13 minutes of eComm Wisdom with Feifan Wang from SourceMedium 🧠 We chat about how his vision has evolved for his business since the launch 4 years ago already. From a Slack app, to a fully verticalized analytics solution for DTC brands, and now building an AI for any store to query their data. Appreciate you sharing your insights Fei!! https://lnkd.in/e3Uuzsy5
Transcript
Thank you for for jumping on and sharing sharing some of your insights in the space. So you've been running Source Medium for a while. For those who aren't as familiar as I am, can you describe in a few words what you're working on? Yeah. Thank you, Tim, for having me on and I really appreciate the opportunity. Excited to dive into all things ecommerce today and data, Yeah. So Source Medium, we provide a fully managed data infrastructure solution that is specifically designed for brands doing between 8:00 to 9:00 figures in annual sales. And essentially what we do is we we abstract away the complexity of unifying your data into this metrics framework that can then ultimately lay the foundation for a decision making framework that the entire company can adopt. That's still a little bit too high up in the air. So if we bring it down to the use cases specifically, you can really look at it in two sides. One is our BI solution, which dramatically simplifies things like reporting and internal dashboard building. And then our data warehousing solution allows for the direct access to our developer friendly datasets and schemas. And then that kind of opens up almost an infinite number of use cases, anything from statistical modeling, the customizing. Metric definitions as well as training your AI agents. So really are you? Do you feel like you're competing more today with a looker type of company as opposed to other like triple oils of the world? Yeah, I we actually see ourselves as actually directly competing with the modern data stack. And that is a concept that kind of goes across industries, right? This is really around businesses that are essentially building their data management infrastructure in house, right? Things adopting things like eye trend to Bigquery to transforming the data with tools like DVT or data Form and ultimately to a lot of the hidden layers of data. Management, right, anything from data observability and quality control to governance and metric definitions and schema definitions. And then BI, which is like we're looker sits ultimately. So I think the problem that we uncovered is a lot of brands adopted these solutions not knowing really how complex and expensive it was going to be to build and to ultimately maintain. So that's where we come in. We essentially focus on what is the output. Of a modern data stack. And that is similar across many businesses that have Omni channel sales, but with the heavy ecommerce focus. So that's where we can come in and just get you from zero to 100 super, super quickly. But you also know that this infrastructure is battle tested across over $2 billion of annual GMB that we process. Maybe to put it, is there like a cool case study that you can highlight from a recent client? Where the impact that you guys delivered is so clear and obvious? Yeah. So my favorite example is we're gonna be working for a few months now with cpap.com. So they are a direct to consumer. They sell CPAP machines and related accessories and products, right. So they were on the gentle with a lot of internally built technology that that includes the data management infrastructure stuff. And so I think they had a lot of pain points around just the cost of maintaining that infrastructure. And the ultimate ROI that was? Hard to see. I think there was just still a lot of friction in terms of like individual leadership as well as operators being able to access the data that they need and can trust to make decisions. So since they were already migrating ecommerce platforms, we had an opportunity to also have them dramatically simplify the data management side of it into source medium. So then what their tech team is able to do now is they can just like specifically focus on investing into. Things that that are specific to their business, right. We also have done a lot of great work with them in terms of getting the number to be boardroom ready to be numbers that FPN can accept, right, This CFO would would be happy about. So they're able to swap out some of our cost components, but take our revenue figures and other components that combines ultimately into the back of the KPI's that being in the entire company should care about so. I think what's really powerful there is think about when the CEO, the CFO, the CMO, all the way down to the e-mail marketing manager and the CX manager using one set of metrics and one set of definitions to make decisions. So I'm really excited about where that's headed. The other example that's been really cool is with our agency offering. So one of our earliest adopters is the snow agency, which is now part of Ave. Z. So they've actually being able to build an entirely white labeled dashboard business on top of our infrastructure, right. So that's provided them with a lot of differentiation and help them significantly increase close rates and just increase the amount of insights and value that they can offer to you to their customers besides like managing media. So that is like a core part of their offering now where they can essentially create these dash very intricate dashboards. That's that they know are gonna perform and then they know they can trust the numbers on to help to further help their customers and increase the value that they provide. I want to know like how when you're competing against the modern tech stack as or modern data stack as you've described, that's a lot of companies that are billion dollar companies that are going to be replaced by source medium. How, what, what have you done on your from a team standpoint and from like a tech build standpoint or product build standpoint to enable you to get all these efficiencies that you can actually strip out a lot of these expensive technologies and just go with source medium and. Yeah, I think the two core competencies that allows us to gather 1 is. We try to spend 80% of our week. Working with customers, my goal is that everyone in the company works with customers directly or indirectly, but everyone should have the opportunity to solve problems with the customers to get directly. And then that allows us to really understand the business specific domain and the problems that businesses want to solve. And then ultimately that allows us to create and recognize the patterns, right? And that's what gets absorbed into the product. So once bug fixes happen or improvements in our definitions when new metrics are shipped or created, right, all of our customers can benefit from that. And whether or not they want to adopt that particular metric where that particular data source is up to them. But what they know for a fact is there is going to be broad acceptance. Otherwise we wouldn't ship it to to everyone at scale and then. The other thing is that we copy other kind of early or other infrastructure companies that have gone IPO, right? So if you look at a company like Palantir, which is what we model our company after or a company like Hashicorp, right, those companies were entirely made of technical staff, right, in the early days, including sales. So our entire company is technical except for one, but even that person is savvy, right, technically. Stop being so everyone can write code, everyone can contribute to the product in a technical sense. And then that allows our R&D to just be basically the entire company all the time by focusing on providing value, overwhelming our customers with value on daily basis. I want to know, like when you went into Source Medium, you had a vision and now you're four years later I'm working with some 9 figure brands. You got a strong team behind you in a really cool product. Did you? Are you like on the path that you had envisioned four years ago, or is Source Medium today actually quite different than what you had set out to build four years ago? Yeah, that's actually such a funny question because when I started the, what I wanted was a Slack bot. The company was incorporated in May 2020. Uh, So what I wanted was a Slack bot. That's the thing. Your Slack that has access to all your data that can you can talk to and give you insights. I didn't know that was called L's, right? So I had no idea how to build that to be honest. And ended up becoming like a daily Slack reporting bot, which became actually semi popular where it had to be taken down. But it's coming back because of the API versions. But I think then that led us to this winding journey to become a dashboard company, right? But really just like ultimately now where we're going. Is that route right? Because the number one obstacle to adopting AI for our organization is unifying all your data, right? So that is like the most expensive time consuming on boring part of it. So now that we've been doing that for four years, I think what that sets us up for is to start really like extracting the value out of that data set in a dramatic way and that is going to be with the combination of enabling. Really use case driven statistical and machine learning modeling that brings open source options into the fray as well as internal AI copilots and agents. And we have some pretty exciting plans behind that. So it's somehow the we came full circle thanks to ChatGPT and then the rest was history. That makes perfect sense that you're going to go back to that route and apply some AI technology on top of it. Especially because you understand data sets of multi channel ecom players pretty much better than anyone. That's super exciting. Do you like, can you, are you guys profitable? Can you just allocate resources that you already have on the team to work on that or would that potentially suggest another fundraise or or something like that? Yeah, I think we have gotten to a place. We weren't always there. I made a lot of mistakes, but we've gotten to a place where we're pretty close to being profitable. The metric that I track closely is a annual recurring revenue per full time employee. We want to make sure that we learn from our lessons in the past and not just throw people at the problem, but really think about automation and really think about using technology to create productivity. Games, but I think what's cool right now is we have our entire business built on top of GCP and ex Coogler as well. So I have some allegiances there, but I also do believe their best positioned in the AI era as a cloud service provider. So we've been getting a lot of businesses on to GCP. We've we've gotten a lot of businesses to switch from other clouds like a WS and Azure onto GCP and our solutions flexible as well. You know, we can enable other clouds using GCP as the foundation. So so Google is slowly recognizing that we're actually working with Google directly on like doing some proof of concepts and by enabling like both Gemini Pro 1.5 right on top of your data as well as open source alternatives like Llama, right? So we'll have pretty hopefully we'll have some proof of concept to show the world too. But the other thing I'm excited about is really to get some. With the machine learning models enabled for brands, there's a lot of things around predictive analytics that people need like time series, forecasting, lead scoring, probability to churn. So we will probably have those capabilities opened up sooner rather than later. So very excited about that. Cool. Well, thank you so much for the for the wisdom. Thanks for the time and great to catch up as always. Yeah, speak to you soon. Alrighty, talk to you soon.To view or add a comment, sign in