Episode - 027: Where we learn from Abhijeet Sarkar, the CEO of Types Sift, about the transformative power of data in the world of business. In this insightful conversation, Abhijeet shares his journey from breaking into the industry with Arc'teryx Equipment to creating Types Sift, a tool that revolutionizes how businesses interact with data.
Abhijeet delves into the evolution of data storage and access, highlighting the shift from pre-aggregated cubes to the more dynamic, bottom-up aggregations enabled by cheaper RAM. He also shares compelling examples of how digital transformation can lead to exponential ROI, with businesses scaling their customer base and improving reporting cadence by moving off spreadsheets.
This episode is a must-listen for anyone interested in data analytics, digital transformation, and the future of business intelligence. Tune in to discover how to distinguish useful data from noise and make your data actionable. Don't miss out on these valuable insights from a leader in the field!
Where to find Abhijeet
- [00:00:00] Introduction and background of Abhijeet Sarkar, CEO at Types Sift, and the genesis of Types Sift
- [00:12:51] Discussion on the evolution of data storage and access, from cubes to data flattening
- [00:37:26] Explanation of the concept of digital transformation and its irreversible nature
- [00:54:53] Talk on identifying low-value activities and automating or streamlining them
- [01:07:38] Closing remarks and where to find more about Abhijeet and Types Sift
So there's two ways that data can help you scale revenue. One is what you just described, which is unlock new revenue opportunities where we may want to collect data from multiple sources. Analyze it and then make a decision based on that. It's not always a clear decision, but sometimes we need to just run an experiment, test something and see if it works.
Sometimes it's for a B2C company, it's launching a new SKU new flavor, new offering, new discount, new promotion, whatever it may be. Test how it works, and then decide to cut it or expand or double down on it.
Hello and welcome to Tech for Finance, where we help finance professionals leverage technology to level up their lives. I'm your host, Adam Shilton, and in this episode we're going to be chatting with Ajit Sarkar, CEO at TypesSift who helps CFOs use data to measure what's driving their EBITDA before Types Sift Abhijeet spent many years working in the business intelligence space and has worked with companies of all shapes and sizes, helping them make more informed decisions from their data.
In his free time, he enjoys road cycling, which in his words is about all the free time you can afford. Um, but before we get into the conversation, if you like what you're here today, please make sure to subscribe to Tech for Finance on your favorite podcast platform and on YouTube. But it's great to to have you on Abijit.
Really appreciate you taking the time. Thank you for having me. I'm excited. No worries. So yeah, types sift. Obviously I know a lot about the business, but the listeners won't. So do you wanna give a little bit more about your background and what led up to you creating types sift? Yeah, totally.
So, uh, my background is in business intelligence, data analytics.
Uh, I kind of got my break into the, into the industry back in, I'd say 2013, working at a company called Artix Equipment, huge apparel manufacturer on the west coast here in Canada. Um, and so our team, you know, spent all day cleaning, preparing data, getting it ready for analysis, and then of course building dashboards for the end users.
And we found that users would come to us to make slight tweaks of the dashboard, Hey, make this change. Hey, make that change. Can you, can you create this? And we thought, I thought, Hey, we've got all this data. It's all clean and prepared and ready to go. Wouldn't it be cool if we could just give them, uh, like a search bar where they could type in the thing that they're looking for, uh, and like ask a question that generates a charter or graph from the clean data and then they can kind of take it from there where they wanna be.
And then my team could just focus on, you know, kind of preparing and cleaning the data. So that was the, the genesis of the idea. You type a question generates data, and then you sift for your answers. Um, and so we took this out to market probably in, uh, 20 20 14, 20 15. And, uh, what ended up happening was I realized at the time, in retrospect, it was a very naive assumption.
I had assumed that every company had a team like mine who spent all day cleaning and preparing and modeling data. And as it turns out, that was not the case. Uh, many companies, their data's in a dozen different places, different granularities coming in at different times, and they don't have the people internally to kind of clean and put it all together.
Um, so forget about building dashboards. Like they just need to help getting that infrastructure set up. So we did a bit of a pivot to say, Hey, maybe we can build a suite of tools, uh, in addition to the dashboarding piece that actually makes that easier for these companies to get their data together, get it clean, get it consolidated.
And then spit out that dashboard. And so that was the, the kind of major pivot that we did. Um, and now the company's grown from there to include planning, forecasting, report distribution, all these kinds of things. But basically to make it easier on these companies to, to make sense out of their data. Yeah, it's
Thanks. And it's obviously working well, right?
So obviously working well. Absolutely. We've been going since 2013, so I'd say so. Yeah.
Very good. No, I, I love, you know, I love small business stories and yeah, that's why, what's, why I have these sorts of guests on, on the podcast, right? Cause I'm, I'm keen to, to tell the story.
So, um, so just a bit of a question from me, and I know we had some, some conversation topics, but we might go a little bit off track. So when we look at data, obviously volume of data now is, is going up and up and up, right? Yeah. Um, So obviously more data to manage, more admin overhead, more low value tasks, and so on and so forth.
So, so obviously there is a business case for these tools like you provide to, to amalgamate that. Um, I won't profess to be a data expert. I, I have previous experience, um, and it's a totally different story. Um, back when I was working in the agricultural space, I had to amalgamate data from I think 26 poultry sheds in Peru.
And I did it, you know, did did it on, in, in one dashboard and it, it was a pain cause it was all spreadsheets and, you know, uh, all of that sort of stuff. So I guess the first question, um, is something that's maybe a bit actionable. So this is maybe more from your business intelligence background rather than, um, stuff types of, so let's say we've got a business and they might not have an e r P system, you know, so they might have, you know, uh, a solution for that, solution for that, you know, so solution for finance, solution for operations, task management, project management, and so on and so forth.
When I tried in the past, whenever I go and look at new solutions, I'll look at the integration piece. Right. You know, so what system will talk to other system? But of course, the limitation you get there is what system becomes the master, uh, master system. Um, and actually a lot of these systems won't offer any sort of, um, visualization or any sort of scenario planning or anything like that.
Right, right. So with your tools and with any of the other tools that you, you've previously in, in different companies, is it best practice to set up something like a, a data lake, you know, to, to amalgamate all of the different data sources before you then connect it to a tool like yours or, or another tool?
Yeah. Are we now saying that there's less need for something like a, a data lake or a data platform and that you can just plug multiple applications into a, an intelligence
tool? Yeah, it, it, it all depends on the size of the company, the level of maturity they're at in terms of their existing systems, and then kind of like how far up the growth curve they, they happen to be.
So I'll make a couple of distinctions and terminology. Data lake is just a place to dump data, right? We just need to stick it in one spot. It's unstructured, it's, it's disconnected. There's, you know, it's just bringing it into one place and, and many times it's not even a database, it's just a place to stick files, whatever, in whatever format they may be.
So that's a kind of a data lake. And if it's never. If it's not properly maintained, it can become a data swamp. So that's a kind of a different sort of piece. Um, usually where the kind of the traditional approach has been to build a data warehouse and the, there's some very strong rules around, uh, what goes into a data warehouse and how it's put together.
The, the challenge with my industry is that the same name can mean different things. So today, data warehouse could, you know, could mean if you're thinking Cloud data warehouse could actually mean the database itself. My kind of more. Purist background in business intelligence is that a data warehouse is a design model, and it's a blueprint of like how your data fits together and then how you implement it could be in the cloud or on premise or whatever technologies it may be.
So usually it's the, the data warehouse is kind of where everything comes together and all your data is cross connected from different places, from your crm, from e r p, from your marketing system, from this and that. And it all knows how to kind of meet at a common granularity. Um, and that has been the case since I'd say.
I think IBM actually invented the data warehouse in 1987 as like the design principle. Um, and that has been the case for. 30, 40 years since then, uh, what we are doing specifically is kind of disrupting that model where, you know, you would implement a data warehouse separately, then you would connect in the tools that you need.
Then you would kind of, kind of go from there and build the sort of business intelligence or analytics stack. What we're doing is kind of saying, Hey, all of that stack is deployed as a single system and then kind of deployed modular. So that, that answers your question of, you know, is the. Is the notion changing to say, Hey, can we just connect directly into these systems?
The truth is you'll still need something to kind of act as like the central source of truth. Um, kind of the, the, the system of record as it were. Every operational system like a CRM or an E r P or warehouse system would like to say that they are the operational record of truth. But the truth is, if you're gonna pull one of those out at some point and put in a different one, as you continue to grow, you need something that's going to outlast them.
And that's typically where analytics, data warehousing, that kind of stuff comes into play. So whether it's in an all-in-one system like ours, or you're implementing it on your own, typically the data warehouse or some kind of analytics store is going to be the system of record for the next 7, 10, 15, 20 years of a company's life.
Okay. And thanks, thanks for breaking that down. And I, and I love the example of the, the data swamp. Um, but, but yeah. So yeah, I think that's gonna be useful for, for people wanting to, to demystify that piece. And I'll add another term in, um, a data cube, which is one that I heard. Yeah. Again, it's, it is probably, probably marketing terminology and, and all of that sort of stuff.
But no, I remember, so I've got a bit of a background in, in dynamics Microsoft Dynamics. Um, and I think it's improved since then. So we're talking years back, right? Um, and we're getting a bit technical here, so it's so better. Yeah. But, um, obviously in the backend of, of all of these systems, you've got a field that is capturing data, right?
And a field will be named something and hopefully the naming convention of that field makes sense so that if you are trying to get data out of a system, it's easy to understand what data is in what fields. But at the time, and, and I think it's cuz the, the fact, you know, it was originally developed in the Netherlands and I dunno whether it was a language thing, um, but to connect dynamics to another system, you'd not just need access to the data, you'd also almost need like a translator.
So there were all these companies that were not just. Extracting the data, but they were also translating all of the data fields. Cool. Yeah. What they were. Yeah. Yeah. So, so, and they, you know, some of the providers ended up, you know, creating the concept of a, a data cube with the idea of a Rubik's cube, that as soon as you have all the connecting points together, you could just twist the cube and make the linkages of the different data in the different systems and Right.
Put a bi front end on it, you know? But I think that's just a data warehouse in a, in a different term. Right. Is essentially what you're saying there. Almost.
So in up until the nineties, this concept, I think even the late nineties, the concept of the cube that you just described was a, that was very much the predominant design and just get a little bit more technical.
The idea was that if I want to, why the name cube comes up, is that if I'm just measuring, say for example, sales. That's a, that's a point, right? If I'm measuring sales over time, so one dimension of time, that's a line. If I'm doing sales by time, by geography, I now have two dimensions. That's a plane. If I'm now throwing in sales by time, by geography, by category or product, I'm now in three dimensions.
That's a cube. And if I keep adding more dimensions, you've got what's called a hypercube, is kind of the name. And so this was very computationally intensive to answer these questions at, at the time, like think about it, in the nineties, we still, there were no SSDs. Internet was barely a thing. There was no high speed internet.
You know, I think we were measuring CPU speeds in the, still in the megahertz or maybe in the gigahertz. And RAM was very expensive, right? Mm-hmm. You, you, you did not have more than one or two gigabytes of ram. Um, and so you had to pre aggregate. The intersection of all these dimensions I had for every combination of sales, by time, by geography, by category, by whatever.
We had to pre aggregate it and store that in a cube. So accessing the sales number for number of jackets sold in Japan in 2014 was a very quick, because we pre aggregated it and we spent six hours overnight pre aggregating things, and we were good. As soon as somebody wanted to ask a different question, add a new dimension, or ask something that wasn't in the cube, we had to tear down the entire cube and then rebuild everything, and that could take days to do.
And then the early two thousands rolled around and kind of like, this is around the, the time of the.com, uh, kind of boom. And, uh, Ram became very, very cheap. And so now we could suddenly start pushing all this data into Ram and keeping it memory, which is very, very fast. But on top of that, we could actually flatten the data.
Instead of it being in a cube, we could actually flatten it, uh, into another format that allowed us to run aggregations from the bottom up. We didn't have to pre aggregate anything. So Cube started to kind of fall away as like a more, uh, legacy approach in favor of these newer, these newer, pro newer being in the early two thousands.
And things have, again, since evolved since then. So, um, cubes are, are quite old. I still, I still chuckle to myself when someone says they're looking at implementing a cube and I have to scratch my head. I'm like, is this, is this 1994? Like, why are you still using cubes as crazy? There's a lot of other technologies out there that will serve it a lot better.
Okay. But, but yeah. Okay, fine. So either way, we need to find a way of making sure data's clean, right? Yes. And that it's in a format that's accessible. And from what we've just said there, you know, if it's a long term play and you know, you potentially looking at evolving your systems over, you know, choose your example 5, 7, 10 years, it might be a good idea to set up some sort of data warehouse whereby you've got a consistent source of information no matter what, what system, how
you're, how you're putting in.
If you wanna have longevity over the next 5, 7, 10, 15, 20 years, that's typically where something like a data warehouse is going to, is going to
help. Yeah. Yeah. But as you say, more and more now, there are solutions like types, if that don't rely on the warehouse to get the data you can bring that into, to the system as opposed to needing to connect to a data warehouse first, right?
Essentially it deploys a data warehouse on your behalf. So it never goes, it never goes away. But, but the more modern tools are making it easier to deploy those things. Data warehouse projects can take 24 months minimum. Um, it costs anywhere from a half a million to a million per year. And then determining, quite often companies don't succeed in the implementation.
They actually experience a negative roi. If they don't do it correctly, they don't have the expertise to do it. So data warehouse projects can be very, very risky. And that's, that's the, that's the push towards these new tools, which we would fall into. The class of these tools that implemented on your behalf is the idea.
I had no idea. Yeah. They're very expensive and they're very, very risky. The return is amazing if you can pull it off. Yeah. But if you don't then the, the downside is quite large.
Okay, fine. So, and this was on the set of question topics, so we can, we can go a bit bit back on track, but thank you for that.
Think it was really Yeah, absolutely. Um, so it comes to, um, Data as noise is, is how I'm gonna summarize it. And, um, had a discussion not too long ago. Um, so had the conversation with, with with Taja that came out to today. Um, Taja. Perrick great. He, he does work in the BI space as well. So, uh, a lot of passion there.
It was really good conversation. Um, but then previously to that we had, uh, tamer at causal who came up. Um, and the principal is the same. Um, you know, it's not just enough now to have good data, you know, it needs to be actionable. Yeah. Um, so as a, as a long way of asking the question, if we've now got so much data, um, and we've got so much to choose from, how do we segment the noise from the stuff that's actually useful?
And maybe you can give a couple of anecdotes on some of the specific use cases, the data you, you've found, but I think it might be useful for under that, for people listening to understand the sort of data that you is just not useful compared to the stuff that really is. So, I dunno whether you can talk about that a little bit more.
Yeah, so I've got a, I'd say, and I'm, you know, I'm kind of putting my finance cap on and I'm looking at this through a finance and business lens. So this is a little bit of a biased answer and I'm sure there are folks out there who will staunchly disagree with me on this. But in my opinion, any, I'll kind of reframe it as like noise versus usefulness to just like what's the highest priority?
Because you could, you could make the case of any kind of data in the company's worth analyzing. It just depends on what's the highest priority of things to analyze. And my opinion is that the highest priority is anything that is going to touch the p and l first and foremost. And then depending on the, the company, either the cash flow statement or the, or the balance sheet.
And these are like the three most important reports, at least for finance, that, that are generated. Um, and so the way that we kind of think of data driven organizations is finance is sort of like the, the core data function in the company, mainly because whatever they're reporting on, they'll have to pay taxes on.
So you have to get that right and they're ultimately the, the stewards of the data. So they're kind of at the, at the core of the company in terms of the data function. And if you sort of go out in concentric circles, the other departments, depending on, depending on the, the industry and the business who's kind of further and further away would become lower and lower priority, uh, in terms of what you're looking for.
So, It depends on how the company makes money. If most of the, so certain software companies where it's, you know, based on usage and the more users using it, the more comp money the company makes, you know, outside of the p and l probably the next most important piece of data to analyze is usage rates or active, active, uh, monthly usage or whatever it may be.
Like how frequently we're able to upsell 'em to the next tier. Um, if you're selling contract services, maybe the next tier outside of that is our sales team. How frequently are they, are they kind of achieving target in these kinds of things? So it depends on every industry. And then kind of moving out from there, what do you think of as sort of like the least important thing to look at?
Um, yeah, I'd say that's probably my very biased answer given that I, I have a finance background as well, and this is a, a finance podcast, so, yeah.
No, excellent. It's a good answer. We, we like it. Um, okay, fine. So, so. Let me, I'll, I'll elaborate a little bit more. Um, just from, from some, some of the stuff that I've been doing recently and we'll, we'll encroach a little bit on, on ai.
Mm-hmm. Um, even though we've done a lot of podcasts on AI and it's becoming a bit fatiguing now, uh, cuz the takeaway is don't panic, it's gonna be built into everything. You know, don't think you have to a master of ai, even though, you know, it's still a skill to be built in the same way that others are.
But there is now, uh, the massive emergence of all of these platforms that are allowing you to visualize your data in a certain way, you know, chat to your data and, and so on and so forth. And, uh, you know, the example you gave, you know, types of right at the beginning was, you know, what, if we had the ability to, to talk to our data, you know, um, to type to get those results, it's not, not changed that much in, in, you know, with, with all of these new AI tools.
But I've been doing some experimentation and I've got dummy data from finance systems. I've got dummy data from CRM systems, and when I do my testing, I, I'll literally just do like a CSV export. Right. And, and I, and I won't do any data cleansing because I'm, I'm keen to see how well some of these tools tackle potentially noisy or, or, or rubbish data.
Right? Right. And what I've found is, and it comes to your point here about first having that focus on what the most relevant priority is. Mm-hmm. Because if you just upload a spreadsheet that's got 30, 40 columns, some data relevant, some data less relevant when it comes to chatting to these systems, um, your results aren't going to be as good as if you had a smaller set of focus data.
Because that AI namely down to the fact that a lot of them don't have context yet. They don't understand your business, they don't understand where you're coming from. They'll make their best guess from the total data set. Right. So you can sometimes get some bizarre visuals. Um, information will be produced that you didn't know exist, could be totally made up.
Right. So I just wanted to enforce there that, you know, When we talk about data, it could be quality data. All of that could be quality data, right? You know? Mm-hmm. Um, it could be completely accurate. Um, but that doesn't necessarily mean that if you feed it into a tool, you're gonna end up with the results that, that you need.
Right? So if I've understood you, uh, correctly there, aachi, what we're saying is that there's potentially first an exercise that says, hold on for a second, you know? Mm-hmm. Let's, let's reduce, let's refocus, and then make sure that we're modeling based on only the relevant data.
Right? Yep. Pri prioritize what you wanna look at, essentially.
Yeah. Okay. Fine. And again, not in the question topics. Apologies. Um, so
Um, so let's, let's take a, a couple of example systems. Um, and again, you might not know the answer to this, but I'm curious because it's, it's something that I've been focused on a bit, uh, recently. So, In an eerp system, um, or a finance system.
I've got, as you say, you know, p and l data, balance sheet data. Um, I've got journal entries, I've got invoice entries, I've got, you know, if I'm using an E R P, I've got stock information, I've got uh, inventory validation data, I've got all of that sort of stuff, but mm-hmm. The majority of it tends to point towards finance in some shape or form, you know, so if I'm, um, putting goods into production, that's gonna take it into work in progress so I can see the valuation of my stocks and work in progress and so on and so forth.
And the same in the finance system, you know, if I'm posting an invoice, then of course that's gonna have a direct reflection on the financial data. Yes. Yeah. What I've been taking more of an interest in recently, and again, there's probably different variations of this, is statistical data. And from what I'm seeing, especially with the advent of finance business partnering, becoming more relevant to, to the finance community, is that.
Finance and now having to step a little bit outside finance only, um, not just in larger businesses, in smaller business as well, to take ownership over that data. Right. What I'm getting at here is there is some statistical data that just won't exist in, in any system. Right, right, right. So in some sense it might need to be manually entered, you know?
Right. So, um, I had a conversation recently, you know, can we input time sheet information and grant, it wasn't a system, it was something like Jira into our financial system so that we can then get, uh, an estimate of, you know, project spend based on the amount of time per individual and that, that sort of stuff.
So that is driving towards finance. But I guess the, the question for you is, Do we need to be thinking about automating the entry of that sort of data? Or is that the sort of stuff that it's okay just to say, right. Well, we've only got a small amount of data. It might be one spreadsheet, we'll bring it in.
I, I sometimes fear that people immediately go to automation and how do we automate this process of getting data? Yeah. I don't dunno whether you've got experience of that or whether it's a scenario that you come across often, but I'm, I'm just curious to ask the question. Your
question makes me think of the old joke of, uh, why, why spend two hours doing something when you could spend a week automating it?
You know what I mean? Like, it's, it's at a certain point, the thing about automation is that it's, there is a certain setup cost to it. It takes time to set up the automation, to implement it, to get it working. Once it's set up, the return that you get on it will continue on for years to come. But it, it, every company has to make a decision.
Around whether it makes sense to automate that one little spreadsheet. If it's literally just the one spreadsheet, uh, that is shared, that can be easily sucked into another system or sucked into the, the system of record, it may not make sense to try an automate that in some way. Um, it might not even be possible to automate that, especially if it's some kind of manual entry.
You'll still need some form of data capture or some kind of portal or interface. This is kind of where a space that we play in is providing those interfaces, but the question still comes down to is it worthwhile? Are the savings there, um, to actually transition from this being one tiny little spreadsheet into something that's more robust.
The, the more compelling reason that we've encountered was if it's a single spreadsheet, um, that only one person in the company knows how to use, and there's enough complexity to what's going on in that spreadsheet, there is value in. Creating automation and you know, putting software around that only because we're trying to de-risk from a business continuity perspective.
We're trying to de-risk that one person and that one spreadsheet that you can imagine that if one person is managing, if their whole job is to manage this one spreadsheet, That's a pretty expensive, uh, you know, kind of proposition. Number one, that person's never gonna be able to take a vacation. Um, they'll never be able to get promoted cuz they'll be stuck in that role.
Or they'll be training the next guy that comes in or gal that comes in. Um, and it's a lot of risk for the company. Uh, if there's just one person who understands a single spreadsheet inside and out, it, it's dangerous. So there's it, it's more than just the cost versus, you know, return of, of automating something because there is a setup to automation.
Sometimes it's just, do we wanna make sure that we've, we've hedged our risk in different areas.
Hmm. Now that, that makes total sense. And, and to come back to your, your previous point, you know, we, we've gotta assess I guess the, the level of need behind this before we, we, we make any decision. Right? So, so the example that I gave there, you know, Very simple example, we want to weigh time spent against project spend so that we can work out on average our hourly rate, you know, and that the, the, the vision for that might be, you know, do we need to put up our prices?
You know, so, so is it a strategic decision that says, well, at the moment with those rates, we're running at a loss, you know, or are we spending too much time, you know, do, do we need to be, um, harder on our customers and not as flexible when they request a change for something? You know, is there a sales education piece that says, look, you, you are always selling, um, under, you know, we, we need to put another 10 or 20 hours worth of buffer in as part of the sales process so that there's not as much of a delivery issue.
But then in that instance, it could literally be data dumped from Jira. Mm-hmm. You know, upload into whatever tool you're using. Right? Right. So, so for me, in that instance, two clicks. Doesn't warrant a whole load of automation, right? No. If, if, coming back to your example about the complex spreadsheet, you know, if it's not just a data dump from, from a, a system and there's a load of complex calculations in those spreadsheets almost to the point where you're doing a little bit of modeling in the spreadsheet for ends up doing in the tool, that's where you've then got to assess, right?
Well, you know, yes, it might be a bit of a pain, but as you say, to reduce that risk to do, do we automate that piece? Right. Um, another one i, I had recently, um, it's quite interesting when you, you hear some of these stories and um, it's, uh, it's a tour company. Um, and again, using basic systems at the moment and during conversations we're, we're trying to tease out, right, well, what sort of reporting do you, do you want to be able to, to produce as opposed to to what?
And, and a lot of the time people, you probably find this as well, people just don't ask themselves that question because they're so swept up in, we're struggling with this now. They don't really have time for the, what would we like to do sort of piece. So, you know, Ted just mentioned it on the last call as well, you know, what are your KPIs?
Where are we losing time? What do you want to be able to see, what in insights you want to be able to drive? And that then led to this conversation about statistical data because they want to be able to do it, uh, by passenger analysis. For example, you know, we sell a tour. Um, how many seats have we booked on a bus?
How many seats have we booked in in a hotel, for instance? You know, they're just not doing that at the moment. But the thing is, they do have that data in their systems. Yeah. Yeah. And, and sometimes it could just be a miscommunication with an IT team or, but people aren't aware of some of the data that they have access to in their systems already.
So that then just comes back to asking the question, look, I'm looking at producing this sort of report. I need data. Do we have it? And then you can make the decision. Is it, you know, is it a live integration between the systems or are, again, we doing just a couple of clicks to get into the system? So Right.
There wasn't a question there, aji, but, you know, just a, a story to, to back that up, I guess.
Well, to build on what you said. You, I noticed you, you used the phrase, what do you want to understand? Do kind of like, and I think the key word to pick on a little bit here is the word want. Mm-hmm. Um, I would almost say, you know, it, it's better to try and understand like, what do we need to analyze?
And that need is either gonna come from something that's happening in the business. A great example would be top line sales are soaring. Net income is flat or is going down, we don't know what's happening in between those two lines. And so we need to break this down and figure out what's happening because we're losing money somewhere and we don't know where.
That's a great example of saying, Hey, we need to do this and that's why we need to run this analysis. Um, so that could be something that just is like a clear need in the company. But another thing would be need is also set at the leadership level. Um, and in every project that we've seen was a strong success.
There was a strong, uh, sponsorship from either the ceo, the cfo, or the c o o where they defined the culture and the need and say, hey. We have this data, there's money being left on the table, you know, even though we're making money and we're growing and everything's fine. Like what are we, what are we leaving on the table?
And that gets set from the top down. And literally every analytics project, especially when, when Arc Alteryx founded its analytics, uh, team internally, it was literally the, the general manager who was like, we need to do this, right? Mm-hmm. And, and I as a, as a young kind of early 20 somethings, got to sit in a meeting with the, the ceo, general manager of the company, along with our, my director of IT at the time.
And we got to plan these things out that they, these kinds of decisions have to start at the top and kind of work their way down. Mm-hmm.
And so it's so valid. And, you know, obviously language is potentially a whole topic of compensation for another day. Right. But I think, I think you, you're absolutely right to break that down because individual wants are often trumped by, Critical business needs.
Right, right. Um, and again, comes back to that, that business partnering piece. Right. You know, so, so if you are an advocate of business intelligence and reporting and analytics, and you're going into a business in a, you know, an fp and a role, or a, you know, what, whatever it happens to be Yeah. If you can suss out what the needs are, you know, and, and, and I see it so many times, you know, people come in, you know, they've run projects before and they're like, you know, I, I want to do this because I think it's the right thing to do.
You know? Right. You wanna do this because that's what everybody else does, you know? Right. Not necessarily the case, you know, so I think you're absolutely right to stress the need there, because comes back to prioritizing, right? Mm-hmm. Focus on the critical needs first, and then maybe you move into the, oh, this, this would be nice to have, would be quite, to have that instead.
Absolutely. So, absolutely. I think you're absolutely right for, for pointing that out. Okay. So coming back to, to your systems for, for a bit then. So, um, I know you guys are keen to, to produce information that supports, um, ebitda, for example. Yes. And, um, in, in the LinkedIn community, I don't know if anything else, but, but I tend to see, um, arguments on, uh, the, the validity of using, um, EBITDA as a, as a measure, um, for, for the value of a, of a business.
Um, my argument is generally, you know, is is there a better metric? I, I don't know. You know, so, so it would be good to get a bit of your experience, you know, why that's a focus for you, um, and what you think the value is in, in that, especially maybe for, for smaller businesses, if that is your target
Yeah, I mean, there's all these different kind of ways of measuring EBITDA and there's adjusted EBITDA and there's EBIT and all these kinds of things, and it's always a little bit different. Um, I, it's, I would look, I'd say it's, it's not the only metric that should be. Taken into account when assessing the valuation of a company.
Um, but I'm quite certain with some ex, with some notable exceptions, uh, I'm quite certain it's one of the most important values or numbers or KPIs that are used because a lot of businesses are valued based on some multiple of ebitda. And the thing about EBITDA that keep in mind when you set aside, how much does it, what does it mean for the valuation of the company and these kinds of things, like at the end of the day, it, it helps you to get a picture and understanding of how efficiently the company is being run with the resourcing that it has right after you've taken into account your variable cost of sales and all these kinds of things.
For a lot of businesses, your fixed cost is such a big and important piece of running the company. EBITDA tells you are, are we as efficient as, you know, the industry average or the industry standard? Can we be more efficient or what can, what do we need to change here? So e just in looking internally, if you have to run the business by the numbers, which we all need to be doing, um, it's a very important number, at least as a good first check to make sure that we're, we're, we're on the right track.
Is it the only number that you should use when valuing companies? You know, there's so many different ways that you can value businesses, whether it's through comparables or through some multiple of ebitda. You know, historically tech companies have been valued based on some multiple of top line revenue.
But I think that even that, you know, given the current landscape is now shifting more towards taking into account, you know, efficiency in EBITDA and these kinds of things, it's, it, is it the best number that we have to use? I won't comment on that, but I will say it, it's not a number that's gonna go away anytime soon.
So it's something that shows up on every p and l and that every CFO has to care about it. It's just not going away anytime
soon. And what, what are the, and you'll have to forgive my ignorance cause I don't spend a huge amount of time in this space, but in terms of key data points, when we are looking to calculate that mm-hmm.
You give a quick overview.
Essentially it's just start at the top of the income statement and, and work your way down. So top line, take out your cost of sales, et cetera, et cetera. You get to essentially what I would term gross revenue. And then you start backing out your fixed. Kind of recurring costs. Uh, and then anything else's kind of like between those two lines.
And you get to ebitda. EBITDA is earnings before interest, taxes, depreciation, amortization. So it's, it's a number that gives you a reflection of the performance of the company before you start taking into account things like, how much do we need to amortize, you know, this investment or this asset, or depreciate that piece of capital.
We paid this on taxes, receive this in interest, or whatever it may be.
Okay. Thanks for that. And as I say, forgive my ignorance and, and anybody who's listening, I, I apologize cuz you, you probably already know all this stuff. So just one of those things. And, and, um, before I forget, have, have you heard of, uh, a guy called Alex Hor Mosey?
I have not, no. So I'll, I'll send you the link. Um, so he's recently started, well I say recently, I think, um, I dunno how long it's been founded for. His company's called acquisition.com. Okay. Um, and before that he, he basically built his business, um, by building gyms. And it's one of those underdog stories.
You know, he got, um, shafted by, you know, another gym owner who, um, basically, you know, I, I can't remember the exact story, but basically he ended up with nothing and then he ended up basically on a cold calling spree and he built a business from nothing. And, and it is, it's one of those really good stories, but he then figured out, actually he's pretty good at scaling businesses, right?
Um, so he sold a few companies and now he's at the point where he is formed acquisition.com. Um, and his LinkedIn tagline is Now, uh, we are buying 1 million to 10 million EBITDA founders. Uh, okay, there you go. Yeah. So, so yeah, I mean, it, it just proves that it is a metric and people are using it as part of these, these decisions.
I guess so, but
yes, matter worse, it's the number that most people look at, um, especially in finance, so, yeah. Yeah.
But no gi give him a follow. Um, he's massive on YouTube now. He's, he's, he's even got to the point where he says that you can't really go outside. And, and I can't even imagine that level of, of celebrity, and again, we're going off on a complete tangent now, but he's, he's kind of gone past podcast.
Cool. Maybe you might be recognized by, you know, one in 10,000 people saying, oh, I listened to your podcast. He's not full on, gets swamped in, in the streets, right? So he is got hundred thousand followers on, on LinkedIn, but, um, he's really good with YouTube shorts. So those, those sorts of quick bits of business advice.
So I know this is a finance podcast, but it is quite refreshing just to have those, you know, quick nuggets of information. So Alex Hormo, that's H O R M O Z i, um, accession.com. So, um, I dunno why I'm plugging because he probably doesn't need it, but there we go. So, um, going, going back to the, the transformation piece a bit, um, I saw I post from you recently on, uh, LinkedIn.
Mm-hmm. And it might, it might be a short description, it might be a bit of a longer one, but it piqued my interest because the, the theme was essentially, um, digital transformation and fundamental changes to work. Yeah. And I won't explain it any more than that. So maybe you can take us through that, where that came from.
Yeah. And just, yeah. So that actually came from a podcast. I was, listen, a different podcast that I was listening to, um, who was actually hosted by our colleague, uh, will Baron, in which he had a guest on who specializes in having transformational conversations. Her name is escaping me at the moment, but she, she gave a definition that was kind of like very groundbreaking for me in its simplicity that a transformation is a change that cannot be undone.
Um, and, and to, to credit her. I'll, I'll, I'll dig up the, the reference a little bit later. But, um, it's a change that you can't undo. You can't unscramble an egg. Right. You can't unmake a cake. Um, it's transformation's very much a one way journey or one direction journey. What's interesting is when you tack the word digital in front of that, which we all tend to accept, means that it has some involvement with tech technology and quite often software, some kind of IT infrastructure.
What's interesting is that from my experience, a digital transformation is one in which you will not want to go back after it has happened. Um, no one in the history of time. Is halfway through an omelet and said, you know, I kind of actually wish I was eating an a raw egg right now. That is, that has never happened.
No one's ever had a piece of cake sat in front of 'em and said, I kind of wish this was flour and milk and butter. That has never, ever, ever happened. And so when companies undergo a digital transformation, such as implementing an E r P or any kind of software or even infrastructure like the internet, right, you, you'll never want to actually go back.
I implementing analytics and, and business intelligence, you'll never want to go back the roi. Is so self-evident that you'll not need to go back. Um, and so that's kind of what's interesting about digital transformation, the fundamental change to work. It fundamentally changes how you, how your employees work, how you deliver on services and products to your customers, how you create value as an organization.
It's a fundamental change that once you kind of walk through that door, you never wanna go back. And one of the interesting things I think that we as finance people tend to kind of fall victim to is before we make a major capital decision, we're always asking ourselves, what's the roi? I, you know, what's the net present value?
What is the discount rate we have to use for this discounted cashflow? Is, does it cross a certain I r r? Yes, let's invest in it. That makes perfect sense for the kinds of investments where the return is very. Measurable on a, on a scale that makes sense. If I'm in, in investing in new machinery, it's gonna output this many new widgets and we can kind of work backwards to figure out what the, the present value of that is.
When you're doing a digital transformation, you're undergoing digital transformation, you're implementing an E R P or analytics or any kind of IT infrastructure. The return is so ridiculously high that the numbers almost don't make any sense. They're almost insane. Mm-hmm. We help one customer scale their customer base by 20 times just by moving them off as spreadsheets.
Mm-hmm. Um, another one, we help them improve their customer reporting cadence by 30 x. Like, the numbers are kind of ridiculous when you look at it and really only serve, uh, to, to reinforce what we already know. Once we've gone through the looking glasses, we don't want to go back. And that's kind of the way that I would sort of term digital transformation.
So my, to kind of wrap up this sort of long-winded answer, it's, it's a fundamental change to work that once you've done. It's so obvious that it was better than what you had before. That we don't ever want to go back and the, the thing that I find can actually impede your decision to undergo a digital transformation are the more classic measurements of roi, which is like, how much money are we gonna get back from making this investment?
Sometimes it's the results are so ridiculous, you never need to ask that question. Mm.
Going back to what you said there about the return that you've helped some customers achieve by moving them off spreadsheets. Yeah. Is that because they've got better access to data, so they've been able to spot maybe a gap that they're not selling into?
So there's two ways that data can help you scale revenue. Mm-hmm. One is what you just described, which is unlock new revenue opportunities where we may want to collect data from multiple sources. Analyze it and then make a decision based on that. And usually the decisions are, it's not always a clear decision, but sometimes we need to just run an experiment, test something and see if it works.
Sometimes it's for a B2C company, it's launching a new SKU new flavor, new offering, new discount, new promotion, whatever it may be. We need to, and we need to do this on a very quick, quick basis, whether the promotion is only gonna last a couple of weeks, or we need to launch a new skew in the next few months, test how it works, and then decide to cut it or expand or double down on it.
Mm-hmm. And so we need a fast, iterative approach. If you have too much data that's gonna fit into a spreadsheet, spreadsheet, if you can't manage it in Excel and a single human can't do it on their own, that's where the investment is made to look for, for bringing that system and that data together. So that's one way of doing it where it, it enables essentially faster decision making is the, uh, is the cliche, but it really does, it, it, it helps you make.
Run experiments and kind of make decisions on a faster cadence. So that's one kind of way of unlocking revenue opportunities. The other one is literally we, we provide data services or we provide reporting to our clients, or we do something where we're, we're generating a report or an insight or analysis on behalf of somebody else.
And if I'm doing that in Excel, and I'm primarily a consulting shop, if I can't scale beyond what I currently have, because I'm doing it all in spreadsheets, I'm, I'm leaving money on the table. I can't grow my business because it's harder to grow headcount depending on the industry that you're in. It can be harder to grow headcount than it is to just implement automation software.
It's, it's fascinating. So, so on your second point though, I've just drill drilled down a little bit more. So yeah, there, there are some processes that can't scale. Everybody knows, you know, that we've, we've created a bottleneck for ourselves. Yeah. Um, I'd say. Sometimes it's easy to make a, a decision to, to transform.
Um, obviously we always need to be wary of the people that are protecting their bit. Right. You know, so, um, I still want to be able to do my process because I'm comfortable doing it. You know, we, we've had those conversations a a ton of times, but in just thinking in those terms, um, and I've, I touched on this in the, in the last call, um, will Barron, by the way, um, heads up, uh, selling my simple, um, he, he built up his reputation with salesman.com podcast.
Really good. Um, but we were talking about the, the 80 20 principle. Um, and everybody knows it, but in my experience, myself included, people generally aren't very good at implementing it because again, we get so swept up in the weeds, you know, just to do a simple exercise to say, let's write out all my tasks, and then let's decide what the highest value versus the lower value tasks are.
Right. I think that feeds into the point that you're saying there is that, you know, if we can map that, you know, revenue making process to the point where we can cut out 80% of the tar, 80% of that activity that isn't actually producing the results that we want mm-hmm. We can, we can then, then scale quicker.
Right. So my, my recommendation if you can, is try and go through an 80 20 exercise, you know, as regularly as time permits. Right. Whether it's, you know, yeah, quarter whatever, and, and just do that, you know, whether it's revenue operations, whether it's, you know, supply negotiations or, or anything like that.
Have a think about, you know, Where is our revenue coming from? Because 80 20 dis dictates that, you know, it is probably 20% of your customers that are responsible for 80% of your revenue. Right. It's, it's also probably 20% of your customers that are responsible for 80% of your headache. Right. So, correct.
I just wanted to stress that. Cause it's something I've been, I dunno whether you've, you found that, um, something I've been thinking about
a lot, literally every company on the planet will agree with that statement. Yes, it is, it is, it is a rule that has stood the test of time. Absolutely. Yeah.
Just going a level deeper though, um, and I, and I've been thinking about this, this recently.
So have you heard the example of the, uh, the two and half thousand dollars espresso machine? I have not actually. You, you've probably heard it in different forms. Right. Uh, and again, it, it relates to 20 80 20. So, um, why would a coffee shop in this instance Starbucks, um, sell an espresso machine for two and a half grand when the majority of their customers pay?
Two or $3 for a coffee, probably not $2 anymore. It's probably like three, three, $3 50 right now. I, i, it's relevant, right? But when, when you look at 80 20, um, and I'll see whether I can share the notes, um, in the, sorry. The tool in the show notes, it's called the 80 20 power curve. And, and, and essentially you can, you can plug in your data to the point where you can say, um, this is my number of, you know, total customers or total addressable market, whatever it is, here's my average product sell value or my average deal value.
And then it will look at the average value across the market segment and it will then do an analysis of, right, well this is roughly how many customers will pay this amount for your service. Right? But then you also have an amount of customers that are willing to pay substantially more cause they love everything that you're about.
You know, and, and I've been thinking about this quite a lot because I think every business can be thinking. What is my two and a half thousand dollars espresso machine, you know? Mm-hmm. Mm-hmm. And you see it all the time with your, you know, your bronze, silver, and gold packages when, you know, whether it's somewhere or something like that, you know, the gold tends to be towards the top end of the power curve for the people that are willing to invest a lot more.
But coming back to revenue, I think ties to the first point you said about having the data is, you know, can we be more agile as much as, I hate the word agile, but there's, there's not really a more fitting one in this, this case, right? Right. Can we be more agile in carving out what that two and a half thousand dollars espresso machine is?
You know, what can we offer that we're not offering already? Because logic dictates there's gonna be at least one person that's, or one company that's gonna buy that. Right. You know, so I, yeah. I think it's interesting. I can't think of a supplier example.
Uh, I, I can, I can think of a great example that we saw, uh, that we're seeing, that we've seen happen, uh, firsthand, uh, in the.
Automotive industry, which is Tesla. Um, they came out with extremely high-end, very, uh, kind of like niche, uh, products, cars, uh, the Roadster, for example. Mm-hmm. Uh, which the vast majority of the population would never be able to afford. Um, but they released that first. There was a group of people who absolutely wanted those and wanted to drive those.
Um, and they did that again and again and again, eventually working their way, kind of down towards mass market. And now they've got the model three and they have like, everyone on his brother has one of these. So they started kind of at the higher end. And the reason why is because they knew that they could.
Make higher margins on that. They knew that the technology still needed to be developed and still needed to be refined, and essentially it was the folks who were willing to pay that premium for a, a higher level of service or a higher experience, or more refined experience or whatever it may be. They were, they were paying for something more, or even just to be one of the first few people to ever drive an electric roadster.
They were paying for that, which then funded the rest of the vision and allowed Tesla to then kind of move down the market towards, uh, you know, mass production, mass appeal is the idea.
Yeah, of course. You know, the example you're given there is it, it working the other way, right? Where you start with a high value product and then you determine what, what the lower end of that, that power curve is, right?
So, so what lower value products can we sell more of at higher volumes? But yeah, that can sometimes be tricky territory, depending on the nature of your solution because Right. You know, you might not have the resource to sell, you know, a, a, a load more volume of lower value products, right? So there's a decision piece that, that goes behind that.
So, uh, but I, I find it fascinating. You often find that, you know, once you've read 80 20 a couple of times and, and that sort of stuff, that you kind of see it everywhere, you know? Yeah. And, and, and I guess to, to, to bring this point home and, and sort of refocus it on the, on the finance piece is, The nature of finance is structured repetitive activities, um, hopefully reducing now that we have better tools, you know, so we're moving into that, that, you know, that, that value add piece.
But map your month end, you know, map the way that you build your reports and then try and prioritize those tasks in terms of, um, high and low. Cause I was, I was having a chat the other day. And it's surprising. So let's say that you've got management pack to get out, um, and then you've got a couple of senior members of the team or an investor or something like that said, um, on top of the management reports, can you just send me, um, this every week?
You know, can you send me this every week? Right. So I guarantee that if you were to rank that, it'd probably be a very low value task. Right. But you probably also find out if you, you know, and, and internally you're probably not gonna track email opens, right? You're not gonna do the confirmed delivery request or anything like that.
But I'd say nine times outta 10, a lot of the stuff that you're doing just doesn't get read or open. And that's, that's nothing against you, you know, you're doing a job, but I think sometimes we need to be hypercritical over the stuff that adds value on the stuff that that doesn't. Right. Again, applying that, that 80 20 mindset.
Right? Yeah. The last thing that I'll say as well, and, and again this, this is 80 22, uh, 80 20 as well, is, um, you can map your salary to 80 20. You know. Okay. Again, more, more relevant to, I guess, um, maybe more junior members of the finance team, uh, that are looking for progression and all that sort of stuff.
But you'll find that when you apply 80 20 to your salary, you'll find that some hours of your day might be worth $5 an hour. Mm-hmm. Some activities in your day might be $500 an hour. Yeah. So, in converting from low, medium, high priority to low value, medium value in terms of monetary terms can sometimes really help you switch towards, right.
Well, I'm, I'm literally working on $5 an hour tasks now. Yes. What can I do to move up the power curve, you know, to ensure that my employer is, you know, seeing a return in the sort of work that I do. So, uh, it works at a personal level as well as a business level is what what I'm saying there. And
to build on that, I think, uh, we know that the, the classic, um, profession which follows this 80 20 power curve, Almost to a t is sales, right?
Like the more effort you put into prospecting, having conversations, you know, opening doors where possible like that directly will translate into more revenue, more commissions, more money. These guys, it's a very clear, you know, one-to-one connection there. Um, and if you're spending a lot of your time on things that are low value, that don't directly relate to, you know, making a sale, then that tends to, you know, kind of put a drag or friction on, on your ability to, to increase your earnings.
There's a lot of professions out there where. The one-to-one connection isn't quite as clear. Mm-hmm. And that's very, very true of back office, um, kind of back office tasks. I'm thinking it HR quite often finance in those cases where, you know, just because you've streamlined some of your own internal process, uh, doesn't necessarily mean you're gonna get a promotion or can I get a raise?
It doesn't, we'd like to think that it always works out that way, but that's not always the case. Um, but to tie into what you just said, there is a personal aspect of it. Um, if you can focus on, Hey, where am I spending? What of my tasks is causing me 80% of my stress in my day-to-day as an AP clerk, as a controller, as a, you know, a finance director or finance manager, like, what's causing me the most stress?
That is causing me to either butt heads with my CFO or butt heads with the rest of the board. Uh, if I am the cfo, what's causing me to work late on Friday nights or, you know, kind of have to be working Saturdays or whatever it may be. Where can I streamline my processes in the event? It, it ultimately it is going to benefit the company.
It, it a hundred percent will, and that's a great way to kind of make a case come, you know, yearly, annual or review time. It's great. It's a great way to kind of, you know, sort of, uh, provide justification for that raise or that promotion or whatever it may be. But certainly on a more short term kind of personal level, that's gonna help you get outta the office at 5:00 PM and kind of have an evening to yourself where you're not frustrated by it.
Especially if you're at a very high growth, uh, company that's experiencing a lot of change and a lot of stress that your 80 20 can be easily applied there to not only help the company, but also really help yourself.
Hmm. And there's, there's a decision step there as well. Once you've done that mapping on mm-hmm.
How you free up that time. Cause you, you might know where your low value activities are, but they might still need to be done. And, and That's right. So, so my advice to anybody listening is that firstly you separate your individual activities mm-hmm. To business activities. So, um, an individual activity could be something like, um, typing up meeting notes, for example.
Right. And then a business activity could be something like producing management reports, right? So, so segment that first, because if you go to, you know, whoever it is that's gonna approve tech or, you know, whatever it is that you're suggesting, it might be easier to start with freeing up time with the lower risk stuff before you start looking at the, the higher risk stuff.
So for, for me personally, um, I will look at the quicker wins at the lower end of the spectrum that then enable me to spend more time freeing up more time, essentially. So I'm, I'm just trying to keep that flywheel going. So for me, stuff like automating me automated meeting notes is game changer, right?
Very low risk, you know, providing, obviously the platform that you are using is, is data compliant and it's not tracking any sensitive data and all that sort of stuff. Um, and I dunno how you think about this aji, but I'm now thinking in terms of previously coming back to books like the four hour work week, you know, Tim Ferris and all those sorts of works where he spoke a lot about outsourcing to, you know, India, the Philippines, wherever the hourly rates are are lower.
That is still an option. Maybe you don't do that with, with business activities, but definitely with some of the personal stuff, if it's nonsensitive data. But we've now got a level now where, um, and I don't think I'm gonna trademark the term cause I don't think it's that good, but AI sourcing. Okay. I like it
I like that. Maybe you should trademark that. It's pretty good. Maybe
trademark. Maybe I'll pause on releasing the podcast until I've got the You
heard it? You heard it here first.
I like it. Yeah. But in theory, we've now got two options when it comes to freeing up our time, you know, because AI is now getting to the point where it kind of acts like a person, you know?
So the meeting notes is a good example, right? That's AI sourcing whereby it's doing what somebody sitting in a meeting typing up minute notes would be doing, right? Mm-hmm. Likewise, um, when it comes to process mapping, you know, when it comes to, you know, HR and onboarding, all of that sort of documentation stuff, we've now got AI that do a lot of the heavy lifting for the human to then do the decision piece after that.
So I'd say that they could be quick wins depending on the tool before you start reinventing the wheel with more complex technologies to, to free up mind, because. Outsourcing to people is never gonna go away. Outsourcing to AI is gonna become more and more relevant. And digital transformation is, as you've spoken about, AGI is always going to be on the agenda if you are a business that's scaling up an undergoing change.
So, yeah, I don't know whether you think in a
similar way. No, I, I agree. And I, I remember watching, uh, an, an interview with Jeff Bezos where questions were being asked of him about ai, and, and he essentially said that, and I think we'll all agree to this, is that AI is a horizontally enabling, uh, te technology.
It, it, it enables, uh, tasks and industries and technologies across the board in the same way that the internet enabled technologies and, uh, and, and businesses across kind of like a horizontally across industries. So, um, this kind of ties into a, an earlier conversation around, you know, AI doesn't have to be scary.
I don't think it's going to be, we're not releasing Skynet quite just yet. But I think that, I think that it's, it's going to help us offload or AI source a number of tasks that historically have been too difficult to, to do or to complete or just time, maybe not difficult, but just time consuming. Mm. And that does allow us to kind of focus on things that I, I think, will always have value and will never go away, which is human to human interaction.
Human to human connection. You know, I'm sure you get like tons of, uh, AI generated or automated sales emails that all kind of sound the same and, and end up in your spam folder. What's so funny is that there was some AI now that's writing these sales emails and it's your inboxes AI that's filtering them out.
So it's kind of AI to AI is like are competing with each other, but something that is like, we always want to. Genuinely connect with other humans, that's never going to go away. This is why, you know, a, a genuinely well-written, you know, honest email will still generate results or a video kind of, uh, letter to somebody will still, you know, form a connection.
Or the best example is going out and seeing people in events or going to, participating in a community of some kind or going to a trade show like human to human is never gonna go away. I think ai, if anything, AI sourcing will allow us to continue to focus on that human to human level interaction. Yeah,
Because if I, if it's quicker for me, Um, to do some analysis, you know, towards a certain business goal, um, and get people in a room so that all of that discussion is extremely relevant, right? That beats hands down, you know, three or four different emails that are all project updates, or do you see what I mean?
You know, if, if you can shift your work coming back to the 80 20 so that we're at the top end of the power curve, so that we know we're adding, you know, high value stuff to, to the business, we can then focus on that, that human-centric stuff, which is the, what's the next move. Right. Um, and that's, you know, what you call it, co-pilot AI sourcing work that happens to be, it should be an assistant, um, rather than something to be feared, I guess.
So that's, but just on, just on that topic. Um, and I'm curious, um, for you guys at, at types if, so obviously the, the data q and a has been built into the platform for, for a while, right? So, mm-hmm. Um, and you don't have to give any secrets away, but in terms of timeline and evolution of, of the system, you know, are you gonna be sort of, um, building AI context and sort of next level into it?
Is, is that a consideration for you guys?
So where we are heading with our product roadmap is we're kind of taking what I like to call the Apple approach to, uh, to building products and technology. So the, the thing about, the thing about, um, the thing about my industry, especially with you're talking about data and analytics, is that a lot of people focus on it as like, Something that's going to enable an as yet unknown future.
We're gonna run some kind of AI machine learning model against our data, and it's gonna spit out some insight that we had no idea about that humans could never have come up with on their their own. And that's a very exciting, futuristic, sort of like science fictiony feel to it. Our approach and our look at data is a lot more pragmatic and a lot more utilitarian.
Um, we think of data as. Utility, like electricity or running water, like you want water to come outta your tap when you turn it on, right? That's, that's the way that we look at data, that it should be a fundamental utility and right and enabler for every company. And so our focus is continuing to push on the, the utility of our products.
And to tie it back to what I said earlier about taking kind of the, the Apple approach here is that if we want to weave in AI in some way into the product, it will be to lower the barrier between the user and their end goal. And so we wanna make it as we want, we want the, the screen of technology to be as transparent as possible so that you don't realize you're using a tool, whether it's a piece of hardware or piece of software.
You don't realize that you're interacting with technology. You just realize you're interacting with. The end goal, the work product or the, the outcome that you're trying to achieve. And so that's, that's our goal and that's kind of what informs our product roadmap. Um, rather than trying to commit to a specific, you know, generative AI model that's going to do this or that with your data, it's like, how do we bring that into the product so that if we do our job right, you won't even realize that there was something there that was making it easier for you.
Yeah, yeah. A hundred, a hundred percent. Yeah. And, and I think that's, that's what you've definitely got right, is Yeah. It it should be totally natural, right? Yeah. So, no, agreed. Okay, so question I always ask, um, and we've spoken about bi, we've spoken about load of tools on the show, right? But, um, and it could be personal, professional, you know.
Yeah. We're talking phone apps, we're talking Google Chrome extensions, we're talking, you know, all of that sort of stuff. If you've got an app or gadget, we had, uh, we had, uh, a Phillips one blade Shaver from Tamer when he on the podcast a few, few days ago. But anything like that, that you couldn't live without that you wanna share with the listeners?
Yeah, so, um, you know, they're not paying me to plug them, but, uh, I, I, if there's, if there is one app or tool that I will always, uh, kind of promote, it's a, it's a tool called Notion. Yeah. Um, which, uh, I think by now, um, most folks have heard of it. It's an all-in-one workspace. But what's very, what I really admire about them is that, They've made it.
So it is exactly that example that I had just given of like, you don't really realize you're using a tool, right? It, it's, it's, it's there to get you as close to the end result as possible. And so internally we use it for standard operating procedures, employee handbooks, you know, sales playbooks, anything that is a, a document that is living and breathing that will change and evolve over time.
And they've just figured out a way of doing this. Um, that is, Better than say Google Sheets or, you know, word documents floating around or whatever it is. They, they've just, they've figured this piece out and they've created a really nice product. We use it at work. I use it in my personal life to manage things.
Um, it's, it's absolutely fantastic.
I'm gonna give a plug, I'm just gonna quickly search for it here now that we're talking about Notion. Cause I use Notion as well. Everybody knows very well, but I, I use Notion as my second brain, right? But there's a guy called Thomas Frank, so it's thomas j frank.com and he's a bit of a notion guru.
Um, so what he does is notion templates. Oh, yeah. Heavily focused on productivity. Mm-hmm. So he does, um, his ultimate task and project management template, which I've got. So that's what I use for my ta, you know, my can ban, you know, low, medium, high, high priority, but then categorized by project, so I can see project progress, right.
Uh, he then does an ultimate brain, which is a chargeable template, but it's a lot better than that one. But anyway, the, the reason I mention it is because to come back to the co concept of AI sourcing and outsourcing, I work with a, with a virtual assistant, you know, and so he helps with the, you know, the softer stuff, the non, the nonsensitive stuff, you know, LinkedIn carousels, um, uh, writing basically copy and pasting texts from podcast transcripts into my website, you know, that, that sort of thing.
Right. Um, and it's all the shared notion, um, task. Yeah. So, so what I do is I use the, the Thomas J. Frank Ultimate Task Framework, and I've added an extra step. So there's. High, medium, low, but then there's also delegate tasks. Okay. So the tasks to go to the va, go to the delegate section, and then I share the note with him via teams, and then he communicates via notion.
So it's, yeah. Yeah. So, so full, full advocate of notion. Yeah. You've got a whole
workflow set up around it. It's amazing that you, the things you can. You can do with it. I mean, we don't, we, we mainly use it for SOPs and these kinds of things. Um, cuz the alternative would be setting up an internal wiki, which would be a huge pain in the butt to, to have to do this.
solves that issue for us. Yeah. Yeah. That's it. And, and you know, we use Microsoft Teams, I've got experience with Slack and all that sort of thing. But coming back to what you say about that ease of use and that just using it without really knowing that you are using it. You can have wikis in teams and, and all of that sort of stuff, but it's just, just doesn't feel like it flows as much.
No. You know what I mean? And we, we could talk about notion forever. So we'll may maybe put, put a stop on it there, but we need a different, we need a whole other episode just to talk about notion. Yeah, absolutely. Maybe we will, maybe we will one day and then we'll definitely make sure that notion givers a bit of cache.
Exactly. Alright, so where can people find out more about you?
Um, honestly the best place to find me is on LinkedIn. Um, so it's, it's Ajit Star Car cfa, uh, on LinkedIn, just kinda social media. The company's called Type Sift Inc. Um, and that's probably the best place to, to connect
and, and to chat with me.
Perfect. So I'll, I'll put those in the show notes as well, but No, that's excellent. Sorry we've overrun. It's, uh, it is been a good chat. Really appreciate you coming on. I've, I've loved it. It's been really good.
Yeah, this has been a lot of fun. So I appreciate, uh, you having me on. No
worries. All right, you cheers.
Alright, take care. See ya.