higher than last year or something, you know, I can start to give you possible explanations, and then you can kind of figure out like, well, is, which one is correct?

28:58 Emma Sjöström: Right.

28:59 David Watt: Kind of a multiple choice instead of an essay response on on your test, right? It's a little bit easier when you have Four Things to choose from.

29:07 Emma Sjöström: For sure. So it could provide it could hold more context around about your business and the outer world in memory and from

29:15 David Watt: Yeah.

29:15 Emma Sjöström: And from the kind of, yeah.

29:18 David Watt: Yeah, it could give you the options to go investigate, right?

29:20 Emma Sjöström: Exactly.

29:22 David Watt: Yeah.

29:22 Emma Sjöström: All right. Well I'm not gonna keep you any longer. Thank you so much. This was helpful and yeah, I hope we

29:28 David Watt: Okay.

29:31 Emma Sjöström: can stay in touch and let me know if I if you have any questions after this. Yeah.

29:38 David Watt: Okay, yeah, happy to help. I'm off next week but if you need to ask more questions week after him, I'll be back. So just let me know.

29:47 Emma Sjöström: Super, thank you so much David. I might I might I might get back to you some point around full costing. That would be very interesting.

29:54 David Watt: Sure, all right. All right.

29:54 Emma Sjöström: Yeah.

29:54 David Watt: Have a good day. Have a good day.

29:56 Emma Sjöström: No, thank you. Good day. Bye.