Kat Orekhova is Co-Founder & CEO of Vareto, a planning and analytics platform for enterprise teams. In her own words Kat has an “unusual backstory” leaving the world of math academia to build Facebook’s first-ever data science team within FP&A. At the social media giant she was Responsible for financial reporting, forecasting, and planning activities for Facebook’s core business. This was followed by stints as Head of Product at IronClad, a $3billion legaltech company, Sequoia Capital Scout, General Partner at Darkmode Ventures and co-founder & CEO of Vareto.
In this episode Kat reveals:
- The challenges of long range planning as Facebook scaled massively and trying to predict growth across countries from the US to India with more than 100 people contributing to the planing process)
- Being at the beginning of data Science at Facebook
- The move from data science and product into finance
- When should companies start looking at FP&A
- Being a Sequoia scout and what it entails and what she looks for in investment opportunities
- How (now) Meta CFO Susan Li “an absolutely outstanding” mentor insisted on a finance team with data science skills led by Orekhova
- The right blend between data science and FP&A who are “Living in Different Tools”
- Her take on the amount of data science that FP&A teams need to boost their career
- The impact of AI in finance
Follow Kat on LinkedIn: https://www.linkedin.com/in/katorekhova/
Paul Barnhurst:
Hello everyone. Welcome to FP&A Today, I am your host, Paul Barnhurst aka the FP&A Guy. FP&A Today is brought to you by Data Rails, the financial planning and analysis platform for Excel users. Every week we welcome a leader from the world of financial planning and analysis. Today we are delighted to welcome to the show, Kat Orekhova, welcome to the show. Thanks
Kat Orekhova:
So much, Paul. Really excited to be here. Uh, excited to get to chat with you today about everything. Yeah,
Paul Barnhurst:
I’ll be fun. I’m excited as well. So, let me give you a little bit of background and in a few minutes here we’ll give you an opportunity to share a little bit more about yourself. So she comes to us from Boulder, Colorado. She’s currently the CEO at Vareto. She earned a bachelor’s in Master’s in Mathematics and started on a PhD. And then she, uh, works, she worked for Facebook in the data science team. She worked for Iron Ironclad in product. And as mentioned earlier, she’s the CEO at Vareto today. So I’m gonna ask you a question about budgeting and forecasting. We like to start with this one with everybody. Tell me about the most challenging or worst budget forecast you’ve ever been a part of.
Kat Orekhova:
That one’s an easy one. So when I, when I worked at Facebook, I started in data science, but then later I, I was in FP&A and we had all of our different processes, all kinds of daily, weekly, monthly, quarterly processes, like any FP&A team. The hardest thing by far though, was actually not the annual planning cycle. It was the long range planning cycle, which we would do at a different time of the year. So we never, never really got a break. And long range planning at Facebook meant planning for the next five years for a business that was growing very, very quickly at a very massive scale. And we had to plan everything from how users would grow, but internationally, like different in the us whereas maybe things were more mature as compared with India, where things were growing really quickly.
Based on that, we would have to plan out, uh, data infras centers and where we would physically build massive, uh, you know, just like massive kind of capital infrastructure. And then, um, also revenue and advertisers and like advertising trends and putting all of that together into a long range plan that then over a hundred people would give inputs into. Um, and, and many, many, many cycles of approvals. It was a very, very, very massive undertaking. So definitely the hardest thing I’ve ever been a part of. We would breathe a big sigh of relief every year when that process was done.
Paul Barnhurst:
I could imagine. How long did that typically take you? I’m curious how many
Kat Orekhova:
Months
Paul Barnhurst:
Work on that long range plan?
Kat Orekhova:
Yeah, well that, that’s kind of the, that was the hard part is we, we had to put all of this together and I’d say like in two to three months get to a, a very solid outcome because then we did also have the annual budgeting process and like monthly rolling forecast. So we couldn’t afford to have this take any more time, and yet it required so much work from so many people. So it was, it was one of those, like you sleeping in the office type of situations.
Paul Barnhurst:
Yeah. Never, never fun. So what was maybe the key, key takeaway or learning experience from going through that?
Kat Orekhova:
Um, well, so one, one big thing that I, I really saw was I needed to get inputs from a lot of different people. And this was a very diverse set of people. You know, some, some folks like the user forecast would come from the Facebook growth team and they would produce it running all kinds of different machine learning models. They would produce it, they would give us a query that we could run to pull the latest numbers. And so that was fine. We could always rerun the query or maybe ask them if we needed a longer forecast or whatever it was. But then other inputs came from different teams that would give it to us in different forms. It could be things coming outta Salesforce, it could be, uh, people sending us different types of spreadsheets with various projections that they came up with. It could be people giving us numbers over Facebook messenger, over email.
And, and then every time any one of these numbers changed and I had to go back to five different people and say, well, this person’s numbers change, so how do your numbers need to change? So much of my job wasn’t even about forecasting per se as it was about data gathering and, and kind of getting everyone on the same page and doing that over and over and over as we iterated throughout the planning and forecasting process. So that was like a big aha moment for me of, it’s not just myself doing data science forecasting. And it’s not just about the technicality. So much of it is about the people and the process. Yeah,
Paul Barnhurst:
No, I a hundred percent agree. I mean, there’s so much that’s about coordination and collaboration and streamlining process. I mean, the gathering process can be a nightmare when you have a lot of people. And how do you make sure you do that in a way that’s efficient, especially when a lot of companies are doing it offline or you’re doing it via email and maybe some spreadsheets and you have it all over the place. It can be a real challenge to consolidate it all. I can remember some late nights when you try to consolidate all and the numbers don’t tie. Just like, okay, are they ever gonna tie? Yeah. Am I gonna plug this? What are we doing here?
Kat Orekhova:
Oh, I, I have so many stories of that, like realizing, you know, realizing at the most critical juncture when you’re in a massive boardroom, that there’s one number out of a thousand in your deck that might not perfectly tie. And just having that, that nightmare of a moment in your head when you, when you realize that that happened, um, I, I’d be, I’d actually be curious, Paul, like if, if there’s a quick story you wanna share about your nightmare moment where numbers don’t tie. I always love hearing those. Oh,
Paul Barnhurst:
I mean, there’s plenty of them. I think my worst nightmare moment, and not so much numbers tying, but, so we’d done this big plan and when I worked, I worked in business travel and so we had a ton of different countries. We had partner countries and we had to load everything. And one year I loaded the co country of Poland in the complete wrong currency, you know, and I was like, oh no, I loaded it in US dollars instead of their local currency. The good news was they were a relatively small country, so it wasn’t a big impact globally, but for them, and we’d already closed things so we couldn’t correct it to them. It was a huge deal, right? Their numbers were completely wrong. And I just remember the nightmare of that ’cause I didn’t, I didn’t catch it till it was too late and have to be like, yeah, yeah, Poland’s all wrong.
We’ll have to fix it at the next forecast. They’re just gonna have to live with those numbers. And so, you know, their variances were a total mess. So that’s probably the worst as far as, you know, just having a major issue with numbers. The other one that always stuck out to me is I worked for a business one time. We were way ahead of budget. We’d put in a very conservative budget deliberately ’cause they hadn’t hit the numbers in like five years. And so of course the leadership decided, well, we’re gonna give you more aggressive numbers for the second half of the year. And so they went through and redid our budget for us and then asked us to report to ’em. The numbers made no sense ’cause I had billed everything driver based, they just used some top-down approach. I literally a couple times said in the comments, I can’t tell you what the variance is ’cause I have no idea how you came up with the budget number. And I’m sure they lovef that going through the deck up to corporate. They never got mad at me, but I’m sure there were a few times they just shook their head, like, come on, you gotta give us better than that. Yeah. So that was probably the two that stuck out to me.
Kat Orekhova:
Yeah, I I love that. Thank you.
Paul Barnhurst:
Yeah, it was fun. It not at the time. It’s fun looking back now, you know, you can kind of laugh about it. I’m sure you had those with Facebook too, where now you look back and laugh, but at the time you’re just like, okay, can I take another day of this
Kat Orekhova:
Absolutely. I think we’ve all been there and that’s why it’s, it’s so great to hear these stories and realize you’re not alone. Everybody that’s ever worked in FP&A and a has seen all of these things. Like, they understand it. And the exciting part for me is being in a position now where I, like every day I’m trying to improve this situation and I’m trying to see how can we automate, how can we get to a, a world where these types of errors get minimized? And uh, that’s, that’s been really fulfilling.
Paul Barnhurst:
I bet. Yeah. No, it’s rewarding for what I do too, of making a difference to FP&A helping people be better at their jobs. Can you just tell us a little bit about yourself and how you ended up where you’re at today? So a little bit of that backstory.
Kat Orekhova:
Yeah, I have a, I have a very un maybe unusual backstory. Very few people have followed the exact career path that I have, but, um, lots, lots to learn from there. Ill start with, I always thought I’d be a math professor. That’s, that’s why I have three different degrees in math. And I was, uh, very far along with my PhD at Yale and I, I was teaching multi-variable calculus to undergrads and, uh, you know, just really very far along on that path and thought that what, that was what I wanted to do. But then I, I started feeling like something was missing and it wasn’t any, it wasn’t any burning flag. It wasn’t something very clearly going wrong. It just felt, it felt like the academic research felt a little unfulfilling and I wasn’t sure what else to do because this was the only career path I’d ever really considered all my life at that point.
So I, I did a little bit of exploration. I delved a little bit into the world of hedge funds and high frequency trading, decided, you know, I don’t, I don’t know culturally that that’s where I wanna go. A friend of mine was, um, like me, an ex math PhD. He was at Facebook and he said, Hey Kat, like I’m having such a great time over here. Billions of rows of data. There’s this new thing called data science. And at the time it was new, and this was before all the, the explosion of data science and bootcamps and everything. Facebook only had 10 people, uh, data science, data engineering, data analytics. They were all, all in one. And Facebook had only 10 people at the time doing anything at all relating to their ads, revenue, monetization data. It was so small, we all fit onto one little floor.
And, and so he said, I’m having a ton of fun out here. You should come out and do this. I, you know, I bet they’d hire you <laugh>. And so I, I applied, um, got a job there. I I thought it would be for a year, and then I’d go back to my academic world. But after a year I was managing a team of data scientists. Facebook was growing everything. And I, I just figured this is gonna be, if I go back and finish my PhD, you know, this is just the biggest opportunity cost to like, this is here and now is so much more fun and so much more fulfilling. So I, I stayed, um, after about two and a half years, I was tapped by the now CFO Susan Li to come join her team on finance and build out this data science FP&A team.
And I could talk about that a bit later. So that was a very unusual thing. I, I will say my manager at the time told me I was crazy for going from data science and product into finance. Most people wanna go the other, the, the opposite direction. But I, I thought that, uh, our FP&A team at Facebook was incredibly strong. Um, I, I figured like a lot of them had MBAs, a lot of them had investment banking backgrounds, consulting backgrounds. I, I thought I’d be able to learn a lot. And I, I did, I, I think about that as one of the best professional experiences I’ve had. So I, I saw it as, let me go, let me build this team, let me have a lot of impact work with the executives of this big public company, and then I can always go back to product, which is what I did.
I then, then built out product teams at Facebook, was ultimately a head of product at a growing startup called Ironclad that’s now a $3 billion company, hopefully will IPO soon. And then I started Vareto. So it’s, it’s a very kind of unusual path of building five teams from the ground up at Facebook, building functions from the ground up at Ironclad, and now building a company from the ground up at Vareto. Um, but that, that mix of data science and finance and product, like, that’s always kind of been there and now I just get to do all three of those every single day.
Paul Barnhurst:
Yeah. Yeah. Two kind of fun things with that one, I remember having the conversation with he, he was my CFO for a while, and then he was the general manager and I was his support. So I don’t remember which role he was in at the time. I think he was the general manager of the business, but one day I mentioned I wanna go over to product and he looked at me, he was like, there’s not a chance you’re gonna product, because just they, we had a real challenging fp a department. He did say it exactly those words, but that was basically the idea. And I was like, ah, man, because I thought it’d been fun to kind of, you know, experience the product side of the business. So I can relate to that one. And, you know, startup environment is al always an interesting environment. It’s very fast moving. Sounds like you have a lot of great experience there, you know, doing that. So that’s kind of my, you know, product experience. But I’m curious, going back, you mentioned you always knew you wanted to do math. So when did you know, like, did you always just love numbers or what told you that mathematics and teaching was what you thought you always wanted to do?
Kat Orekhova:
Well, I, I think this is where we sometimes believe we have more, more choice or we’ve made choices ourselves that are actually so impacted by the environment around us. My, my husband’s a lawyer and his dad was a lawyer, and he never considered going into academia. And in my case, my parents were both kind of the, you know, masters PhD, like math sciencey types, like white lab coat, like literal scientist types. And I, I, I mean, no one forced me to go into the career that I did to go into academia, but in retrospect, I think that’s just what I was seeing around me. And that’s just what I thought, well, this, you know, this seems good. This like, seems like a good profession. But, um, I, while I love, I do love math and I, I spent many years doing it, I, what I realized I love even more is it, and part of that you get in math, part of it you don’t, is just problem solving that actually then has an impact on the world.
Because the, the, the thing is when you fall in love with mathematics like calculus, multi-variable calculus, it’s centuries old. Then you get into number theory and topology and algebra in grad school and you know, that’s still like centuries or at least decades old. And by the time you’re doing any kind of real mathematical research, it’s so theoretical and it’s so specific that very rarely can it have real world impact. And I think that’s what I was missing. And as soon as I came to Facebook every week I could go work with some engineers to say, Hey, can we log A, b, c things for, ’cause then I’m gonna start building all kinds of analysis and dashboards and whatever on top of it. And then based on that, I would have recommendations of like, we should build this feature, or we should work with this partner, but not that partner. Or maybe we need to change our pricing or, and I would see this immediate impact not only on the business, but on the people around me. I did a thing and they were happy and they were like, great, like, we’re gonna make more money now, or our team is doing well. And I really wasn’t seeing that in academia. So it was, um, like that was that problem solving, but there has to ultimately be a, an impact that you’re able to see, like, that was the thing that I needed.
Paul Barnhurst:
Makes sense. You wanted to be able to see that direct impact and be able to experience that. And you just, it’s different in academic. I’m sure you could see some with students sometimes and things, but it’s not the same especially on the problem solving side. Right? Mm-Hmm. <affirmative> like you, you mentioned, I’m sure the research that’s going on today is very theoretical because math has been around for thousands of these years. How many things are you gonna really be studying that aren’t theoretical at this point, that are unique? Exactly.
Kat Orekhova:
There,
Paul Barnhurst:
You know, there’s probably not a lot out there. Yeah, that makes a lot of sense. You know, one of the other things I know you do is, I know you’re a general partner for Dark Mode ventures and you invest in startups in addition to having your own startup. So what do you look for in a, in a company and when do you recommend these companies start thinking about finance and FP&A and you know, those type of things for their companies?
Kat Orekhova:
Oh, yeah. I, I love this question. So I, uh, I should mention that I started investing as a way to prepare for being a founder. I, I spent 10 years actually preparing to be a founder, and that was part of why I joined finance, part of why I did all the well, did all the many things I did in my career to get more exposure, but I also started angel investing and working with startups even while I was still at meta at a big company as a way to learn more about them and to learn about fundraising and, and how they operate and how they hire and, and how you really get to product market fit. And I, I just like, I saw this as a big learning opportunity. And so now when I, when I look at founders after years of having done angel investing, having then been a Sequoia Scout having one fund, now having a second fund, obviously what I look for has matured quite a bit.
Um, I, I think the thing that’s always been the same throughout is a big emphasis on the people. And everyone will say this for early stage investing, you know, the, the founders matter a lot. Uh, idea is important, but who the people are matters. But I guess the way that we look at founder and founder, founder competencies and what, what does someone actually need, not just in terms of can they do it, but motivation? What is gonna make them want to do it for years? Like, get up every day and like work hard and because it can be really hard. Um, so these are, these are kind of things that we think a lot about now, almost like an interview process for founders in terms of finance and FP&A for companies. Um, this is maybe a bit of a controversial view. I I know there’s a lot of, uh, there’s a lot of people out there kind of saying, how do you, you build an all-in-one so that even startups at their earliest days just have like a system that can do finance for them and, uh, maybe help them plan at earlier stages of the company.
I, I actually think there’s something to be said for having people do something manually or in a very hands-on way, especially when things are changing quickly to really understand what are, like coming back to something you said earlier, Paul, like what are the drivers of the business? What really matters and what doesn’t? And so for any startup that’s less than, let’s say 50 people or less than series B doesn’t, doesn’t really have a, a not just a, a growth stage kind of scalable motion, but really a, a very repeatable, um, motion where it makes sense to have bottoms up planning and things like that. I personally say like, use spreadsheets and, uh, and, but focus more on what matters for the business and like nail that don’t get too bogged down and the BVAs, uh, where you’re, you’re on operating on such a little small revenue that it doesn’t really matter. Focus on the bigger picture of like, what’s gonna drive faster revenue growth.
Paul Barnhurst:
[Datarails ad] You know what it is, like 13 different spreadsheets emailed out to 23 different budget holders, multiple iterations, version control errors, back and forth updates. You never really feel in control of the consolidation and collection process. Yep. I’ve been there. Stop. Breathe
Data rails is the financial planning and analysis platform for Excel users. Data rails takes data from all your company’s disparate sources. No organization is too complex, consolidating everything into one place, secured in the cloud. Now all your data finally talking to each other, everything is automated back into your report in Excel. Cashflow, FX conversion, intercompany transactions now automated and up to date, drill down and variance analysis in seconds. Don’t replace Excel, embrace Excel, turn your Excel into a lean mean fp and a machine, find out more@www.data rails.com.
There’s a lot of truth to that. And yeah, I have a lot of people talk to me and it’s like, hey, well when do I get a tool? And I’m like, it when the complexity warrants it, you’ll know is usually what I’ll say, right. We’ve all experienced that, and that can be at different points for everybody, but there is some value, like I said, focusing on the strategic and the big picture versus thinking, I need a tool from day one. Are there some places where it makes sense? Sure. Are there some great tools out there that you can use in, you know, in different places? Yeah. But you know, like just even our sponsor focuses on the, you know, on the small and medium market, and I encourage people when you’re ready to go, but you know, it’s hard to meet a spreadsheet, especially, you know, those very early days from a cost standpoint and other things. So you just have to think about when does it make sense?
Kat Orekhova:
Yeah, absolutely.
Paul Barnhurst:
I I I appreciate that. Data
Kat Orekhova:
Datarails is actually a great example of people are still on spreadsheets, but they’ve got the benefit, like some incremental kind of benefits on top of that with maybe pulling data in and out. And I, I think that could be really great for small, medium businesses. So I don’t wanna imply that there’s, there’s no value at all in any kind of software if you’re a small company. But more just, just like you still have to ultimately own, I think the strategic planning yourself, you can’t outsource that to a tool early on.
Paul Barnhurst:
And that makes a lot of sense. There’s the strategic part of the process, and I agree with you, Datarails has a great, you know, great platform and there’s, there’s definitely opportunities out there, but like you said, it’s the strategic part and the learning that takes place in those early days, so that makes a lot of sense. So I’m curious, you mentioned one other thing, I think you said you’re on your second fund, is that what you said for your Dark Mode Ventures.
Kat Orekhova:
Yeah, correct. We recently announced Fund two.
Paul Barnhurst:
How exciting. So what’s your favorite part of investing in startups? What’s kept you doing that even as a CEO? Like what do you love about that?
Kat Orekhova:
Yeah, that’s, uh, it’s interesting because I definitely thought that when I became a startup founder, I would stop investing. The whole point of investing was to learn and to prepare. And now that I was a founder, I should just focus on that, right? So I, I intended to stop, but by, by that point, I had very good deal flow. I, I had worked with enough startups that they recommended me to other newer startups. And, and so even without me doing any outbound work, inbound, great deals were kind of coming to me and people just saying, Hey, I’m starting a startup. Do you wanna invest? Do you wanna help me? And, um, and so I thought, okay, well let me, let me just do that only the inbound side of it, like let me really limit my time. And then kind of around the same time, Sequoia and, and others funds that I’d been sending some companies to and helping facilitate a, a lead for, for a seed round, let’s say.
Were just saying, you know, you’re sending me great companies. Do you wanna work more formally with us? Do you wanna be a scout and invest our money? And then, you know, like kind of have a more formal relationship. So I decided to do that, and I was a Sequoia Scout for a couple of years, and again, really tried to, they, they do a very good job trying to offer you optionality, but without the commitment of having to make investments. So that was a good way for me to feel things out. Ultimately though, I, I ended up doing the fund and sticking with it because I realized it did make me a better founder. And while there is a bit of a trade off of time, it’s the same reason why so many let’s, like finance leaders and others take board roles in other companies, it makes you better as a CFO or as a VP finance, it makes you better at your job if you have to look at other companies from a 360 view where you don’t have the same type of day-to-day, you know, nitty gritty, and you’re forced to understand how product and go to market and how all the things, all the dots connect together.
I’ve just seen on the same way the investing perspective gives me that and actually has enabled me to do a better job with Vareto. And, and frankly, I just enjoy it. So I love working with founders. It’s, it’s been a real privilege.
Paul Barnhurst:
It’s great. So I appreciate you sharing that, and it does sound like it’s obvious that you love it. I can tell the passion there. So I’m gonna switch gears a little bit and ask a question more. Getting back to FP&A, so you’ve worked in product, a good part of your career, you’ve been head of product at a company. How do you see product and FP&A working together and any advice you can offer on how FP&A can better support product from your experiences?
Kat Orekhova:
Yes. I, I especially saw this at Ironclad where I, I had the, the PM function and also the beginnings of our analytics function for the entire company. And, um, we, we also had our first ever finance lead join around the same time. Um, that person didn’t report to me, so, you know, I think reported to the CEO directly maybe at that point. But it was just so early where you, you only had a few PMs and analysts and a single finance person, and that was when we put together the first ever infrastructure for, for our company of like, how did the not, how did the dots connect? And for example, we realized with some very basic analysis that sales was selling all use cases equally. And meanwhile customer success was spending 10 x the time on serving the procurement use case as serving a sales use case.
And that doesn’t really make sense. Maybe you need to then charge more for the procurement use case or figure something else out, right? But it, it was just this like misconnection or lost opportunity, let’s call it, between what sales was doing and what customer success was doing.
And we only found this out when we actually did the analysis. And the reason I’m talking about this as analysis is, ’cause I think often the finance team is, on one hand, they’re sort of seen as like, well, you’re part of G&A, you’s like finance and HR and legal and ops, and you all sit over there. And then there’s all these other teams that actually build the product and nd like sell the product and, you know, make, like, make the business go. But actually, especially when I was in FP&A myself, what I really saw was finance thinks about the entire business.
They, they have to understand so well what is going on with product and with go to market and how all of the, all of those different segmentations are looking and connecting the dots as a whole. And so in that sense, like the analytics team that we formed at at ironclad was, it wasn’t a finance team per se, but we were doing much of what a finance function might otherwise do, just in that sense of like, let’s look at our business at different levels. Let’s understand where there are gaps or opportunities. Let’s surface those recommendations up to the executive team and let’s, you know, let’s, let’s take action on it. So I just really saw that that kind of, um, interplay between finance, product, go to market, really just connecting those dots was, uh, was absolutely critical there.
Paul Barnhurst:
Thank you, and I appreciate that. And there’s something you said there that I’ve always, uh, really liked is, you know, finance in particular, fp and a is one of the only functions outside of the CEO, the office of the CEO, the CFO that has a 360 view of the business, right? Especially in a small business where you’re working with everybody. And so there’s opportunities to bring things up that others don’t notice because sales is focused on the sales engine and marketing is focused on the marketing engine and product on the product, whereas finance can bring that analysis and those opportunities across all of it. And I think we’re seeing more and more of that today, especially since covid where people are realizing that, you know, FP&A and finance can play a much more strategic role than historically. Like you think back office, right? They’re just, you, you go over there, you don’t make money, you just kind of do your job. You’re a cost center. And I think that we’re definitely seeing that change. I, I like how you said that. So I, I’m curious, you spent a few years at Facebook. You mentioned a little bit about that story, but can you go into details why did they create a, you know, a finance team that had data science skills? What led to that? What was kind of the goal of your team and how did that come about?
Kat Orekhova:
Um, I mean, huge credit there too. Susan Li, who’s now the CFO and is just absolutely outstanding. I’ve, I’ve had many great managers over the years. I’ve been very, very fortunate to learn from so many incredible advisors. But Susan is just, she is special. She’s in a category all on her own, just someone that I, I really look up to and aspire to. And she had this insight, like she made this happen. She saw that on FP&A at Meta, and this is like, you know, like meta was already just gone. Public, let’s say, was growing a successful business, making a lot of money. But, um, despite all the wonderful people they were hiring, very smart, very hardworking people. Meta could get the best of the best, but they were still FP&A type of people. They weren’t data scientists and, and they didn’t have any relationships with like the product teams that we were mentioning earlier.
So they didn’t really know what the state of the users and like what product launches might cause users to engage more and spend more time, which meant we could show them more ads. They just didn’t have any of those internal relationships. And, and so Susan saw that the FP&A team would do their best using the data that they had to create revenue forecasts and to do various things. And, but then ultimately something would happen. And kind of like, again, coming back to what you said, Paul, the, the drivers weren’t there. And so, um, there, the actuals versus the forecast would be quite a bit off and the entire executive team would be worried. And for a while this was happening on a weekly basis. There was like a revenue war room, and it wasn’t clear whether is it that users are switching to mobile?
Is it that, um, advertisers, you know, the brands using that Facebook or maybe doing something different? Is it that we’re just changing our own algorithms for ads or for products that are like, what are the drivers of this change? Should we be worried, should we be taking an action? Those were questions that were being asked on a daily basis, and war rooms were happening every week, and it was just really hard. So I think Susan saw the need to try to bridge that gap and to get someone that did have more of that product and data science, not only technical ability, but I had the relationships because that’s for two and a half years I’d been working in the ads, data science and product world. And so Susan kind of figured like, okay, if I bring cat over here into finance, and we build out a team here that’s focused on almost bridging the gap, like that’s at, maybe it’s not gonna work, but at least it’s something new <laugh> that we should try because, uh, the thing that we’ve been trying so far, sure, um, it’s, it’s just, it, it hasn’t been working.
So, uh, so she, she did that. She kind of came over, um, asked me about this idea again, like I mentioned, my manager thought it was a crazy idea at first, but I was really intrigued. Um, we really talked about the role, like we talked about it being a two year thing where I would come in, I’d build a team, I’d do a, a few, you know, um, we ended up automating a lot of, uh, big processes. We ended up taking forecasting from monthly to daily, making it accurate, like having it be more drivers based. Um, and then I, I had hired my replacement, who’s actually still the leader of that team today. So, uh, that was, that was kind of how that transition came to be. And fortunately, I think Susan just, she was so, she was, she really saw the future, right? It, it worked out very well for us and we had a number of years of very good forecasting.
Paul Barnhurst:
That’s great. And whenever you see a inspirational leader that can look beyond and kind of see where things are going, can make a huge difference. I mean, there’s so much value in that, having that ability to kind of be forward thinking. But I’m curious, you know, obviously you guys approached forecasting from a data science prote perspective different than FP&A. So was there kind of challenges as you tried to bring in more data science or how did that go about of kind of blending the processes and know improving the forecasting? Because I’m sure there are very different methods that you were using compared to what most those traditional FP&A people were used to.
Kat Orekhova:
Correct. Um, that, and that’s actually one of the trickiest parts is when you have folks in data science and then you have folks in FP&A they’re, they’re living in different tools. Like as a data scientist, my tools were Python and Rand terminals and FP&A lives in spreadsheets. And so even if you could do a certain thing better, if now the rest of the fp and a team can’t access the thing that you did because it’s, it’s at a Python script, it’s helpful, but there’s a limit to how helpful it could be. So establishing a good bridge where the whole workflow can, can operate together, but without forcing everyone in FP&A to learn Python and, you know, uh, kind of vice versa, that that was a big part of what we had to do. So I, I, I didn’t just, um, you know, hire people to write python scripts.
While we did do quite a bit of that and we built some very sophisticated machine learning models and everything, um, I also, you know, owned this giant 50 tab kind of revenue forecasting model that took as outputs some of the things our Python scripts did, as well as, uh, various other inputs we get from the accounting team and others. And ultimately, like that was something that the rest of FP&A could interact with because it was more in a format that they could understand. Um, and that, that was honestly too, one of the gaps that I saw where, like I was, I was very fortunate and Facebook is a special company that’s printing cash and can afford to hire a data science team within finance. ’cause let’s face it, most companies, that’s not an option. So Facebook was lucky to do that, but even there, everything we built internally in a lot of our internal processes, it, it was never kind of completely to the level of what I thought could be done with a purpose-built tool. And in particular, that collaboration layer, like that’s still largely remained spreadsheets and emails and, and PowerPoint decks. And that’s kind of the part that I always thought, you know, it’s great that we did all this automation and forecasting stuff, but if I could also build a collaboration layer to solve that additional set of problems, like the inputs gathering and, and the approvals, like that would really, that would be ideal. And, and that’s kind of one of the main things I’ve been thinking about ever since.
Paul Barnhurst:
I’m curious, stepping back, you mentioned you did a lot of algorithms. I know Facebook has released an open source profit algorithm that’s used for some of forecasting. Is that one of the things you were working on there? Is that part of Yeah. Kind of the forecasting you were doing? I’m just curious,
Kat Orekhova:
Uh, yeah, actually that, that that team and, and that model, um, they were working very closely with my team as well. So the, the way Facebook is structured was there was the finance team, but we also had a kind of core data science team for the entire company. And several people on that team worked really closely with us on some of the more technical aspects and models. And yes, they ended up later taking some of that as inspiration and building kind of external open source versions. I think one of them might have even built a startup around some of this. So it, it’s a lot of really, really exciting stuff was done there. And it’s, it’s just important to separate out though the, there’s like the pure forecasting of imagine I have all of the data, how do I do the technical aspects? We did a lot with seasonality, for example, and other overlays that was very interesting, made our forecast a lot more accurate.
But then there is sort of the, well, but I, I just need even macro inputs. Um, like there’s, there’s data about the future that nothing about my historical data could possibly predict that a lot of that data lives in people’s heads and it’s maybe it’s even strategic decisions about where you wanna take the company. And so that, how does that data ultimately get those drivers ultimately get into your model and kind of mix in with the technology with like the more advanced forecasting techniques. Like that was one of the very interesting things that our team tackled.
Paul Barnhurst:
Got it. And kind of speaking of that, that leads to a question I think is along those lines. How do you know FP&A and data science professionals work better together? Any advice you’d offer from your learning experience? You know, if companies are thinking, Hey, I wanna have, you know, a data science person in fp a, we’re definitely seeing more of that. How do they make sure they’re kind of, you know, maximizing the benefit of their skillset and working together to improve that forecasting process? Any advice you’d offer?
Kat Orekhova:
Yeah, absolutely. I mean, in a way that’s sort of what we did where my team was kind of a mix of, I’d say more technical and more traditional FP&A, but then we worked with this core data science team that was very technical, like passing in, you know, a set of engineering interviews and everything. And we would also work with a, a BI team of more like data engineers, kind of helping build certain data pipelines for us and working with these other kind of partner teams. Um, one of the things we found was it’s so critical to lay out the right context because it, it’s very easy to just assume, especially when some of these folks are more or less full-time assigned to work on the FP&A set of problems that they know what fp and a is all about and they understand what the executive team cares about or, but they don’t <laugh>.
And so, so sometimes like things like, um, whenever we’d introduce a new ad type, we’d have to make sure that, that the logging for it and ultimately it would get rolled up in a certain correct way into our actuals, uh, without context as to what is or isn’t important for, for the type of reporting that we do at a business level. It wouldn’t be obvious to our BI team necessarily to, to do that and to know what to look out for or like at what, which product features or launches matter and we need to do some work and which ones don’t. So the more business context we could give the data science and, and BI teams, the more they could use their judgment. ’cause otherwise where you, where you end up is in a situation where the finance team has to know about every single technical thing that could happen. And while maybe they’re not writing the code, they have to really understand a lot of it, uh, enough to give very explicit directions, which I think kind of defeats the whole purpose. So the more business context you can give, the better
Paul Barnhurst:
Makes a lot of sense. I would totally agree. The better you have that business context so you can kind of work together. So I, I appreciate that. I think that’s really good advice. One other question I have is, you know, I occasionally get people approaching me and for a while they got a lot of people think I need to go into data science, right? Like, I need a data science degree, even if I’m working in FP&A, everything’s going toward the data science. What’s
Kat Orekhova:
Your thoughts on that? Uh, you know, I think it’s, for anyone that’s really ambitious, and especially if you want a career honestly in almost any anything, but that’s gonna be in a more strategic role, you don’t need to know detailed level coding, but I think you need to at least understand what coding can do or what data science generally is or what it can do enough. So to potentially be able to hire people to do that or to know what kind of context to give them. So it, it’s, what I would say is, um, you know, very tactically I would recommend taking one or two intro courses to data science. You know, just, just like get comfortable with the basics of manipulating data, doing some basically even like linear regressions, some, some kind of, uh, basic forecasting. It doesn’t have to be anything crazy. I don’t think you need to go in and learn all, all the depths of, uh, machine learning or generative ai, but enough just that you understand the concepts and again, what what matters and what doesn’t.
And then where what I kind of see things going, especially with just more and more things being automated and more, um, kind of more applications of AI being used is even in software engineering, we’re sort of seeing this, that software engineers right now might need to write all kinds of code to produce, let’s say a dark mode button. Like if you want your website to have dark mode, light mode, it’s all kinds of like code that an engineer would’ve written historically. But nowadays you can actually ask chat GPT to write this. It may not be performance optimized. I’m not suggesting we all go out and do this, but the point is, often if you can say, nowadays I want to write a script that does x, it’s becoming increasingly easier to do that. Same thing in data science. Same thing in software engineering. Any, uh, anything you can imagine. So what matters is almost the, uh, the architecture of knowing I, I need a script that does this, I need a script that does that, and then how do I put those two scripts together? Because in the end, you can’t quite ask had ChatGPT to write you like the whole web application. It’s not there yet. Um, it’s more on a micro level.
Paul Barnhurst:
I appreciate that. You know, funny enough, so I, when I did grad school, you know, data science really wasn’t a thing yet, you know, I mean yeah, at early stages of it. And so I did an MBA and a master of science and information management and you know, I remember looking back, it was five, maybe five, 10 years after I had graduated seeing they had an analytics program, you know, data analytics. And I’m like, man, I wish I could have done that because that would’ve been more what I wanted, but it just wasn’t there yet in the universities. It really cropped up about a couple years after I left, I really started to become bigger. And so I can appreciate wanting to be able to at least understand it. You don’t gotta go deep, like you said, you don’t need to be a coder and spend all your time in Python, but having an appreciation can really help.
Kat Orekhova:
Exactly.
Paul Barnhurst:
Is kind of what I was hearing there and I would agree with that. And you mentioned a little bit about ai and so how do you see AI changing the role of FP&A moving forward? And this will be a two part question. What should FP&A professionals do today to be prepared to capitalize on AI? Because it’s not going away away, right? We all know it’s here to stay. It’s just a question of what does it look like in the future? So I know you presented at AFP on this subject and we just love, you know, your thoughts on that.
Kat Orekhova:
Yeah, a lot of people are asking that, and not just in finance, but across every function in in marketing and sales, what is gonna be automated? Yeah. Whether that’s AI specific or just kind of more general automation, and what are people still gonna be doing in five years, 10 years? And to me it really comes down to, again, that kind of like the, the business strategy, the business logic, figuring out what to do is always gonna be the job of humans. They’re the ones that make decisions. And a lot of the decisions we make aren’t pure optimization. Like you look at companies and there are many successful types of businesses out there. There are some that say we’re limiting ourselves to 50 employees. We’ll never grow above that, but we wanna see how much revenue we can get, just 50 employees. There are others that make a different decision to grow as quickly as possible, even though, you know, revenue per headcount is not gonna be as good.
There’s just so many, like, it’s not black or white. Humans make choices, humans drive the visions for their businesses. And, um, and I think in a, in a very similar way with FP&A, humans are the ones that work with business stakeholders to, to make these types of decisions and to say, well, here are our sales targets marketing, do we feel comfortable or not with you doing this big event this year given this budget?
But where I do think AI and automation really steps in is in a lot more of those more granular pieces, kind of the, all the like little scenarios, all the little, well, if marketing does or doesn’t do this conference, what impact does that have on expenses? What impact will that have on sales given reduced downstream leads if we don’t do the conference? Um, just really like all those, um, all those things where fp and a would previously have said, let me get back to you, and then three days after a meeting would’ve come back with a spreadsheet and an analysis. All of those things will become in the moment, like, let’s just run a scenario, let’s just get the answers we need and let’s make decisions faster. But at the end, human beings are always gonna be the ones to make decisions. I’m, uh, I don’t believe companies are gonna be AI driven, like cars are gonna be, uh, self-driven. It, it doesn’t work that way.
Paul Barnhurst:
Yeah. And that, that makes sense to me. I, I could see that, and I tend to agree with you. I think, you know, the humans will still make the decisions, but will be much more productive, have the data much more readily available, and be able to be better equipped to make those decisions better to analyze the data. I think that’s, you know, where chat GPT can be great in some of those areas is helping you get something ready. You still have to decide what to do with it, but it can make a big difference in, in that preparation process.
Kat Orekhova:
Just just wanted to add that I, someone else had said it so well at the AFP conference, this idea that manyFP&A folks today who are, who can be very strategic, are spending 80% of their time in a, you know, kind of like in a, a back room by themselves running various analyses because there’s just so many scenarios that, that we need to be running at all times. But as all of that gets automated, the role of those folks and in general, the role of most FP&A professionals will be more to have conversations to be, in a way in the spotlight to have discussions to talk about the trade-offs and the decisions. And so I see this kind of shift from, not to say you don’t need your Excel skills anymore. I think you will always need some level of technical data skills, but you will spend less time practicing those skills and more time practicing a lot of the soft skills, communication, negotiation kind of alignment with stakeholders. And so I’m seeing a lot of folks actually focus more on developing those skills as well, not just the technical.
Paul Barnhurst:
I agree with you. I’ve, I said, you know, a lot of that data prep, those type of things will get to the point where a lot of that will be automated. And as you mentioned, the soft skills, the human skills, the people skills, those things will become more important. So we’re coming up on the end of our time. So we have this section called the get to know you section. How this works is we have four questions we’re gonna ask you get no more than 30 seconds to answer each question. So that is, you know, kind of relatively quick and it’s just to have a little bit of fun and get to know you on a little more personal level. So the first one is, what is something interesting about you that maybe we wouldn’t find online or that not many people know that you can share with us?
Kat Orekhova:
Well, I grew up in New York City, uh, with skyscrapers. Did not see a mountain until I was 18 years old when I did a road trip across the United States. And, and then it was love at first sight. So that’s why I, I moved to Boulder, Colorado later with no, no family and no official relationship there whatsoever. But I, I live in Boulder. I love my mountains. It feels very, very natural for me being here in the, in the center of the US.
Paul Barnhurst:
I love mountains as well. I grew up in Salt Lake, but I’ve always been a big outdoors person. A lot of mountains camping, hiking, so I can totally appreciate that one. So next question, if you could meet one person in the world, dead or alive, who would you meet and why?
Kat Orekhova:
There are so many people. Um, uh, I, I think if I had to pick one, Mozart would be fascinating. Just the kind of music that he wrote, especially at such an early age is astounding. I, I don’t, I don’t know how many people even today or since then have been able to replicate that. So just understanding how such a young boy is able to do something so incredible that has touched humanity for centuries afterward. That would be fa fascinating.
Paul Barnhurst:
I, I could see that one. It would be fascinating, especially as a kid to watch him work and just see that his thought process, ’cause yeah, it stood the test of time and he did it at such a young age. It really is amazing. All right, so this next one we’re gonna ask you is, what is the last thing? And I’m gonna ask specifically, what’s the last thing you used generative AI that you asked it about something related to finance, FP&A Excel. What’s the last thing you used with generative AI for in that realm?
Kat Orekhova:
So I, I do, I do a lot of, uh, research and testing or I try to using chat GPT, some things it’s very, very good for, and other things not so much. I was, I was trying to, I was trying to get it to help me brainstorm topics around actually finance and ai. And it was not able to give me a whole lot. If, if you go and you search for machine learning and finance or AI and finance and you, you try to figure it out, not a whole lot, which I think just goes to the internet in general, not having a lot on that topic, but the level of interest compared to the level of knowledge or information available, it’s very big discrepancy. So we’re working to fill that.
Paul Barnhurst:
Interesting. Yeah, I, that had been an interesting search and I would agree with you. We definitely don’t have as much information as I think could help all of us. So the last one here, obviously, you know, Datarails sponsors the show Great Excel tool for small and medium sized businesses. And so one of the questions we ask everybody is, what is your favorite Excel function or feature? Favorite thing about Excel?
Kat Orekhova:
So I will have to say the keyboard shortcuts. The fact that there is like a World Excel championships and there’s so many lovers of Excel, it, it’s Excel is not just a tool. It it can become something that you can be good at. Something that almost like a gamer or like a chess master, you can master Excel. And the fact that you have all these keyboard shortcuts that not only make you insanely efficient, not having to click through like a whole, you know, 10 different clicks into some modal somewhere, but you can, um, it’s almost like a secret language that some people know and some people are very, very, very good at it. It, it makes it, it makes Excel almost a bit like a superheroes tool if you know how to use it. So I absolutely love that. And I, I, I think like to all FP&A folks out there, if you’re not using Excel, um, demand that the tool you do use cares about keyboard shortcuts, right? Because it’s, it’s such a superpower.
Paul Barnhurst:
I have to laugh ’cause that’s what I’ve been doing all day to today. ’cause I’m prepping for, uh, the FMI, which is a modeling test and it’s all about speed with that one. ‘Because you get four hours to build an entire three statement model. So yeah, you gotta, you have to use shortcuts or you’re not gonna pass. If you’re doing the mouse, you’re gonna be there forever. So yeah, there definitely is a whole world of shortcuts and it’s amazing to watch some of these people how quick they can do something. It is, it is pretty amazing. So I have just two questions left here for you. So first one is, what advice would you offer to someone starting a career in fp and a today? Any advice you’d give them?
Kat Orekhova:
Probably similar advice to anyone in general, but especially FP&A given, like we said, even an analyst tends to have more of a 360 view of the business. It’s a very rare position to be in. Um, most important would be focusing on learning. I think historically FP&A has often hired people for processes and even an analyst will typically get hired in into one business line and you tend to do a lot of that back office spreadsheet stuff and you don’t, you’re not running the budgeting process, usually the manager or the director is. But the more that you can, I guess try to get your processes done as, as quickly as possible and try to learn more about those interactions between finance and the business. Try to take advantage of being in finance to get the context of the business as a whole.
If you work at a public company, listen to the earnings calls. We used to do that as an FP&A team altogether. It was really, it was really special. Just like take advantage of the fact that you are in a very rare position to have access to so much information and even as an analyst to the leaders of the company in a way that most people, like just a data analyst, a junior PM a junior engineer wouldn’t have that level of access. Um, so just really, really trying to learn as much as you can early on. And like we mentioned, look, to develop not only your core finance skills, but also some technical knowledge that not beyond a certain amount is needed, but at least some is good. And, and then soft skills as well.
Paul Barnhurst:
Got it. Yeah, great advice there. And I agree with about learning the business and being curious. So much value in that, especially in an FP&A role. So last question here. If someone wants to get ahold of you, what’s the best way for them to do that?
Kat Orekhova:
Um, email or LinkedIn. I’m just kat KA t@reto.com. So super simple. And I’m on LinkedIn as well, so find me there, connect with me. I’d love to hear from you. Um, just, I love chatting about this stuff. So all questions, comments are open.
Paul Barnhurst:
Great. I know we’ve had a number of conversations on a lot of this stuff, so thanks for joining. I really enjoyed learning more about your, uh, story and experience working in data science and fp and a and bridging that gap and then, you know, starting your own company. And so thank you so much for joining us today. I really enjoyed the conversation, Kat.
Kat Orekhova:
Likewise. Thanks so much for having me Paul.
