How to consolidate cash plans across multiple companies
Episode 3· 24 min· Charlie Liu and Eugene ChengCash planning · Multi-company
Chapters click to jump
From this episode
Rebuild it on your own numbers
The build guide is a short PDF of what we built on air. It covers:
- The fix in three sizes: a 30 minute Monday check with no software, the budget cycle rule and the system
- Three templates: the brand cash plan, the group cash view with the cases and the AI reading, and the funding decision
- Two Claude prompts, ready to paste: one reads the five plans together, one writes the note to the backers
- What they return: the answers from the live demo, so you can check your own run
Free. One email with the PDF, no newsletter unless you ask.
The problem
A made-up food group (five brands, one shared finance team, about $50M of 2027 sales at a 13% EBITDA margin). The frozen pizza brand lands a national club store listing. Every brand's 2027 plan makes money. Added up month by month, April needs $9.4M and the group can draw $8.0M, a $6.5M bank line plus a $1.5M cash reserve. That is $1.4M short, and $1.9M short of the $0.5M buffer the VP Finance keeps.
- April need
- $9.4M
- Can draw
- $8.0M
- Short
- $1.4M
How we fixed it
- 1One standard cash template for every brand, with its own drivers: customer days, stock, builds, fees paid ahead and supplier days.
- 2The five plans added up month by month. Only April breaks the $0.5M rule, when launch stock, a spring build, a promotion and a cheese make land together.
- 3AI reads all five plans, names the causes biggest first and tests four cases against the buffer rule. Three pass.
- 4The VP Finance picks the balanced mix: $1.5M more headroom asked in November, suppliers moved to 60 days, April keeps $0.7M of room and the launch stays on time.
Want this on your own numbers?
We build fixes like this on real data for finance teams every week. Talk to our team about yours. Not ready for a call? Send us your problem and it could be the next episode.
Transcript
Word for word, split by chapter. Lightly edited for reading.
Eugene Cheng and Charlie Liu, live on LinkedIn. Kestrel Foods is made up, and every number in this episode is generated.
0:00 Welcome: one VP Finance, five food brands
All right, everyone. Again, thank you for joining us for another episode of our speedrun where we listen to your challenges, things that you're doing, and see how can we help you solve it in a couple of days or in a week. Basically, we speedrun your financial challenges.
0:16 The good news: a big retail deal for the pizza brand
So what we're looking to do today and to tell a story here is that imagine yourself as a VP of finance, and then one of your companies, let's say you manage over five companies, one of your companies who sells mixed pizza, mixed frozen pizzas comes to you and say, great news. In November, Costco or Sam's Club actually came to us and say, we would like to stop procuring your frozen pizza and put it in stock. Generally, I'll say that's just exciting news. But what they didn't know is that this, you know, seemingly a very exciting story may actually lead to some financial ruin.
And today we're going to look at what that looks like and how can we sort of medicate and how can we work around this type of situation. So with that said, let's see who we are today. Today we are Maya, VP of Finance for Cresto Foods. They manage five brands from frozen pizza to curried meat, cheese, sauces, frozen burgers, soups. And because Maya is just a one -man army, her finance team manages all five brands. And together, they come up with about 49.8 million in sales, with 13 % EBITDA margin, and every brand makes money, so very healthy.
But being the responsible of EBITDA finance that she is, she also has a line of credit or backing from banks with $8 million in cash that she can use to weather any storms and also if there are any expansion plans. And again, being the responsible person that is, she actually wants to keep a buffer of about $500 ,000 or $800 ,000 just so that she doesn't need to go to the bank that often and make sure that she stays far away from are covering in the agreements.
So that's where we are at today. So with that said, let's see what happened. As we mentioned, one of the brands, the pizza brand, actually came up and said, you know what? We have a really big expansion going on. We would like to have some capital and start planning and what that's selling to the Costco or Sam's food is going to look like. And then Maya being the VP Finance said, great, now that's all five brands, let's do some planning together. And that's exactly what she did.
2:42 One standard template, each brand with its own drivers
She asked everyone in the space to actually come up with each brand's cash plan. So in here, every brand comes in and actually start planning in a centralized model. And in this centralized model, because every brand is a little bit unique, they have all the nuances, but Maya has managed to build a model that fits majority of the brand, but also be able to choose different scenarios that they can pick. So right now let's say we're looking at a base plan where everyone starts to plan their stuff.
Let's move, let's look at the pizza plan. And we'll see that the owner of the previous event has done a pretty diligent job in planning for it. So Minh, where he was just start planning by net sales by different channels and start planning the margins by different channels as well too, what the cost of goods sold out, what some of the cost drivers and what some of the cash needs are. And so not only have they done their own decision planning, they also gives you little bit of explanation of what is it that we're looking to do.
And then not only does his one, this one team does it, every team does it as well too. And we can see in here, this team or this brand has a pretty healthy margin, even with the expansion. Eugene sorry, can you just zoom in a little bit more? So just for the folks who's Yep. watching on phones and all that stuff. It's pretty little bit hard. Yeah, Yep. that's How is much this? better. Thank you This is what happens when you work on a much bigger monitor.
well We're just looking at the Yep, pizza brand right now so we're just looking at the pizza brand, where as a brand owner, he goes in and put any sales that is highlighted in blue, supplied by Maya. Now we can also toggle to a different brand, let's say for example, the Kuremeets brand. This is where a different founder or different business owner comes in and also fills in his own details. And again, the first structure is very standardized. Such as now, when all of this is consolidated, Maya doesn't need to go and spend a lot of time consolidating all the numbers.
Any type of updates here, is then reflected automatically in Maya's reports. So, everyone has done their own diligence, and everyone has come up with pretty relatively healthy results. So, Maya's going to now take all these and consolidate them, and see what happens.
5:24 Adding it up: April needs $9.4M, the group has $8.0M
Minds look at their plan for the base plan, basically the base plan of what everyone has submitted. And let's see what happens. Every plan, every company has come in asking for a bit of cash, which is generally pretty normal. But look at what happened in April, we're actually feeling our upper roof, upper limit of how much cash we are allowed to borrow or how much cash we are allowed to spend. And this makes sense because this is exactly what the pizza brand has told us. They actually want to expand nationally.
And as a result of that, there's a lot of extra spending, a lot of prepayments. And even though we expect the sales to really come in and cover the margins, there's a lot of upfront costs with listing and also putting your goods in front of in a big retailer shop. So in here, this is sort of where the alert comes in. As much as things are looking good for the company, we are at a really difficult spot of not having enough cash to fund it, and this is where Maya comes in.
So we saw that just now by combining all of the good news, so basically five good news together, by combining them together, we actually come into a really weird or complicated situation of we are 1.4 million shots, short of the buffer that we want, and we now have to go to the bank for more money. We have two choices of going to the bank for more money. We can then either have a conversation in April to have the bank to say ask for more money, which could be leaving it at a less advantage or might be as a more difficult conversation just because of the timing of it.
Or we can wait until November, because we actually meet with the bank about once every year or once every six months just to gauge our conversation to see what the business is. So we can have this conversation of having the hard conversation now or delay it
7:26 The levers: delay, squeeze, supplier terms or the bank
until later. But if we delay it until later, we just won't have the cash to cover it. So what do we do? This is where some of the modeling that our team has done together. So what our team has done now is basically to go through every scenario and talk to each of the business partners to say, hey, what if we do things differently? So now we can actually create different use cases to say, what happens if we slow down the expansion? What if we don't roll out to the big spot retailers in November?
What if we have a situation where we squeeze everyone a bit, say for example, everyone to reduce spend, to cut costs, or what if we would just simply ask the bank for more money. So now here we can create different scenarios and different options to try to weather the storm. And after combine all of this, let's see what it looks like. So this is where once everyone has put in the analysis, put in the job, we can now start using some of the latest technology such as AI or other tools to start deciding and looking into how can we get out of this situation.
So in here, I've created, in here, if I want to see, I'm using my choice of code, which is my short of AI, which is Claude, but doesn't necessarily need to be the same. You can use ChatGP or your AI of choice as well too. So I'm going to go in here and say, okay, let me run an analysis on, this is the cache analysis is the, oops, I cannot find it. One second.
9:08 Our flavor of tools: Excel, a planning platform and Claude
Does This everyone is what happens. know what kind of just presenting what these cases actually look like with Kind of tools that we like we all love Excel We like Claude What you saw earlier is a platform combat it up. We like you don't have to use it You could test a lot of this with all that as well It just helps instead of having like five different Excel models for different scenarios for us. It's just a toggle all this there are different flavors of this that can work for you we're kind of just showing our flavor and if you have any questions comment we're happy to kind of showcase what this looks like in other different flavors as well
9:50 Demo: a Claude skill runs the funding analysis
And to add to that is that what I've created here is basically what we call a skill in Claude. Now, a skill is basically you can think of as just a package instructions on how to do certain skills. For example, I'm open to run a skill called Analyze Funding for the Crest of Foods. I generated this skill simply by putting a lot of prompts together to help you help the AI understand the context and the business and what is it that I'm trying to do, how do I do the funding analysis?
And now, as opposed to typing it all over again, we package it into a clock in those clock skills such that every time we do something, it's a little bit slightly more deterministic because now running off the same context and same instructions are supposed to slightly different in how I prompt it. So now that everyone, every company has put in their assumptions and looking for ways to squeeze our cash use a bit more. And now we can run the scale to do the analysis. So this scale would probably take about two to three minutes to run.
So as it goes through this, it will go through, as Charlie mentioned, our planning database on all the exemptions, all the history, all the drivers, and all the comments that people have put in, and start to put together a plan together. But as we know that this is, everything here done is just pure AI. And as humans, we can't really just 100 % just take it blindly. So there's simply a certain layer of approvals or understanding of what it does before we act on it. So what I'm going to do now is while this thing runs, I'm going to go back into my kind of instance, and let's see what kind of input it comes up with.
And in here, a lot of these have been preloaded just because we don't want to waste your time watching a problem run for two two to three minutes, I'm going to suggest generally some of those results ahead of time. So you can see in here, this is the different cases that it has come up with. In the base case, we're missing 1.4 million, or in the case where we're trying to slow down our business, and we're still off by 0.4, but then we saw in cases where we're going to just squeeze everyone or borrow and see and both based on you can see in that those ones actually wants what what is satisfied our our covenant our covenant requirements but at the same time we don't want to just want to squeeze everyone again this is some of the ai generated um results so we can't really just take it with a grain of salt we do have to fully analyze and understand it so in here i was able to see what Eugene yep can you zoom in wow still too small I think it's just because of those texts and
12:36 AI lays out the options, finance picks
uh they're just remembered sometimes i see people yeah watching from a phone right yeah Cool. So yeah, so in here, the AI actually tells us, you know, what are some of the impacts and what are the costs for it. So now as a human, I can now then decide what are some of the things I want to do, what are some of the levers I will play with. And then I can then also take it back to our senior leadership and decide what are some of the actions that we want to do.
Well, for each case, so let's say for this case of slowing down, what is it that you want to do and does it still meet our governance requirements? And for other cases, doing explanation, almost like building a business case of what is it trying to do. So now once all the details and all the analysis done by the AI comes back, we can now start to make that decision. And we can see that very cleanly in one of these. So in here, we see what some of the recommendations are.
13:39 The decision: a balanced mix
We can then now come up with basically different cases of saying, for example, do we want to delay the launch by six weeks, do we want to pay the ingredients or price later? So now as humans, we can now search to almost like a beaver to decide which option do we want to go with. And, of course, we also want to be the one to be able to satisfy our requirements for eight million funding requirements. So now, as we go talk through our business, talk to each business partner, try and figure out the plan that works, and just to spoil the ending, what we've decided with is that we're going to go with a balance mix, where we do slow down some of our payments.
We actually see if we can do some negotiating with our suppliers at the payment terms, But also at the same time ask for investment later in the year. So now it's by combining these different behaviors that are recommended by AI, but also approved by human, that we can actually really come down to a use case that actually passes our recommendations and passes our requirements. So this is sort of what in a nutshell is what we're looking to show today. Basically a combination of planning by each individual business, combining inference input into a single consolidated place using a standardized template or input, having them all saved into the database, just analyze to see what that, what inference input together look like.
So it's not just each individual company, what the results are, and then consolidate them into a field that your team can then make decisions off. And in this case, we see one where we have five successful businesses together, they can cause issues within the financial confines of our governance. So with that said, the last piece is that we use AI to do a little bit of analysis to see how we can get out of this and what are some of the recommendations that AI comes up with. And then finally having us as a human to approve or reject, we showed up these options you want to go with.
15:55 Excel is not going anywhere
I love this this is what you meant by the kind of storytelling interactiveness. It's a really nice kind of combination of kind of the standard in finance, which is Excel. We all love Excel, or hate, but you can't live without it. Even though everyone's talking about Claudia, and we use it day in, day out. But at the end of the day, any sort of fundraising covenants, any sort of investments, M&A, due diligence, all that's still in Excel, so it's not going anywhere. But at the same time, leveraging an EPM tool where it holds the data, give governance.
A lot of this control that Eugene's talking about by locking it down, it has that as well. You can do it in Claude, right? You can connect to Excel, build a whole database, try to figure out an audit show. I'm not sure if most folks in finance account is comfortable with building that. That's why EPM tool is still helpful, especially on kind of collaboration and plus from different folks. Eugene mentioned we have different locations here, different assumptions of the drivers tracking that down. A large part of these days I talk to CFOs is, I actually think Eugene and I have been in this space for a long time now.
We've always had that conversation, which is to buy or to build, which is part of it. But I think here is now it's the world's changing. It's more about the maintenance, making sure that it's maintained properly. It's audit. The audited part is always there. So a platform like an EPM platform that's helped, but at the same time, a lot of your analysis right here, um, happy AI to help you with that. We see a whole spectrum of different finance folks with different comfort levels. Um, You can absolutely use an AI -native tool like Claude, like what Eugene's doing, but a lot of these EPM tools also have native AI tools that can help as well. Different versions, different preference, different comfort levels, but this is a great visualization so far of what this looks like, Eugene.
18:05 Who changed what: the audit trail
Yeah. And also the other part is that as much as AI does a lot of those decisions and also advances for us, we still need a way to communicate back and forth. And there's still no better ways than doing it all on a centralized platform because sure, you can do a lot of this communication with people in Teams, Chat, Slack, or even by right Claude itself to do a lot of communication. But six months from now, you're not going to remember where that can be. Where is that thread?
Is it an email? Is it a chat? And having a platform where you centralize a lot of communication also helps tremendously. So now, six months from now, it's not a project to look for all your proof. Six months from now, you need to look for how certain number came to be. It's right there in the platform versus a whole a project just to take out your information. Cool. So that's where I am at. Charlie, do you have any comments, anything, something that I've seen, you know, like this in your experience as well?
19:09 Many brands, one cash pool
This is a common problem. I mean, these different locations, kind of different entities, different businesses, all kind of managed from the same cash pool, different templates as well. I think this is helpful. So I wonder what, if it makes sense, I should go back to some of those other kind of templates earlier, just slow it down and walk through them a little bit more. I know we're jumping in Yeah, absolutely. and out really quickly. Yeah, Absolutely. So in here we have, because we have different, different, different companies.
So let's say, for example, you might have one that is for frozen, frozen pizza, one for fun, for cured meats, cheese, sauces, all of all, They all have their own unique instances or unique drivers. And as I mentioned, so skim this through very quickly, but we can actually decide what some of the different narrows want to go through, whether it is the base case or slowing down or squeezing everyone. So we can just say, let's see what this is in the case where it's slowing down. We can go ahead and see how will some of these numbers change as well to compare the
20:14 Send two tabs, get ten back
difference in narrow. So, now, as opposed to, you know, having four or five different spreadsheets floating around in your Google Drive or your OneDrive of what the difference in errors are, or looking at version 1.0 versus version 2.7, now everything is in a single spot within your platform, and as you can, as users go in and make certain changes or certain assumptions to their business, anytime that we save this data, and then also save consolidated into centralized database such that when Maya or the VP of Finance of these group of companies need to see where things are at, they do not need to ask you, hey, do you have the latest set of version?
And if you don't have the latest set of version, when can I get it? And when I do get it, is it still in the same format that was provided? How many of you have seen instances where you give someone an Excel file, simple to that, but then they give you this home on a perfect 10 top file that looks nothing like yours. Now you need to spend time, hours, and your day just to do the consolidation. By having a platform, a centralized, innovative EPM platform, you can do all of those consolidations automatically and in real time as well too.
And also be able to see it as an auto trail to see exactly at a certain point in time what changed what and what was the number at any point in time well too. So say for example, we have one in here where it is, they do pantry. So what are some of the net sales by different channels? How will we cut it in margin? Cause good soft margin? What does the gross margin look like? And you know, one thing that's really active or a huge, huge piece in the food industry is what are some of the promotions?
What are some of the coupon codes that are being used? So you can do a lot with these planning as part of your calculations and as important to any business, what's with the cash drivers, each business is going to have their own unique trash drivers as well too. So how many days does it take for inventory to come over? What are some of those customer payment days? So the beauty of this is that you can actually bring in any type of model or create your own type of model within Claude and Excel and bring this in a platform and have it drive your business decision and drive your forecast. I think so. I think so.
22:46 What is next: the same food group for a few weeks
So I know Eugene, maybe talk about the series as well. I think part of this is we're changing up the format a little bit. We're over the next couple of weeks. We're kind of looking at the same business, right? And looking at different use cases with different problems they have. Maybe, can I explain what that looks like? So basically what we're looking to do is just to help us, help the audience understand and give some continuity of we're going to start to use, I mean, let's say the same business or similar business, but using it for solving different problems.
Again, if you have any type of business or problems that you're looking to solve, please send us our way. We can look to see how we can help you optimize your specific challenge. And in this case, we're working on a food grocery or a business with 500 % lines. We'll continue to use that. While we're not too sure what the next ones will be, but it will be going to be something similar. Basically, it will be a problem that we have consolidation, reconciliation, or what multiple entity business is very commonly seen in some of the Office of Finance activities.
Yeah, everyone, we'll be switching up industries, but for now, for a few weeks, we'll be in this industry. So for anyone out there that's in that food kind of beverage, kind of food adjacent business, let me know where are the common things that's keeping you up at night or working late on a Friday or a weekend. Let us know. Eugene and I will show you what that looks like. Within a few days, after in a week, you see what it looks like. Really excited for this.
And this was fun. You do Thank you. And I always enjoy businesses like these because it's a lot more relatable. You know exactly, it's the same reason I have for hospitality. You know exactly what you're looking to, what is it that you're selling and in terms of grocery, it's something that we all relate to.
24:35 Wrap-up: send us your problem
awesome. Cool. Well, with that said, thank you everyone. And then I hope to see you again soon next week. And then again, if you have any type of business, office of finance related things that keeps you up at night, please send us our way and then we will try to speed on it for you next week. Thank you for watching.
Next live: Episode 4, Wednesday 10:45 am ET
Got a problem like this one? Send it in. If we pick it, the fix is free.

