#253 The Infrastructure Supporting the Data Revolution with Saad Siddiqui, General Partner at Titanium Ventures

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

In some ways, you do need to do the job today, but I think if you don't think about what's going to happen the next five years from now, you're going to be making platform shifts every couple of years that are really costly and just slows down the organization in a massive way.

Welcome to Data Framed. This is Richie. The data space is moving fast enough that you need to continually improve your capabilities. That is, you need an ongoing data transformation program to stay relevant. Any data transformation program involves changes to technology talent and processes, and today we're going to focus on the technology side, from core infrastructure upwards. There are many tools available promising to improve your data performance. So the key questions are how do you decide what to buy and how can you implement them in a way that has an impact on your customer experience or your business financials? It can also be tricky to know which data use cases to target for an overhaul, whether you want to automate your existing processes or chase that moonshot. Our guest is Saad Siddiqui, a general partner at Telstra Ventures. Over the last decade, he's managed a $100 million portfolio of data infrastructure investments with a focus on tooling for business operations. He sits on the board of a dozen startups and advises many more. Previously, Saad worked in mergers and acquisitions at RBC Capital Markets and was assistant vice President at Merrill Lynch's Global Wealth Management Division. And he was awarded the President's Volunteer Service Award for his fundraising work with Blue Engine, the Pakistani Rehabilitation for the Disabled foundation, and other charities. So let's hear his advice on how to transform your data infrastructure. Hi Saad. Welcome to the show.

Thanks, Richie. It's good to be here.

Cool. So, just to begin with, what does data infrastructure mean?

So, data infrastructure generally means collecting a lot of data and getting insights and actions out of that data that you're collecting. And during that entire process around collecting all the data and getting insights in an action plan to act on that data, what are the different components that you need to make that data as useful as possible. And this can cover going a little bit deeper. Everything from how the data is transformed, how you secure that data, who has control around changing that data and trying to understand that change over time and then finding unique ways to get insights that can impact a business outcome or a life outcome that you may be working on.

Okay, that's kind of interesting that I guess in my mind I sort of thought, well, infrastructure means like all that really low level hardware stuff like, oh, something's happening in cloud, but actually there's a lot more layers of infrastructure goes right up to the software stack as well. So there are, there's a lot more depth to this. And I like that you talked about business impact. So maybe can you just talk me through like what business problems are you trying to solve when you are going about improving your data infrastructure?

There's a bunch of different use cases that we're kind of looking at. Right. So we're seeing a lot of different use cases around the finance organization, around product and marketing. That's always been an area where data has had a massive impact. We're seeing a lot of impact around the sales and business operations side of things as well. So across functional units and also across different verticals that are getting impacted by the more recent trends in data. So it kind of, it touches every single organization within an enterprise and also different sectors as this ecosystem is evolving.

That's kind of cool that it's basically every different sort of a team. Like you mentioned marketing, you mentioned sales, you mentioned like whatever industry you're in, then more data, better data infrastructure is going to be beneficial. I'd like to talk a bit about how you get started. So suppose your CEO says we need better data infrastructure. What do you do first?

I think it is really important to get a first understanding of where everything is sort of sits today. Right. That is one of the more important things because there's massive data sprawl within the organization. There's a lot of shadow IT people have, are using some of the newer technologies like OpenAI and trying to understand like, okay, where does all the data sit today and how are people using it and what are the outcomes that the CEO and the management team is sort of looking for. Then you kind of go down the list and trying to figure out like, okay, like what are the platforms that the enterprise needs to bet on that can allow for those outcomes that can be built for scale. So, so as the organization scales, they have better Sort of understanding around the data usage, not just today, but five years from now, 10 years is hard because the data doesn't scale linearly, it scales exponentially and sometimes logarithmically. So I think just kind of like having an eye around scaling is really critical and then securing that data set. Right. So just making sure that the core IP that is within that enterprise and more and more, in my opinion, data is becoming the core IP for every enterprise across the entire enterprise sector. Just kind of making sure that the crown jewelers are protected and people know like who has access to what and kind of protecting against those breaches around either data getting leaked. And there's a bunch of issues that we've sort of seen in more recent years around that. And data, like a bad data getting infused into the data sets that can corrupt outcomes. And some of the language learning model stuff that we're going to talk about.

Today, that sounds slightly terrifying, like the idea that, well, okay, maybe your data is going to grow exponentially over 10 years. Do you need to plan for that when you start building stuff like have you got to try and make a prediction about what's going to happen in 10 years or can you work with a shorter timescale to get started?

I think it is important to making sure that you can deliver on those outcomes. But you almost have to think through your data is going to grow pretty substantially, especially now more than ever before. All the data that was generated from the beginning of time till 2020, we're generating in a year now. It is quite tremendous at how quickly we're collecting data across the enterprise across functional units. And it's not just the IT side where they're getting pinged around alerts around application performance. We're also seeing that sort of happening on the security side where everyone's getting alerted on all these different vulnerabilities that they have. It's also happening on business performance as well. Right. So every time someone clicks on a button and how it interacts with an application that informs product decisions and collecting all of that data and those interactions are going to increase as the UI sort of changes over time. What are the kind of questions that people are asking? What are the kinds of answers we're getting? How long is someone spending time with a video that is surfaced and that sort of data that is being stored now more than ever before? You almost have to have an eye for scale and have an eye for what are the key components that you need to sort of like extract from that data. So in some ways you do need to do the job today. But I think if you don't think about what's going to happen the next five years from now, you're going to be making platform shifts every couple of years that are really costly and just slows down the organization in a massive way.

Okay, so maybe you don't need to be able to predict exactly how much data you're going to have in 10 years, but you want to be aware that there's going to be some sort of limit to how far you can scale with whatever technology you choose. And you want to be mindful of when that's going to happen. So you're not to going, okay, so can you talk me through who needs to be involved in this? So if you've got a data infrastructure project, what sort of teams or roles are going to be involved?

I think in general, there's a bunch of teams that are now being involved that historically haven't been involved in the past. So traditionally there's generally the CIA team, the chief Information office, like that sort of like controls all the data that is typically involved. Right now, more and more we're seeing a lot of different business units having their own internal teams around data as well. So the CFO may have data engineers on his or her team, the product teams have their own data teams. And so you almost as a central organization on the cio, their job has gone from let's sort of control all the data, is to empower all the business units to be able to get insights out of the centralized data set. So that is becoming more clear. And that sort of goes from product engineering to marketing to sales across the board. In addition to that, we're also seeing the CISO is more involved now more than ever before, the security teams. And that's because they need to have a better insight around who has access to these data sets. Like, understanding which accounts are becoming vulnerable in terms of, like, who are the most critical people that have access to the most amount of data or the most critical data, that becomes incredibly important as well. So the CISOs are becoming more involved in decisioning as well. It doesn't matter in some ways if you can sort of have the best data architecture. If the security team doesn't bless, becomes really hard to make a bet on a bad platform.

So actually, you mentioned quite a lot of the C suite there. So you mentioned the Chief Information officer is perhaps responsible for the IT side of things. You got chief Information Security Officer. It's going to be responsible for, I guess, data security and Then you mentioned chief financial officers kind of want to weigh in on the cost of things and then the product managers. And then that's before we even get to like chief Data and Analytics officer, chief AI officer, maybe even chief Technology Officer. So it feels like a lot of the C suite could be involved in here and that could get tricky. So can you talk me through like who needs to be accountable for which bits of this then and how they go about? Well, how you stop a C suite fight?

Yeah, 100%. I think that's sort of something that we're sort of seeing across the board where honestly, enterprises are trying to figure this out because to your point, there's a lot of cooks in the kitchen, so making a bet on a specific platform that suffices everyone's needs becomes really, really hard. In my mind, the most important sort of factor is like, and this is based off some of the conversations you've had with some of the guests in the past, which is where we've kind of gone from a centralized CIO Org in the world of business objects and some of the old legacy BI tools to the complete ownership of data across the enterprise with companies like Tableau that sort of had a lot of data across every business unit. And in some ways it's coming back where centralized control is coming back because of a lot of the high profile security events have had over the last couple of years and people need more control. So in some ways the security and the data platform decisions are still being made by the CIO and CISO and over time, some of the end applications, the decisions can be made by the functional units. So the product teams may leverage the data set that the centralized teams have, but they may decide a different BI and analytics tool compared to everyone else. And some of the other teams may not be involved in those decisions. So as long as an enterprise has a really good understanding of the data assets that they have, that is one of the most critical things. And on top of that, the teams can sort of enable self service uses to the other teams that need access to it.

That just seemed like an incredibly important point worth mentioning, that it's a very good idea to understand what data assets you have before you start doing anything, because otherwise everything's going to be a mess. Okay, so you talked about business impact and it's important for data to have impact on the business now it feels like the further you get away from commercial teams and sort of the further down the technology stack you go, the further away you are from these obvious business metrics. For success. So how might you go about pitching a change to data infrastructure when it's not just like, hey, here's a direct impact on revenue or costs?

I think that's a very important point. I think as an organization grows, people just get hired to do a specific job within IT team or software engineering teams, and they don't understand necessarily the impact of the actions that they're taking. So I think it is really critical before any sort of decision is made around the company objectives, it is important that the entire team understands why they're taking the steps that they are. Right. So for example, if you're moving off of a legacy platform onto Snowflake or databricks, why is that an important thing? Is it on a cost saving side? Is it speed of execution? Is it getting better insights out of the product? What is the reasoning behind that platform shift? That's not an easy thing to do. So I think what we're seeing is the best organizations are actually connecting business outcomes and communicating with the entire team. Why those steps are really important, because some of these decisions aren't easy. Right? When you sort of migrate an entire data architecture from a legacy sort of platform into a newer platform, it takes months, sometimes even years to do, especially for some of the largest enterprises. So I think it is really important to understand for the entire team, why are they making the steps that they are and connecting them to those specific business outcomes. And generally the best business outcomes that people get behind are around cost savings and revenue drivers. So they need to sort of see like, hey, we're taking these steps for X, Y and Z reasons.

Okay, that does seem really important. That really can tie back any decisions to some kind of business metric, even if it's slightly more abstract than just a purely monetary thing. I guess one thing I've seen is like, anytime there's a technology change, it's going to break some of my workflows. And then I'm sort of slightly cursing whoever made the decision to do this, even though I know there's probably a good reason for it. Can you talk me through how you go about bringing people on board with these technology changes? How do you pitch it to the rest of the organization?

The most important pitch ends up being like, this kind of goes back to the initial point. Okay, if your CEO comes to you and asks you like, hey, make this data infrastructure better. Tying back to that sort of initial conversation we had, now that we have a better understanding of all the different data assets that we have, all the strengths and weaknesses in our platform, it could be speed, it could be cost, it could be a lot of different things and the changes that are being made as part of it. Right. I think sometimes engineering teams tend to kind of go down the rabbit hole of like, hey, this is the super cool new technology and we need to sort of migrate off of everything to let's say an iceberg based architecture. So if the CEO and the engineering team made a shift a bit ago, like maybe two, three years ago, onto a GCP or Azure based architecture, they're just going to question everything that you're doing. You're like, we just made the shift a couple of years ago, why do we need to do this again? So unless it is clear that there's a meaningful ROI around performance driving revenue, around cost savings, it becomes really hard to get some of these decisions across. And in my mind that's sort of like one of the most important things. Also, like to our conversation earlier, it is really critical to think about scale over time because data infrastructure moves pretty quickly and I suspect with the new language learning model architecture that's coming up, it's going to move even faster.

You mentioned the idea of like moving from say a legacy data warehouse to databricks or Snowflake or one of these kind of modern cloud platforms. Would you quite often just changing the data warehouse isn't enough. You might want to say, okay, let's change our BI tools, let's change some of the other infrastructure around this. Is it better to try and change everything in your technology stack at once or would you recommend doing one piece at a time?

I think it really depends on the organization itself. It depends on where the pain is being felt. Moving to a different cloud platform allows you to scale much easily than some of the legacy systems. If the pain is being felt around scaling, then that's what you need to do. And seeing how the BI tool is reacting to the new influx of data volume, if it doesn't scale, then you kind of need to sort of make a decision around how do we change the interface by which the data is being absorbed by the organization. So I think, to be honest, I'm not giving you an answer here, but I think it really depends on the organization itself and what are the key priorities that they have and what, where can they get the biggest ROI around the data infrastructure.

Okay, that was artfully avoiding the answering question. I like that. Yeah, it depends on the organization, so I certainly can see how. Yeah, you probably want to think about what your goals and then pick the technology based on your goals rather than just making a decision outright. So I'd like to talk a little bit about timescales. So you mentioned the idea. Okay. If you just change your data house two years ago, you're probably not going to find much favor if you suggest changing it again. What sort of timescales are you looking at for overhauling data infrastructure and how often do you want to change things?

I would have told you it is probably every three to five years. But as language learning models have become the number one priority for every organization, if you made a decision two years ago or a year ago, and your infrastructure doesn't allow for the utilization of some of these newer technologies, and you probably need to move faster and make sure that your organization can leverage some of these newer technologies because your competition is not slowing down because you made a decision just a year ago, you almost sort of have to have an eye on the market and seeing what's happening in the broader ecosystem and make a decision that way. So in my mind, I think it is really critical to, from a timescale perspective, marrying that with the shifts that are sort of happening in the market, because you can kind of make a shift in some ways now. And then if that doesn't adhere to some of the innovation that's coming out, your competition is going to catch up to you much faster and may even start beating you in some ways.

Okay, yeah, so that just seemed like a tricky thing. So suddenly generative AI is everywhere and your infrastructure is not keeping up. How can you keep your organization or your infrastructure flexible enough to deal with changes in market conditions?

I think in some ways you almost have to reassess. This is such a big trend. This doesn't happen every day. This happens once every 20 years. So I think if your architecture today is not conducive for leveraging some of these language learning models, you have to sort of rethink your entire strategy. And I personally think that language learning models in general, if you kind of think about it like salespeople, will be able to qualify sales needs significantly better than before. So what is the value of that? You're actually narrowing in on a specific customer that has a need that you can sell to today, versus canvassing 100 customers in the past. In the case of customer success, we're seeing a lot of companies that are looking at changing customer success, reducing costs in call centers, and getting better results out of the call centers than they've ever done before. So in some ways, you almost need to think about like, okay, this magnitude will shift. That's sort of happening. If you don't kind of get on now, you are missing the belt in a meaningful way. So how do you sort of like take advantage or make sure that you're sort of ahead of these sort of shifts? I think it's honestly tied to continuous education and just kind of making sure that you're not falling behind compared to your competition. Because this is one of those things that like happens once every 20 years or so.

So that's good that you're not going to have to react to some crazy change in the market conditions every quarter or something. It's like it's a bit less frequent than that. Do you have any tips around project management or processes that are going to enable you to be more reactive?

I think one of the most important things in my mind is kind of making sure that you're collecting as much data as possible, but not necessarily like noisy data. The data quality piece of it is going to be really critical. And collecting the best quality of data and then tying it to language learning models is going to become really important in the next generation as well. I think it's going to start informing your language learning models in a more meaningful way and it is also going to make sure that you're able to service your customers significantly better than before. So data quality becomes really critical. Trying to understand what is the data set that you're actually collecting to train these language learning models and tuning those models to the next generation of problems that are sort of coming up is going to become really, really important as well in the next generation for the next couple of years.

Related to that, are there any skills that you think you ought to have in house in order to be able to work with better data infrastructure or create better data infrastructure?

I think I kind of go back to the continuous education side of things. Right. So I think it is really important to kind of be on top in it. In the past, you would say sort of assume that if you had a strong data team within an organization, that should be good enough. In my opinion, that's not good enough anymore. You're having impact of data is being felt throughout the entire organization. And so you almost have to have your marketing team, your sales teams, our sales operations teams, your product teams or product operations teams get exposed to the next generation technologies in some ways so that they're not very siloed. They can sort of see the impact of the new innovation that's coming up. But also how can they leverage data in a more meaningful way to impact their functions? Basically, okay, that's wonderful.

I love that you sort of mentioned continuing education. There's a very dear to our hearts at datacamp and the idea that sales marketing people like all these commercial teams, they ought to have some exposure to data and all these sort of technologies around them. Do you have any examples of this, like where you've seen commercial people have that level of data literacy and they've had some sort of benefit?

Yeah, we're seeing that actually in a bunch of our portfolio companies. A lot of our portfolio companies are sending their. They have educational programs for their product teams and their finance organizations as well, where they're making sure that they're across some of these newer innovations. And what we're also seeing is there are a lot of newer applications that are coming up that are reinventing some of these categories. We're seeing organizations that are coming up in FP&A. We're seeing sales and marketing has a ton of companies that are coming up. So the education is also in some ways happening by where the vendors and startups are sort of going. So in some ways you almost have to make sure that you're across the newer sort of technology companies are coming up. You don't need to make a buying decision now, to be honest, to be honest, that might actually help people like me. But I think folks need to be across some of these newer technologies either in an educational fashion or be exposed to the startup ecosystem. And even if you think a startup tool is a toy, it is really important to understand why the founders have conviction around this idea. Why do they feel that it's going to have an impact in the use case that they're sort of focused on.

I'd like to know a bit about some trends that are happening within data infrastructure. So just at a high level, what do you think are the most important trends at the moment?

I think there's a few trends that are kind of really interesting to us here. The first one is around data security. I think as more and more organizations are leveraging language learning models, we're seeing that data security is becoming really critical for that. So people are exposing really confidential data to companies like ChatGPT and having control over that within an organization is really critical. So we're seeing a lot of startup activity in that ecosystem around making sure that your private sense of data isn't exposed to a third party and where you may not have control. The second thing that we're seeing is data quality and transformation and some of those capabilities. The middleware of data is becoming incredibly critical because now that you have your data secured. We're seeing a lot of issues around hallucination. People making decisions on outcomes that may not be correct as the models are hallucinating. One of the biggest problems around that is around data transformation and observability and making sure that everything that you're sort of collecting and training your models on is as close to real as it possibly can be. And the models aren't being trained on a data set that may change the outcomes that you're sort of looking for. And in some ways this ties around data quality and lineage and security is access control. So we're also seeing a lot of startup activities sort of like happening around, like who has the right access to, to the data sets that are critical to the company. And as you're changing those pipelines, it becomes really important to kind of making sure that there's no file data that is entered into that data set. So yeah, I think that's sort of like some of the more important trends that we're sort of seeing in the broader data ecosystem.

Okay, so I guess the three trends really there. So bigger focus on data security, bigger focus on data manipulation and transformation, and then also around data quality. And I suppose governance is the broader idea around that.

Exactly.

Interesting. So of those, which would you start with? Like, if you want to try and improve your data infrastructure, which of those three things would you care about first?

I think it depends on the outcomes that you're looking for. In some ways I think it would be naive to believe that your employees aren't using some of these technologies. So I would probably start with the security piece of things and then also trying to make sure that once you sort of figure that out, then you move on to data lineage and quality. I think all of these are critically important. The quality and lineage impacts outcomes in a more meaningful way and data security makes sure that you're not on the front page of the Wall Street Journal because there was a leak because of an employee mishap.

All those three things are actually, they're slightly removed from those sort of business impact things. So I'm just wondering, are there any ways to change your data infrastructure that's going to have a direct impact on the customer experience?

I think there's a bunch of really interesting companies that are coming out that have an impact on specific customer related use cases. Right. So around product analytics, there's a bunch of companies coming out that are, are going to have an impact around sales and customer success as well. So if you can build a product and this is maybe for the startup ecosystem here, if you can kind of build a product that leverages language learning models and you're able to collect a data set around a specific use case. Right. Let's say sales performance and you can kind of use that to forecast your sales in a more meaningful way. There's companies like Gong and some of the more established guys that are basically using some of these technologies to inform their users on what's the best way to deliver outcomes and what's the best way to deliver, like compete against their competitors. Something like that is really critical. So yeah, I think we're seeing a lot of like really interesting technologies coming up for specific use cases that sort of touch different customers as well.

I really like the idea of just basically using data in order to compete better. And so that's a kind of, I guess, offensive use of data. And the flip side to that is defensive use, which is about dealing with, well, complying with regulations. And yeah, I think particularly in a lot of regulated industries. So we're talking about finance, we're talking about healthcare. There are so many regulations you have to deal with around data. Can you talk me through how better data infrastructure can help there?

What we're sort of seeing is like there's a bunch of companies that came up almost a decade ago around process automation. These are companies like UiPath, Automation Anywhere, a bunch of these RPA tools. That's an example of a use case. And a lot of their early customers were customers in the insurance space and they were looking to improve their operations, whereby a customer, for example, uploads their documents for an insurance claim and, and kind of going from initial information to claim reimbursement. There's all these different pieces where an organization is collecting data and making decisions on that data. Stuff like that gets supercharged with some of these language learning models. Right. So now you could kind of figure out exactly what part of the organization or what part of that process is the big bottleneck and what is the, of the process that costs the organization the most amount of money. And you can build a plan, you get visibility on the entire process and you can now build a plan around how to mitigate some of those bottlenecks and cost overruns in specific parts of the process as well. So that's an example of how AI is impacting, let's say, the financial services industry. We're seeing similar things sort of happen in healthcare. We're seeing similar things happening in a bunch of different verticals even in construction technology. Right. So in construction tech, we've made an investment in that space where we're seeing, let's say you have an H vac outage. How do you understand what were the things that would have done to the H vac before this? Now you can actually take a picture of that H vac machine. You can actually get insights around, like what are the last maintenance things that were done. You can understand exactly what are the things to do to fix that H vac machine or pipes are incoming. You can basically get a better understanding in real time using computer vision and AI. And then it's connected to the backend around what is the cost of these operations. So you need this sort of pipe, this sort of cog that you need to fix that H vac machine or some of these pipes and get insights, get an outcome significantly faster than you would have in the past where you wait for someone to come for over a week and then they tell you like, hey, we're going to be back next week to try to fix this thing. You're able to see some of these impacts happen in a much faster way.

Okay.

I like the idea that you're getting that impact on the customer just by being able to diagnose problems quicker and then avoid some of the admin work of trying to get stuff fixed. So yeah, that's going to speed things up and make everyone happy. So on that note, it sounds like a lot of the use case here is around automating things. So have you seen any other examples where automation has been possible through better infrastructure?

Are there areas that you'd like to kind of COVID Because in some ways, like everything is being reinvented.

Okay.

I love that.

It's just everything is being automated. That's kind of brilliant.

Okay. Yeah.

I don't know whether there's any specific examples you've seen where it's just like there's been a clunky process that you've seen automated.

The area that people are sort of going to the most is around the RPA use case. And I'm happy to describe it again, Rich, if you like, maybe just talk.

A little bit about venture capital. So since you're an investor, can you just talk me through how you make use of data during the investment process?

This is an area that we've made a massive amount of investment in over the last six, seven years or so. And I think the way we sort of think about data within our practice is the identification of interesting startups on the front end. And these are based off of 60 plus signals that we track Everything from web traffic to employee growth to funding, a bunch of different metrics that we track there. In addition to that, we also tried to understand we have specific platform. In our platform technology we have a concept of thematic search where you can type in data transformation and it will pop out a list of all the data transformation companies and then also rank all those companies as well around who's doing well and who's not doing super well based off of these signals that we have. And then once we've made an investment, we have benchmarking tools that are again over 70, 80 different metrics that show the health of the company across the entire everything from revenue growth, how efficient a company is around cash burn around all these different facets that are critical for the business. And as these companies scale and look for more funding, trying to understand what are the things that the company needs to fix over time. And we've sort of built this benchmarking tool as well that does that. So in some ways we've got everything from initial identification to like helping our companies operate better. The entire process we're using modern data infrastructure for.

Okay, that's cool. Data is being used like throughout the whole investment process. I do like the idea of using data to be able to rank companies and see how well they're doing. Actually, the other thing you mentioned was about using data in order to help companies operationally. Are you able to share any of the details on what sort of help you might be able to provide and what the sort of data indications are that a company needs help in some area?

So I think the one of the most important things is like to first identify. So when you're operating a company as a founder, you're spending a lot of time in your own world, right? So you're like, hey, this organization may be rough. This organization has room for improvement. As VCs, we're talking to over a thousand companies a year. And what we are able to do is help the founders and the management team understand what best in class looks like and even within our portfolio, understanding who are the best companies. So let's, for example, let's say a company that we're working with has a customer churn issue. They do a really good job in terms of selling their product, but over time, the customers don't stay with them for too long. Based off that, we can identify companies that have had the most improvement over the last 12 to 18 months or have had the highest customer retention rates in our portfolio and connect those organizations together. Be like, hey, do you have 30 minutes of your time. We're just trying to understand this problem and fix this. And we're sort of collecting a lot of that data on our side and building a data set that can inform those companies in a more meaningful way. So it's not just like advice that we're telling our employees, hey, you need to and growing at 100% or 120%. It's not that. It's like actual insight that we're collecting and connecting our portfolio companies together based on the challenges that they're facing.

So before we wrap up, can you tell me, are there any particular companies that you're most excited about right now?

Yeah, there's a few companies. So we've invested in a company called Incorda. It's one of the most interesting data query platforms that serve out there. They've got some of the largest Fortune 10, Fortune 50 companies as customers, and they're helping them on everything from supply chain, FP&A, and a bunch of different parts of the organization and getting insights faster than any other tool that you can find out on the market. And we also recently made an investment in a company called Coalesce, which is a data transformation tool on Snowflake. So if you're a Snowflake customer, you probably know of Coalesce, and they have hundreds of customers using the platform today, and using some of their metadata technology, they're able to manipulate their data significantly better than any other tool out in the market. So, yeah, those are two companies that we're super excited about and doing super well.

I've definitely seen Coalesce around at a few conferences. A lot of advertising going on there, so that's interesting. All right, so just finally, do you have any advice for any companies who are wanting to. To make better use of data?

Yeah, I think at this moment in time, it is very important to just keep your eyes and ears open. The technology stack is changing by the day, by the minute, by the hour. It is really important to make sure that you're building a flexible process and infrastructure that can sort of scale better and leverage some of these newer technologies in a more meaningful way. You don't want to be locked into a solution that may be outdated in a couple of years.

Okay. Yeah, it does seem tricky. Like the technology platforms, they're always moving. So you got to keep building things, keep doing things to keep up. All right, super. Thank you so much for your time, sir.

Thanks so much, Richie.