Graham Stanton on the CFO Tech Stack

Graham Stanton Introduction

[00:00:00] Joe Garafalo: Hello and welcome to another episode of The Role Forward podcast. I’m Joe Garafalo, and this episode is brought to you by Mosaic, a strategic finance platform that transforms the way business gets done. Today, I’m here with Graham Stanton who co-founded Peloton in 2012.

[00:00:17] During his time there, he wore multiple hats, overseeing everything from technology to legal, to the people team, performance marketing, and most notably for our podcast, he oversaw finance. Graham, Graham scaled Peloton from the startup back in 2012 to the household name that it is today, and he’s taken all that experience and co-founded a company called Avise, which is aiming to unlock the power of accounting to help drive business forward. Graham, thanks so much for joining me today. We’re super excited to have you on The Roll Forward podcast. Can you give the audience a quick introduction and a little bit of background on yourself? 

[00:00:53] Graham Stanton: Yeah. Well, thank you, Joe. I’m really happy to be here and yeah, sure. Yeah, and I, that, that was pretty good intro right [00:01:00] there. Yeah, I’ll throw in that, yeah, I, I started my career as a software engineer before moving to the business, to the business side a few years out of college. But yeah, I always kind of took that software development, analytical, you know, view of things, and, you know, maybe that gave me a lower tolerance for using bad, bad software for anybody.

[00:01:18] Joe Garafalo: I love that. And, that kind of brings us to our topic for today’s conversation, which is really focused around the CFO tech stack and back office technology. Graham, in a Forbes article that was, that was highlighting Avise, the new company that you’re building, you mentioned something about having a career-long frustration with, with back office technology and with finance tooling.

[00:01:38] Can you explain a little bit about that frustration and how it led you to start?

Frustration with ERPs and Other Back-Office Software

[00:01:41] Graham Stanton: Yeah, absolutely. I, and I, I, I can say in general, yeah, I always had it, the sort of, like, bad view of back office software, that my, my first paid programming job was an IT department that, at some corporation where, yeah, the people developing the software kept blaming the [00:02:00] users for bad data in the system, and now I want to pull my hair out saying, “Oh, it’s your bad software, and that’s leading to that.” But, you know, that, that was late nineties and unfortunately nothing’s improved since then for the most part. You’re starting to see some green shoots now. But yeah, when I, when I was at Peloton, you know, I was kind of right in the middle of it and everything, everything we used was clunky, I mean, we, we used QuickBooks desktop, which, you know, did its job, it w, it, it was fine, but nothing more than that. Anytime we tried to go beyond what it offered out of the box, you know, we hit some real, real big issues, you know, and then at one point we had to upgrade to an ERP and that’s when I learned that sort of real enterprise technology is worse than anything I’d seen up to that point.

[00:02:52] Joe Garafalo: For sure. Talk a little bit more about that ERP transition. I’ve, I’ve been through some, some really long implementations, I think. [00:03:00] When they go well, it takes at least six months, and when they go not so well it could take up to a year. How was your transition? 

[00:03:06] Graham Stanton: Ah, I think it was two years or, or more, depending on what you have, what you count as done. It was, I mean, it was a big disillusionment. I mean, I didn’t, I’d never used it in the ERP before. We’ve gotten to a point at Peloton when we had a fairly complex business and it was, it was tough, it was tough to keep track of everything, you know, both from the standpoint of getting the books done, to actually understanding the business. And I, you know, I foolishly thought that moving to an ERP would solve both of those problems, and I mean, just right there in the name, Enterprise Resource Planning, I mean, that’s what these systems are designed to do, and yeah, from my perspective, moving to an ERP really just added a huge headache and it was a necessary evil. There, there was no alternative. Any alternative would have been worse, but, you know, in terms of closing our books, in terms of getting our accounting done and in terms of understanding our business, [00:04:00] all that progress was done based on custom internal work and yeah, the off-the-shelf software, you know, the very expensive off-the-shelf software that, you know, required, you know, really, hoards of people to operate, a bunch of in-house hires, implementers, consultants, to consultants and just on and on, it didn’t add to that process. It was a useful repository of the end result of some of the data we generated, but even then it was only a summary so we could produce our financial statements from it, but we had to do a lot of custom work to get the data and, and then any analysis, like the granularity, wasn’t in the system so we needed to use other platforms anyway.

[00:04:38] Joe Garafalo: Graham, I’ve had a lot of the same frustrations around the ERP system, being more of this backwards-looking newspaper from last, not really driving any of the analysis, especially not in real time.

[00:05:39] What were some of the other tools that you used at Peloton that helped plug the gaps that the ERP couldn’t fill? 

[00:05:45] Graham Stanton: Yeah. I mean, I, I really liked that view of what an ERP is, and it’s, it’s, it’s so slow and, and it’s, it’s really just the headlines, and that’s assuming you’re using it right in the first place. Yeah, you can’t use it to drive your business. So, at Peloton we [00:06:00] had, I’ve mentioned we had a number of internal systems, you know, external systems.

[00:06:03] So, we had, we built our own e-commerce system, you know, we integrated with Stripe to take payments, you know. We had our own, we had our own system obviously for, for tracking workouts, yeah, for doing the subscription billing and tracking, you know, various external systems for deliveries, et cetera.

[00:06:18] And, yeah, really our only hope to close the gaps was to pull all that data into a Data Warehouse, you know, w, which was very separate from the ERP, and that required a dedicated team and just to get the data in there, and then we put, we put Looker on top, and yeah, we had countless, we had countless Looker reports and, you know, I think, at one point, I counted 800 Looker views to go into that, yeah, hundreds of back end reports on top of that and maybe thousands, yeah. And then, yeah, that was only kind of, sort of usable for the business team, so we also had data analysts whose job it was to [00:07:00] pull reports on request or, you know, kind of configure it exactly as, you know, as people needed and, and that got us our real-time information. It meant that the, the turnaround on new analysis was, you know, depending on the proficiency of the user, maybe a little slower than ideal, and yeah, one of the, one of the weaknesses was it didn’t necessarily tie to the core financials just because of the way the data flowed over. And, it was all in theory accurate but that led to another big headache and when we were getting ready to go public and needed to guarantee that absolutely everything tied down to the penny.

The Modern CFO Tech Stack

[00:07:35] Joe Garafalo: Totally. Especially with all the accounting to gap and cash affects that, that happen through multiple sets of data, it’s really hard to, to tie that granular transaction data to your headline financial numbers and have that beautiful bridge that connects it When you think about, like, the, the pieces of the modern CFO tech stack, what are some of the characteristics or features that you think will make the [00:08:00] back office less painful in the future? 

[00:08:05] Graham Stanton: it’s a good question. I mean for sure that, you know, that realtime view is going to be critical and any company that, you know, actually changes from month to month, you know, isn’t going to be satisfied looking at last months of data or data that’s two months old, if the close actually takes an additional month, and I think, you you, know, the, the other part really is usability, and I think that we’re getting to a point where, yeah, people aren’t accepting that, “Oh, this tool gets the job done, but it’s a huge pain to do it,” and especially startups or modern businesses that have a lot to get done, and there’s not, yeah, there’s not going to be so much tolerance for, yeah, you can technically do this. I mean, it feels like it should take 20 minutes, but it’s going to take you all day or all week and you don’t have it, you’ve [00:09:00] got things to do to actually drive the business.

[00:09:02] Joe Garafalo: Yep. I love that, and even when you mentioned a tool like Looker, which we absolutely love, there is a pretty steep learning curve to get a tool like Looker up and running to have the, the LookML to write those queries that produce all the metrics data. So, I think for the future of the modern CFO tech stack, that skillset to wrangle and manipulate that data needs to get a bit easier, and it also needs to be in the hands of people like us leading the finance team versus the engineering folks who, who, who do know how to manipulate that data at a mass scale. Another topic that I really like that I saw you talk about in the Forbes article is this concept of consumer grade enterprise technology.

[00:09:43] I think one of the things that Peloton has really nailed was making a fitness industry, which was, you know, kind of old, older, and, less, less futuristic feel really friendly and feel really new and fresh. That same approach, I think, that you took to Peloton needs to be taken to financial [00:10:00] technology and specifically that of the CFO tech stack.

[00:10:03] Can you talk a little bit about consumer-friendly, design-based product for finance? 

[00:10:07] Graham Stanton: Yeah, absolutely. And, yeah, it, it’s definitely a parallel. Yeah, I think it’d be, in the fitness space, yeah, there, you know, to the extent that there were any electronics integrated into equipment, it was blinking lights in a circle, and in 2012, I went to an equipment conference and one of the manufacturers very eagerly showed me their new iPad app that was literally an iPad app of blinking lights in a circle, and it’s the attitude of like, well, you’re doing a workout because you know you’re supposed to, or you have to, so you’ll put up with anything and, and it’s the same attitude for back office software. You know, it’s like, “Well, you know, these people are paid to do it.”

[00:10:47] You know, w, we actually heard recently from someone who met with a sales rep, a sales representative from one of the top, you know, seven-figure per year software where the salesperson outright [00:11:00] told them, “Well, we’re not gonna win any usability awards,” and, but, you know, they, they get the job done technically, supposedly, and, and I, and you know, as I was talking about a little bit, you know, with the previous question that, you know, modern company, people don’t have the time for that, and people are losing the patients for that, may have, and we already have this world of consumer software that, yeah, that, that’s always been the case where consumers, they have, will only, will only use software that is intuitive, that they enjoy using and if it’s not, they’ll move on, and people are starting to have that same attitude to, toward the software that they have to use for work. So, I do think that there’s a real opportunity here just to kind of take some of the same people who have built consumer software, some of the same mindset and be inspired by these people and say, “Well, let’s just pretend our end users, you know, in a work environment are just as picky as end-users in a consumer [00:12:00] environment and build it as if it was for the consumers.”

[00:12:04] Joe Garafalo: I love that. It’s such a great philosophy and it’s, it’s, it’s funny that that wasn’t a thought with all the prior technology that, that was built, and it does look like, at Avise, you guys were able to bring over Eric Wong, who was the original UX creator of Peloton. How’s his transition going from building fitness software to finance software? 

[00:12:23] Graham Stanton: Yeah, I think, I think that one was a real coat, yeah, Eric, Eric is an old colleague, an old friend of mine. We, we first worked together and I think it was in 2007 and as ISE, yeah, he designed all the original screens for Peloton, you know, and the evolution there for a while, he designed the Peloton logo and yeah, he’s got a real eye to consumer experiences and he had never built back office software, he never built any financial software and had no experience at all, and, in this category [00:13:00] and we had advised that that was perfect. You know, we needed, we needed a set of eyes unpolluted by all the bad choices that have been made over the past five decades.

[00:13:11] Joe Garafalo: I love that because when you have, when you have that, that first-principles approach, you can make so many things that are just historically accepted better because you’re looking at them with a fresh set of eyes, and I think that’s exactly what financial software back office, CFO tech stack needs for this future wave of technology to, to really win the market, which brings me to our next question here, which is, where do you think, where do you stand, I guess, on the, the software versus Excel debate? Do you think that the future eliminates spreadsheets or just makes up for their weakness? 

[00:13:47] Graham Stanton: I think, yeah, I think the question is a good one, but it’s also sort of a false dichotomy, right? Excel has its place. Excel is amazing software. It’s taken over the world for a [00:14:00] good reason, and, yeah, I was still, I mean, I spend a lot of time in Excel or Google Sheets or some, or whatever spreadsheet tool I’m using and it’s great for prototyping and it is, it’s great for one-time analysis.

[00:14:17] It’s great for that one process that this one company has, that’s just, that is completely unique to itself and doesn’t fit anywhere else. And, but, yeah, anytime it’s being used like a database, when, you know, as a source of truth, I think that’s a big red flag that software would be a better choice in this, in this scenario, and I know that that’s often the case with financings, it certainly was for us a Peloton, and, and, and anytime you have an overly complex excel spreadsheet that’s just being used over and over and over again, and that’s another sign that, you know, maybe this could have been better done [00:15:00] in a tool, and I think, as the tool is mature, you know, people would just naturally reach for Excel less and less, and it will just live where, you know, where it belongs, around the edges.

[00:15:12] Joe Garafalo: Yep. I love that. I think it’ll always have a place in, in the CFO’s tool stick, tool kit, but when you only have a hammer, everything looks like a nail and I think that’s the world that we’ve had to live in with kind of the shortcomings of, of technology, and a lot of folks have been burned by those two-year long implementations of existing legacy, financial tooling, and I think there’s going to have to be kind of this, this education and this next wave of, of marketing that happens to, to get CFOs thinking that, like, there actually can be a better way and excited that, that you’re along the journey, helping, helping build that next wave.

[00:15:46] Next question for you here, and this is really for a lot of the mosaic customers that are listening. A lot of our customers have different revenue streams. It’s hard enough to get the CFO tech stack to produce the outputs that you need when [00:16:00] you have one software revenue stream. You guys at Peloton had several. There was the hardware component of the bikes, there was the software, there was the live classes, and everyone here, if they don’t already have one, should have some, some Peloton apparel ’cause it’s great. Can you talk about how the CFO tech stack, how you made it work for so many different revenue streams? 

[00:16:24] Graham Stanton: Yeah. I mean, it definitely wasn’t easy and, yeah, I’ll say that they’re out of necessity when, as we’re building up the finance function and, and we were, you know, actually maturing our financials and, you know, even, you know, moving towards going public. I mean, we started off by doing a lot of our, a lot of our revenue recognition, yeah, booking, yeah, just outright booking and revenue based on Excel spreadsheets. And, you know, we, we got to a point when, yeah, so the, the big Excel [00:17:00] spreadsheet now for all hardware revenue would crash someone’s laptop half the time, so it’s a, yeah, they’d open it and then wait 10 minutes, either would work or their laptop would, their Excel would crash, they’d kill the process and try again, and then we partitioned it, w, we turned it into multiple spreadsheets, and if, obviously that doesn’t work for analysis if you want to know, it’s kind of, how’s the business doing, you want to start tracking, you want to start tracking revenue per customer, maybe you’re curious about, you know, people acquired via this marketing campaign, “Can I estimate their LTV?” And then, yeah, all that, if all that data is locked in those spreadsheets and not getting into the financials until months later and that’s kind of useless, so we did a lot of this analysis in, you know, in Redshift and Looker, and just, yeah, we have custom sequel first, parts of it and it was this big intensive thing, and when it worked, it was kind of unfortunate, but it took so much effort and [00:18:00] when it came time to go public and, and make this process SOX-compliant, we had a whole other team of engineers build it, you know, a new sub-ledger, yeah, for all revenue that then made its way also to Redshift and in summarized form to our ERP, and, so yeah, giant headache and, yeah, just, just a lot of kind of duct tape along the way, and I think a mature process in the end, but in hindsight, it’s kind of sad that, you know, Peloton needed to be on the precipice of going public to actually pull it all together, and just the number of people involved in that is, you know, painful to think about.

[00:18:41] Joe Garafalo: Yup. And, and that is, I think the the world-class finance function, is when you can take data from all your point solutions, use ETL to pipe them into a database, database, write code to join and blend all those different tables together to give you the right views so that you can see everything in one place [00:19:00] and close to real time, and that world-class structure is every company is trying to build it, and it’s, it’s every company rebuilding it from scratch each time, so, I think, for the CFO tech stack, there’s this chance to build kind of that metal layer that will just snap into all those tools and automate a lot of, like, that, that sequel writing and creating those views so that you can get that stack up and running faster from day one, instead of building it, you know, leading up on the roadshow towards the IPO, so I think that’s a really big area for disruption in the CFO tech stack, and it’s, it’s very aligned with what we’re trying to build that at the moment. But, Graham, thanks so much for talking about some of the pain and the struggles of the CFO tech stack at, at Peloton. It’s, it’s really exciting to hear. Question for you now on, like, just summarizing everything, if you have the right finance and accounting tech stack before, before that IPO, what were some of the things or some of the insights that maybe you would’ve done differently or changed? 

[00:19:59] Graham Stanton: Hmm. [00:20:00] That’s a really interesting question. Surface level improvements, the, would have been, the CEO never would have needed to think about the letters ERP at all. The fact that our, you know, our CFO tech stack implementation ended up being, like, you know, one of the biggest initiatives for the company as a sign of how broken the whole ecosystem is, or at least was, so that, that already would have been a big deal. But, you know, if we’d had that, you know, if we’d had that level of insight much earlier on, you know, I, I think about just how dramatically better our operations would have been. I mean the, you know, w, we were a very fast-growing company and, you know, when you think of the size Peloton is now and used to, you know, you think if I, think, “Oh, you know, Peloton maybe spent 10% too much in this area early on in its existence,” [00:21:00] yeah, that, that’s a rounding error in the financials of Peloton today, so it’s just kind of, like, you know, grow fast and, and, and do what you have to do, but the rail doesn’t work out that way. I mean, we, we always struggled to raise money until we had, you know, a very large number of paying subscribers, you know, VCs always doubted the business model. Everything was a struggle and, yeah, in hindsight , a lot of that struggle was just kind of poor visibility into the core economics of the engine. They, you know, we knew, you know, we knew what was important for the company and kind of how to prioritize, but in terms of presenting the business, you know, if, yeah, if we would have had that insight in terms of, like, “Okay, this is how to do manufacturing. This is, yeah, how to do your, how to manage our third party logistics relationship,” yeah, I think back to marketing and, and the, [00:22:00] just the sheer amount of work we did on insight there, if, if we had that properly coupled to the rest of the business, you know, we’d probably would have spent less on marketing in some areas and dramatically more than others, and that, they, I get frustrated at, at the time sync due to having bad insight and bad software to use, but the impact of the business is so much bigger than that. I mean, it could have been that, it could have been the difference between success and failure that I think, you know, we’re lucky we pulled through, you know, the degree of difficulty, it was so hard, I can only imagine the companies out there that they would be very successful today, but, you know, had to close shop before they ever were given the chance.

[00:22:42] Joe Garafalo: That’s such an insightful perspective. I would’ve never guessed that with all the success that Peloton has and our friend Jack McCullough has, has a really good quote, he said, “CEOs should value a CFO that is a really good storyteller,” and you could only tell that really good story of the [00:23:00] success that Peloton was having if you have good insights into the numbers and you can piece together all of those analytical insights about the supply chain, about the three PLS, about the marketing data to, to tell that story, so it’s really cool to hear that, if you had better technology, that storytelling would have been easier.

[00:23:17] Obviously it doesn’t matter to all those VCs, I think you just grouped them very wrong, which is, I always love that, but congrats on all the success. I think, last question for you here today is, we always ask this to all of our guests, if, if somebody starting their career in, in finance and aspiring to be a CFO or Head of Finance, what’s one thing you recommend or wish they knew when they were, when you were starting your career? 

Career Advice from Graham Stanton

[00:23:43] Graham Stanton: Yeah, I think just to, just to stick with the theme of this conversation, I think a solid understanding of the, I guess of data, to do it more broadly, you know, through what we’re doing at Avise, so I would tilt that a little bit to the, it’s good to understand [00:24:00] accounting, but ultimately accounting is really just, get the data underlying the business and, you know, and it, flowing through to then, you know, the story of the business that you can only tell a good story about, about how the business is doing based on the underlying data, you know, that, that’s kind of the CFO’s job to be able to interpret the data and tell that story, and I think there are a lot of operational aspects of being a CFO that, you know, that, that are independent, but almost all of them are held by having that really good base for knowing where everything comes from.

[00:24:38] Joe Garafalo: Yeah, I think that’s so true. I mean, a degree in finance, a degree in accounting, those things are table stakes now, and it seems like the future is actually being able to understand the data and all the different data sources and how to piece them together to tell that, that story. So, I 100% agree. Cool, before I let you go here, Graham, what is your favorite type of ride over [00:25:00] on Peloton? I, myself am a low-impact guy, but curious to hear what you enjoy. Okay. 

[00:25:04] Graham Stanton: No. Well, that’s a good question. I mean, I love all of them. I love the, I love the low-impact rides as well, and I use them as pillars, as recovery. I, I’ll go to the extreme opposite though. Yeah, I love the HIIT and the Tabata, just for the, for how good a workout you can get in, and, in such a short period of time, but I made the mistake of doing those two minutes, yeah, too often, and the, I really got to mix it up.

[00:25:26] Joe Garafalo: For sure, for sure. Well, very cool. For folks that are looking for information on Avise and we over at Mosaic, or our customers and are really enjoying what you guys are building and unlocking that, that accounting function to drive business forward, where can folks go to learn more about what you’re building at Avise? 

[00:25:44] Graham Stanton: Well, we are, we have just recently launched our Beta, so, you please go to our website, avise.com, A V I S E, and, yeah, just drop your email address in there, or, you know, feel free to email me directly, graham@avise.com.

[00:25:58] Joe Garafalo: Cool. Well, Graham, [00:26:00] thanks so much for being here, really enjoyed the conversation and looking forward to staying in touch shortly. 

[00:26:05] Graham Stanton: Yep. Thanks so much, Joe. Really enjoyed it. 

[00:26:07] Joe Garafalo: Thanks Graham. Bye.