Alex Epshteyn, Zignage

The 16:9 PODCAST IS SPONSORED BY SCREENFEED – DIGITAL SIGNAGE CONTENT

If a company wants to hang its business hat on the proposition that it is very good at visualizing real-time data to screens, it helps to have a big, very familiar client that heavily uses that sort of thing.

A small New York City start-up called Zignage has that in the New York Stock Exchange - providing and maintaining a platform that shows the numbers and trends charting on screens around the hyper-kinetic trading floor in Wall Street.

The company grew out of an NYU media lab and spent its first few years working mostly behind the curtains, developing signage and data-handling capabilities to software firms and end-user clients. But a few years ago, the company made the decision to develop a brand and start selling its data-centric capabilities directly to end-users.

I had a great chat with Alex Epshteyn, the CEO and Founder of the company, about how it got started, where its headed, who it all serves, and how there can be a huge gulf between software shops that can take a number from a shared data table somewhere, and running mission-critical, hyper-secure visualizations on a stock exchange floor.

Subscribe to this podcast: iTunes * Google Play * RSS

TRANSCRIPT

Alex, thank you for joining me. Can you give me a rundown on what Zignage is all about, how they started and how long you've been around? 

Alex Epshteyn: Absolutely. Thank you for having me, Dave. Zignage started in 2009 formally, and we started at the NYU Incubator while I was doing my graduate work at the Media Lab in NYU and suffice to say the company was more interesting than the graduate works. So I started doing that, even though I'm from the east coast and this doesn't typically happen, it kinda happened here. So initially, conceptually, we were gonna get into the digital out home space and we were gonna build an auction backend that people can bid for spots on digital signs. So kinda a slightly novel idea, especially in digital signage and we couldn't do a big enough raise, and then we found a number of these sort of remnant advertising platforms coming into the market and we decided, since I have a pretty good little black book of enterprise clients, and we built the platform to about 50% at that point, in mid to end of 2009, let's try our hand at some enterprise folks, and what ended up happening is a trajectory that basically pushed us for about eight years, which is we built a middleware and a toolkit, essentially our own toolkit, that enabled us to build very quickly CMSs and builds and anything related to that, data bindings for third party systems like CRM systems and CRP systems, a variety of backends essentially, and we essentially entered OEM space.

So we built products for other companies. Some of them were large, some of them were small. We had a tremendous amount of NDAs and non-competes, as you can imagine. These companies would not like you to advertise your own stuff while you were building it for them and typically we would have maybe one or two of these customers at the same time. So from 2009 to about 2017, maybe a little bit later even, we basically did work for third parties and we built a lot of different solutions, and around 2018, we decided that we were gonna attempt to productize. That means, essentially build our own, front facing, become a brand, and move away from a pure sort of project solution, even though we had a product in there. But it was a product for us, not so much for the end customer and to get into the market and so we did, and in the meanwhile we had two direct customers during almost all the time.

NYU was one. We had a number of schools at NYU that we were able to pitch, and successfully had running, so NYU Law School, NYU Engineering School, where I was a student and then NYSE where we initially partnered Thomson Reuters. So Thomson Reuters did the data and, most of the application stack actually, and what we provided is a device management framework and advanced players to run the WebGL and all the other things that they needed to run for the New York Stock Exchange. This was under the NYSE-Euronext regime, which has since been bought by the Intercontinental Exchange. This was in 2017, which was a formative year for us. As I mentioned, NYSE under the new ownership came to us and said, “Look, Thomson Reuters is relatively expensive and essentially they're reselling us their data, how about you guys take on their responsibility?” You get nine months to replicate and you get this support contract that basically takes over for them, at a discount for them but it was a nice option for us. 

We took on the challenge. Because we were able  in these intermittent years we built up so much experience and know how to deal with realtime sources, realtime data sources, and WebGL specifically to make things pretty bulletproof whereas perhaps some other HTML5 technology that is fairly popular in digital signage would is maybe not robust or maybe not as performant. So we took that toolkit and applied it to over essentially at the New York Stock Exchange and took the contract over and successfully we did that. So at the New York Stock Exchange today, they're actually running two separate solutions from us. They have our more standard on print solution for their marketing group and then they have a much more customized, almost like an OEM version for their trading real time data, which are now classed as a number of financial data widgets. 

So if I'm at the NYSE and I'm looking down on the floor, or I'm walking around the floor with all the guys with the funny jackets and everything, those various dashboard screens that I see with all the pricing indicators and everything else, that's all being driven by you?

Alex Epshteyn: That is correct. So everything essentially above the workstation level, everywhere above the trader level, if you just look up above the 5’8” level from the ground, you’ll basically be looking at our solutions. It actually is a full gambit of our capabilities. We have synchronized video, real time widgets for financial data consumption, charting types of things and a lot of different ticker technologies that we've custom built and some of our generic ones, and streaming as well.

The only other company that works with us at the site is Haivision, so they provide the backend system and supplementary streaming solutions. So we consume their feeds and also feed them. 

They’re a video distribution company? 

Alex Epshteyn: That's right. So we're actually partnered with them. So they're one of our partners in space. We like working with them, they are a nice Canadian company to say the least and I know some of the original folks that sort of constituted the company and they have grown as a company tremendously through the years. So we really like working with them. 

Yeah, this must have been a really big holy shit moment for you guys when you got that deal because it's not like winning a hundred locations QSR chain or something, this is the New York Stock Exchange. It's on the TV every day with endless photos and everything else, and it's mission critical. Like you can't say, oh, we're just doing a software update and we'll be back in 10 minutes?

Alex Epshteyn: Indeed, and the escalations we get are pretty hardcore. We have just a few minutes to get things going, and philosophically, we try to blend some aspects of redundancy with a lot of resiliency because redundancy itself, some folks who deal with these sorts of mission critical situations, could itself present its own set of problems, right? So you want the system, the platform itself to be as resilient or high availability as possible to use a term out of the server space. 

So yeah, it was a huge thing for us and ultimately, we specialized in a lot of financial services and non-retail banking is a more generic category or an area we do very well in and we work with some integrators in the space that are known for it as well in terms of channel. Currently our CTO is actually the chief architect of the Thompson Reuter solution. He came on board with us a year ago, a year and a half ago as a full time hire. He was a consultant for many years after Thomson Reuters got customization space, and he worked with us for a long time and then finally our CTO to do other stuff, and Steve came on board. So we're very well positioned for this work. 

So for your company, if you had to do an elevator pitch saying what all you do, what do you rattle off for them?

Alex Epshteyn: I think what we would do is, as you mentioned, mission critical type of usages, whichever vertical, right? We've done things with SCADA. We've done things in transportation that I wish I was at liberty to say, maybe soon, and it doesn't have to just be financial data. It could be sports feeds. It could be building services, things of that sort that are critical for the use. That's one of our specialty points. 

The other is, I would say, while we're very happy to have relationships with a number of hardware companies, we still have really some high end hardware that we field. So what we do is, for very demanding applications, not necessarily mission critical ones, but those overlap obviously, we provide a full-stack solution, and these players, we're getting into the realm of show control type of players, really beefy and professional level graphics capabilities. So we do sell those. Those are fully our stack, and this way we can guarantee basically the solution as opposed to having us do a certain portion system integrated to another and so forth. 

The last thing I would say is while we still support some level of OEM work, we currently have two customers that we work with. Our business model changed a bit in the last three years of supporting them. We have our standard SaaS business and in some cases we modified it for on-prem. So it's already flexible, but we also have a platform as a service offering to really support those OEM customers. So it's a lot less expensive in volume, very scalable, and I would say those are the things that really make us stand out. It's real time data, data visualization, full-stack solution with hardware to do very difficult things often, and finally, configuration where people assume real, ad-hocs customization. There's an assumption there, right? If you're doing something very bespoke, the assumption there is that it's gonna be insanely expensive and take a long time to build and that's true if you haven't built two dozen variants of it and you don’t have a toolkit to basically assemble it from parts like a LEGO set, which we do.

I would assume that your calling card when you go in to talk to opportunities, when you can say, yeah, we do the New York Stock Exchange, we do all the data handling on that, and you could imagine it's more than a little bit secure and mission critically oriented. I suspect that makes the target customers feel pretty comfy?

Alex Epshteyn: It does, and even before them, it makes consultants who put us on the bid lists and generally are interested in finding parties that can actually fulfill the scope, call us. So we don't really advertise much, and that's gonna change, I think, maybe next year. We're gonna do maybe a marketing splash at some point next year. 

Right now, it's all word of mouth, and we do get a lot of calls. There's a lot of projects we actually pass on because they're not in our sweet spot and they're distractions, but the projects that we do take on are often difficult. We even do work in retail, as I mentioned to you, and the types of deals we take in are always really heavy data integration, visualization, where they are very automated workflows, there's almost no humans involved where the humans are basically special events, and then the system essentially corrects for automation again. 

Yeah, I've been writing about data visualization for 6-7 years now, and when I started writing about it, it was pretty rare and beyond FIDS displays and things like that but it's now pretty standard. I'm curious because you guys are obviously super deep and experienced in that area, when you see all the other software companies saying, yeah, we do real time data, we can do realtime data handling, we can integrate, we have APIs and this and that. 

When you get into a conversation with a prospect, how do you distinguish what you do versus other companies who say, yeah we do all that too, cuz I suspect it's different? 

Alex Epshteyn: It is. One of the first things we've put on a table is that we can mostly guarantee our resolution time SLA, nobody else can pretty much. Most people will be aggressive, pick up the phone and work the problem, but the way that our stuff is built, we can fix the problem. We can guarantee fixing the problem within a certain period of time. Now it's not inexpensive, sometimes it's actually affordable for a lot of types of businesses where a fully custom solution would not be. 

The other one is that most data visualization takes a lot of shortcuts, it really leverages, not to get too deep in technicalities unless you want me to, basically JavaScript and CSS, the mainstay of HTML5. But all of our data visualizations are built in WebGL. It's like the difference between driving a car on the road and driving a bullet train on tracks, right? There's no interruptions to the bullet train. It'll just go and it'll be on schedule. There's no interruptions. There's no jitter. There's no movement. That sort of paradigm. So we like to guarantee behavior of our data visualization, especially dynamic like charting or graphing libraries that we use and implement. It's actually extremely difficult to build something that you would think is easy like a ticker or crawler. 

Whatever data that's feeding it, I’m sure we both have probably seen a lot of instances where it stutters, it has problems, it doesn't refresh on time and doesn't deal well with different fonts and whatnot. That's just not true of our solution. Our solution is, I would say, cutting edge on dynamic data visualization. 

So for an end user or for an integrator, they have to educate themselves that just because a company says they can do real time data doesn't mean they can really do it. That means they might be able to reflect a number that's in a data table and show it on a screen, and that's quite a bit different from what you're talking about. 

Alex Epshteyn: It is and maybe the third aspect is most of the companies we work with already have accounts with the big data warehouse places like Refinitive, IBS, and a number of others, so we already are super familiar with these back ends. In fact, we have things that monitor the APIs. We routinely do a lot of monitoring of real time or just dynamic sources. So this is a huge value add in the industry, and I wish more providers would do that because ultimately, if you are a data fed platform, it's up to you to tell the customer something's failing on the back end because they won't know, they'll assume all sorts of things, but you need to critically have the tools inside to tell what's going on, and if you build it out in a smart way, you can also alert the right people at the right time that something's happening and to look into it. So you can be proactive about it. That's the third item, I’d say. 

They also change like the schemas and everything without telling people, right?

Alex Epshteyn: That's true. But it's a super exciting space. Once you have the core technology built out. You could really do a lot, in terms of, consuming this kind of data and I think generally, signage, we're in a slightly privileged position regarding this, but I think there's a move into industry towards generative and procedural content away from more Codec-heavy content. Although, there's obviously gonna be overlap for many years for both.

We certainly support Codec playback in a variety of ways, synchronized, on different players and so forth, and there's nice innovations like AV1 coming onto the market nowadays. But you could do so much more with generative dynamic content, it's a big difference. For instance, we had a client that wanted us to expose much more of the controllability of a layout, standard design tool inside of our platform. Now, typically we would not wanna do that because there's some nice tools on the market like Premiere, like After Effects, real tools that they generally use. But the problem that certain customers power users I would say are having is they don't wanna have to export an After Effects file and have it encoded in something, that's time, that's sometimes money because they do it externally because they don’t have a kit on-prem, or in the cloud. 

So what we've done is basically have a simpler version of something like Adobe Premiere or After Effects that lets them make quick changes in some key framing or some transitory effects and they don't have to put the whole thing into a codec. So that seemed to really resonate with certain power users that we have and directionally, it's the area that we'd like to innovate in.

Is it important to make a distinction between generative data for business applications and generative data for artwork? Because I see a lot of video walls out there that are set and forget. They're driven by generative data and it's just these abstract visuals that are swirl and kind of bloom and everything else, but that's very different from, I think what you're talking about, which is what on the screen in terms of charting or what appears is based on what the data is influencing, it's it's shaping what appears? 

Alex Epshteyn: That’s correct. A lot of general data is canned, right? It's almost like a video basically, and some experts, some design shops typically would change it for you, and it becomes evergreen content, day two, three, and day four. What we try to do is something a little bit different and we work with some really nice design companies as well. So just to be completely clear, we don't do the design ourselves. We typically either partner with a company that's really good at it. Sometimes the company brings us into the opportunity, right? 

The consultant can also spec us to partner with somebody or the end client may have relationships with companies that do this very well. But, I would say the formulation, the recipe for this kind of thing, to make it dynamic is a few things, and that's where this sort of generative content becomes more like a Zignage type of problem, as opposed to something that you could hire a design house to basically build for you, right?

One is that you could update content even if the filters or the generative piece is running. Separately you might be able to in CMS have the tools to change the filters of the generative option, just as I explained prior, and finally have trigger conditions. We do mostly casting, right? There are some great companies in space. I think they're very good at that kinda stuff. They do a lot of smart interactive signage. We do a little bit of that, but we mostly do narrowcasting. So in our world trigger conditions come from some sort of backend system. It could be a calendaring system, it could be something smarter, right? Where it's not just a boolean condition. It could be a multivariable that basically has to click off a list of things that can happen. And that's really where we can add a lot of value and it overlaps with the kinda work we do with the New York Stock Exchange. We generally term it as business logic So we really do some smart business logic and I think it's actually, there's a lot of growth in that area once we apply modern sort of machine learning to it to make it extensible to go further. 

But with that kind of approach you have an ability to modify a piece of content continuously, right? It's a living piece of generative content, even if it's not dynamically fed with financial data sources, or sports data sources. 

I haven't seen your user experience, but I'm guessing people listening to this are thinking, this is really interesting, but I'd be terrified to try to use this software. What’s it actually like?

Alex Epshteyn: You’re not gonna be terrified because we are one of the proponents of nearly or fully automated systems. So often what we do for non-power users is to give a build out to the software that our customers use, and then everything is essentially this business logic that I'm describing to you. 

It’s kinda like a headless CMS?

Alex Epshteyn: It's like a headless CMS for the non-power users. For the power users that really like their tools like Adobe, or you could just use a Dropbox or some sort of hotfolder mechanism. We're also partnered with a number of DAM solutions. There's a lot of workflow that happens in digital asset management solutions, including tag based workflows. 

We do a lot of tag based workflows nowadays, where we consume the tags that are done in a DAM, and essentially they find their way onto the right players at the right time, and on the flip side, we do have a standard suite. It's actually going through a major overhaul at the end of the year, what we call Z Cast 6. It does have a number of these power tools. But our CMS generally follows a certain idea. It was popular for a while and it's hard to execute unless you have our kinds of customers, which is what we call an additive UX. So it's the opposite of something like Microsoft Office, right where you have a billion features and there's a long learning curve if you wanna learn everything.

What we do is really try to identify the user story behind what needs to be done. We create the access controls that really expose certain parts of the CMS, and even within the same context, add or remove tools as needed. That creates a situation where there's almost really minimal training. I think one of the biggest problems we're trying to solve for our direct customers, or channel customers is the attrition that happens in major enterprises for users of digital signage, right? Like one of the biggest problems we face even in huge banks is the fact that digital signage is consigned to a webmaster subcategory. Like they manage the CMS that's published on their portal, and then somebody in that team or a few people in that team handles digital signage as well. So that's historically been a problem for our whole industry, and what we're trying to tackle is kinda remove both the friction of adoption and also try to give them the tools that they need, and if they use tools, bridge those tools, that's our philosophy on that end.

So what's the structure of your company? Are you a private company? 

Alex Epshteyn: We are a private company. We're an LLC in New York, and we're about 20 people. Most of our development used to take place until very recently in Ukraine because one of my partners and I from there originally. So as this topic is in the news, unfortunately, forget about our team. The fact is cities in the eastern part of Ukraine are partially destroyed but luckily a lot of the folks that we would use are in the Western part of Ukraine now, and we continue to use them but not all of them unfortunately.

So you're having to manage your way through that along with other things, right? 

Alex Epshteyn: We did, and they're very talented folks. We have worked on so many projects. 

Yeah, it's interesting. I was trading LinkedIn messages with another company and he was talking about operating out of Odessa and they're still like opening QSRs and things like that and putting in menu boards. 

Alex Epshteyn: Good for them. That's exactly what they should do. 

Yeah, and I was thinking, boy, all the other challenges you have out there, like supply chain and everything else, layer in a hot war on top of that. Good lord.

Alex Epshteyn: Our problems are very small compared to the real problems in Ukraine and the world. But it's a small world. You sort of face these things as they come. 

Well, hopefully someway or other, it gets resolved. I'm not quite sure how, but this was great. Can you let people know where they can find your company online? 

Alex Epshteyn: Sure. It's Zignage.com 

So signage with a Z on the front? 

Alex Epshteyn: Correct. The last word is Zignage. You find me on LinkedIn, Alex Epshteyn. That's where mostly we do our sort of minimum branding that we do. 

All right, but we'll be looking for more later in the year, right? 

Alex Epshteyn: Absolutely. We're excited to make some announcements in the transportation space, some more in the financial industry and some more in retail.

All right. Great to hear it's going well for you. Thanks so much for spending the time with me.

Alex Epshteyn: Thank you, Dave. My pleasure.