Veritone, Inc. (VERI) Earnings Call Transcript & Summary

Hi, good afternoon. You have Steve [indiscernible] from UBS here. I wanted to introduce to you guys the management team from Veritone. On the call today, we have Ryan Steelberg, President; Mike Zemetra, CFO; and Brian Alger, Head of Corporate Development and IR. So without further ado, I'll turn the call over to them. Thank you.

Great. Thank you for the time today. I look forward to providing an update on Veritone and go into more detail about the exciting performance of the business and as we look towards closing out a strong year. So with that, first, just we're going to be providing some forward-looking statements today. I think recording just started, so I'll do the little preamble again. But anyway, thank you for the opportunity to speak to you today at the UBS conference. We're excited to provide a corporate update on the business, going a little bit more inside of all the exciting recent advancements and progress that we've made, and also really give you some forward-looking opportunities to see what we're excited about moving into 2022. So with that, I just wanted to make sure that -- we all appreciate that we're going to be discussing some forward-looking statements and providing forward-looking statements and some guidance that -- so please evaluate and take appropriate protections and as it relates to the risks associated with it. So with that, let's kick it off. Veritone, first and foremost, is de facto leader in enterprise AI software services and applications. We are a software and technology-first company, it's from our original pedigree, and we still maintain that very passionate focus as we've been able to grow this business successfully over the last few years. And the business has been around now for several years, I think, we first came to market with some of our first prototypes in early -- in 2015. So today, we have roughly a little bit more than 500 total employees that spanned, really, North America and now, post our PandoLogic acquisition, a major strategic hub in Tel Aviv in Israel. We have about over 1,800 total customers that span all different aspects of business verticals and market segments, and we're going to go into those in more detail later. And we do have hundreds of partners, partners that we sell through, partners that we co-sell with, integration partners, technology partners. So again, we are, at times, can go vertical. Just Veritone by ourself end-to-end into delivering a solution, but we also have built the right tools and interfaces and integrations that we work with a multitude of different partners, both reseller, consulting partners and technology partners. And we'll talk a little bit more about those as well. And our mission is to continue to be the leading AI software company in the industry. And thankfully, we've been able to create a very exciting, successful business that is also very passionate about bringing to market AI and an enterprise scale level for good. It is important that we maintain transparency, that we are open and build systems to present and snuff out bias, if there is a bias in our markets. AI is both exciting and scary to a lot of individuals, and so we look at ourselves as, very importantly, one of the early adopters and pioneers in the space that we need to set the course. So what you'll see here is not just us having strong discipline on being very transparent, on how we invoke and leverage AI, and how obviously, we've been subsequently licensing it and deploy that to our customers, but also in the products that we -- and solutions we built, including applications like Redact and other -- or the basis of the entire application is actually to provide transparency, for example, what we're doing in law enforcement. So again, I think what we're excited about is we built a successful business that is truly a leader in driving AI adoption, but we're also able to maintain our core values, what we call [ very values ], that we are really a leader in AI for good. And that is not just a purely edition statement, it's been memorialized into the actual profitable applications that we're licensing itself. So this is what makes Veritone so excited. I talked about it first, I'll state it again, is we started and have a long track record and pedigree of building world-class scalable software and SaaS solutions. From that experience and with our talented team, we've developed and expanded our platform, which we call aiWARE, to be the de facto leading operating system for AI in the marketplace, to help us really drive enterprise AI adoption. But more importantly, that's step one, it's really the transformation of these businesses, right, enabling them to hyperscale the ability to invoke and leverage AI to truly transform their business. That could be, again, for an efficiency gain or outright revenue growth and market expansion. The market we live -- that we're working in is a very large market. You guys know the studies as much as we do, but the enterprise AI market is big, big and very big. And if you actually look at some of the derivative opportunities of how AI is sort of interfacing with these new opportunities like the Metaverse, right, and synthetic content in digital avatars in conversational AI, we think that literally this is a $1 trillion industry, right? And Veritone is right in the middle of so many of these big market segments, but we're doing it with a focus from a technology and a software perspective, first and foremost. Because of our strong focus on products and applications and software-based workflows, this has really helped us sustain our growth going -- looking back, past many quarters in several years. But we're also looking at and will continue to look at very strategic M&A opportunities where -- whether it's technologies or companies that have interesting market penetration or relationships that will help accelerate Veritone's adoption in those respective markets. I'll kind of finish it off by stating that, again, we are a founder-led entrepreneurial team. We have strong, very data-centric expertises that we've been able to leverage those as we've sort of been able to bridge the gap between pure algorithmic data-centric based models to true AI and machine learning-based opportunities. And I think we've done that very effectively. And we have a track record of sustained growth year-over-year coupled with very attractive financial margins and in a financial model that we look -- that we think is very sustainable, but also one that we can we believe strongly we can continue to accelerate. I think with that, I don't know, if Mike, you're on, you're going to take this slide, but I'm happy to cover it for you. But again, at Veritone at a glance is we are the leading enterprise AI company, again, we're licensing our software and platforms and enabling hundreds over 1,800 customers, that are making it easier for them to jump in and adopt AI, but also on a platform that allows them to scale the investment of AI. Almost every single one of our larger customers started, I'll say, it's somewhere in the continuum of an investment into aiWARE to start building and deploying AI-based solutions, integrated into their own processes and technology stacks. And many of them have continued to grow. Meaning we're doing -- they're investing more of their processes and AI-based workflows on the aiWARE stack, which is leading to not only increased revenues on a per customer basis, but it's allowing us to expand the capabilities even of AI work based upon the feedback working with these customers to continue to advance the platform to make it even more scalable. The proof is in the indications of our momentum. We continue to post quarter-over-quarter strong growth and obviously, year-over-year growth, both organically and inorganically through combinations. So again, we are very excited about our performance. We see our pipeline being stronger than it ever has been across our different verticals, which gives us confidence that as we look to Q4 and finishing up this quarter and into 2022 and beyond, we expect great things both technically from Veritone, but also great performance fiscally and strategically. A few recent highlights, and we have many that we can talk about, but let me highlight a few recent and public announcements that we've made about business and technology at Veritone. I'm going to state again how excited we are about being really a first multi-cloud FedRAMP solution for AI that's been awarded an ATO. And that ATO, with the primary sponsor of the Department of Justice, is not just a multi-cloud acknowledgment and authorization to operate but it also expands our ability to service every single division and group within the Department of Justice. So looking back in rears, yes, we've been in FedRAMP, but it was limited primarily to one cloud-based infrastructure, a; and b, we were limited to the state attorney's office. Now the FBI, the Bureau of Prisons and all the different divisions of the DOJ are now able to be serviced. And obviously, we've been preselling in that area, so we expect great things from the DOJ and that relationship in terms of business and revenue in the future. A couple of other ones I hide out is the Metaverse is all the rage these days, and we've been involved in it for a while now. As in our media and entertainment segment of our business, we have many very strong relationships and clients who are the leaders in many aspects of the content ecosystem. The Masters Golf Tournament, [ NTAA ], CBS News, Bloomberg, CNN. These are customers who are already aiWARE customers of Veritone. Many of them are also licensing customers of ours, where we help deleveraging our platform to facilitate licensing deals. And that's allowed us already to have brokered with real material dollars in terms of contract value, transactions that allow us to use AI virtue bridge, in this instance, the Pac-12 content to the Metaverse in the context of the creation of NFTs at scale and through partnerships. Many of you have read about our NAV Product of the Year award that was bestowed upon MARVEL.ai. Our proprietary voice as a service solution that allows you to create synthetic voice. Yes, this is a turnkey opportunity for having dynamic synthetic voices in the Metaverse, or I'll call it in the gaming type of world, the unreal engines, and on an NVIDIA ecosystem, the Omniverse. But we've also shown how you can leverage at scale synthetic technology, synthetic voice outside the full Metaverse. If you guys listen to the radio and you hear DraftKings or many of our other brands, the ad copy you're hearing now is all synthetic, right, programmatically developed and generated and distributed by Veritone aiWARE and Marvel.ai. And then recently, we talked about it again is the close of PandoLogic. An exciting, obviously, financial, very beneficial acquisition that we made a few months ago. I'm excited to state that we've been successful in not only closing that transaction, but really starting to integrate it at a management level, at a sales level and looking a little bit into the future of technology integration as well. This deal was immediately accretive for us on many different fronts. But again, just a summary of the transaction again was it was roughly $150 million in total value, $50 million in cash and $35 million in stock upon the closing with an additional $65 million in cash and stock based upon very clear and specific financial earnout, right, and performance metrics. Again, and we're confident and we're very hopeful that we will have to pay them every penny, right, which obviously would bode well for the rest of the business overreaching. This was immediately accretive in terms of the financial impact to Veritone well. Coming here almost immediately, infusing a run rate of $50 million in revenue with very high EBITDA profit margins, and so we're very thrilled about that. We're looking forward to discussing and presenting our results here in the first part of next year as we present our Q4 earnings. But strategically, the PandoLogic also enabled us to build out and expand our technical team, adding over 7 people in product and engineering to our portfolio of brilliant talent, but also now enabling the company to spearhead more into many different types of business verticals. Almost everybody in the space these days is struggling with hiring. And PandoLogic is the de facto leader in programmatic AI-based hiring and job listing management and listing advertising. And it's something that we're -- we've been looking at for a very long time. We obviously have good insight into sort of the output and through the -- and application of what they're bringing to bear. But again, it's been an exciting business for us. We expect great things from PandoLogic in that vertical, but we also are very excited about being able to tap into that expanded market opportunity for us to continue to not only sell PandoLogic, but also to expand on other aiWARE-enabled solutions that we can cross-sell to our respective customer bases, so thrilled about that. Mike, why don't you cover your on -- do you want to cover just the recent convertible offering that we just -- I'm happy to cover it.

Yes, sure. No problem. Yes. We're super excited. This month, we closed the -- we closed a convertible deal. The initial deal was targeted at $150 million because it was more than 3x oversubscribed of $200 million. Key terms in the deal is a convertible premium of 133%. We did purchase a [indiscernible] premium, which says if our stock doubles, that we can have it converted at 175%. And which basically means north of $2 billion in terms of market cap. Overall, coupon 1.75% interest, only 3-year [indiscernible] with the [ make haul ] and no covenants. Payable in cash at our discretion, and total cost of capital and just slightly below 4% and 3.8%, including the pro forma rate. Net proceeds to us were just slightly north of $180 million. And on a pro forma basis, it puts us north of $250 million of cash and the use of proceeds, general corporate purposes and growth, including strategic acquisitions.

And I think with that greatly improved balance sheet and obviously, the recent performance and expected future performance of the company, I think we're in a very strategic and healthy position to really dominate the space. Not just against our peers, which we're very confident about -- but also just independently against any benchmark. So again, a healthy balance sheet, fast-growing company, expanding market opportunity and TAM. So I think, again, at any point in time snapshot, this is the most exciting it's ever been at Veritone, and I think we're just scratching the surface here. So now we have the capital resources to really focus on accelerating that growth for the next year plus. Okay. Second big opportunity here is -- we talked a little bit about it is the unstructured data market, That $100 billion TAM of market opportunity, really, where Veritone started focusing on early on, is an area that we're going to continue to remain very focused on, continue to invest in and hopefully continue to lead by a wide margin. So unstructured data sets is still the super majority of data sets out there as compared to structured data sets. For example, video footage, audio, even on the legacy documents that I'll say, may be in an alternative non-digital format, right, that have to be digitized via OCR or other things. So there's tons and tons of data that needs to be ingested, digitized, harvested and interrogated and analyzed with AI and we're -- and we believe we have the best platform and technology to accomplish that and the proof is in our performance. But what I want to touch a little bit on is this is really Phase 1, right? It's not just about the incorporation of starting to ingest and analyze or take advantage on structured data, it's the applications workflows to do next. And I think also what we want to impress upon people is the aiWARE platform is not just the ingestion in cognitive layer of applying AI to it. It is also natively integrated with the workflow and the application layer, right? Automate Studio, our proprietary low-code workflow platform and an environment, allows you literally to end-to-end ingest data sets, apply AI, come up with predictions and models and recommendations, for example, in media and entertainment. But then also seamlessly leveraging Automate Studio, the ability to build the workflow and the piping to deliver the end value and solution to the customer to solve the business problem, right? We -- the mantra we used years ago still holds true today, nobody is paying for 2 [ scoops ] of AI, right? AI is simply an element, it's a tool, a powerful tool. But again, it's the resulting workflow and application that presents the great value. And AI where the platform is -- was uniquely built and specifically built to accomplish just that. So a quick view of our technology stack. This is our AI proprietary platform, is really -- and I'll just go through the different levels of the stack at the end at the very, very top of the stack is the application layer. So this is ultimately how most of our customers are realizing the value of all the cognitive work and the cognitive engines and workflow that aiWARE enables. This is the user interface for some of our customers. We have over 10,000 individuals in media and entertainment end users who are sitting at their desks or on their phones and they're interfacing through an actual hardened application, right, entirely built on the aiWARE stack. In this instance, it helps them identify programming trends, advertising trends, advertising efficacy. In the context of law enforcement, the end application is used now by thousands of sworn officers, right, to help them sit through and redact or analyze body camera -- body and dashcam video footage, right? So again, the important thing is, and really, why we use the tagline that aiWARE is the OS -- is the operating system for AI is we purposely understood to build and build the interfaces and the development framework to deliver the end application, which we're -- again, most of our customers are realizing the value. But we've advanced a lot over the years. And obviously, now we have a very robust API layer in Automate Studio. If you go back one more slide, I just want to cover up with a couple more things. That allows, really, people to focus and build an end-to-end based solution. But 2 other points here, we'll go to the next slide, is 2 other major attributes of the aiWARE stack is we are completely cloud and platform agnostic. Almost every section of this, the application layer, integration layer, the cognitive layer, the data lake layer, all of these are modular and can be containerized and deployed into, really, any environment. Today, the entire stack is running on the biggest cloud solutions, both public clouds like Azure and AWS, but also government clouds. We're also in this entire stack was deployed in FedRAMP, right. And now multi-cloud in FedRAMP, both on Azure and on AWS in FedRAMP. This entire stack has also been deployed completely on-prem, which is very, very unique in this space. But also, we've created a modular sense now that if groups are showing up with an investment or they want to say they've made a significant investment to build a data lake around Snowflake, we seamlessly integrate with that data set. So meaning, they get to leverage every aspect of aiWARE but then still do seamless integration with an S3 framework or AWS or a Snowflake investment or even a [ day breaks ]. Next slide, please. So our focus on product first, we've seen some, obviously, major advancements over the last several years as the technology and platform have progressed. We started this entire business like many great businesses where we were trying to solve a very specific, limited problem that -- and it was basically trying to solve problems for one of our own businesses. And it was really trying to identify all our representatives, a needle in the haystack of unstructured content. We were trying to sift through thousands and thousands of different audio and video feeds to find out when specifically certain brand or a product mention was occurring organically, and it was a huge complicated problem. And that was it. And that was the exception kernel of the idea. We were successful at that and boom, very quickly, we saw and sort of envisioned aiWARE and what it could do for a complete multicognitive framework that will allow you to really look at and analyze in neural time any data set. Apply virtually any combination of proprietary or off-the-shelf AI models against such, and then ultimately deliver and build world-class and highly scalable solutions and applications against that aiWARE base -- I'm sorry, that AI machine learning base. It's advanced, and I think the biggest one to point out is, really, and this was about 1.5 years ago plus, but going from aiWARE 1 to 2 is really making sure we nail it real time. And a lot of groups focus on sort of batch-based processing simulations and in other machine learning and AI-based solutions based on batch-based data. So much of just from our background, we needed to almost -- we had to have a clear strategy for real-time and stream-based content. So again, today, we'll -- we process tens of thousands of hours of real-time stream ingest data. Again, both text-based stream ingest, but also audio and video every single day. And in parallel and in real time, we book a suite of our cognitive models to process index that content as it's being ingested. It's radically different. We do believe we're by far the most efficient in the industry. We're doing that and from an end-to-end perspective. And that really just open up so many more use cases for us. And then aiWARE 3 really was our push to become platform agnostic. As I mentioned earlier, our ability to deploy on-prem or in any cloud or private cloud environment, which was critical that we achieved, particularly for our Government and Legal & Compliance and Regulated Industries. And then just we continue to advance. We love software. We love building products. And obviously, this has been memorialized into, frankly, the expansion and build-out of very powerful tools for aiWARE such as Automated Studio but also complete proprietary new solutions and applications. MARVEL.ai, our proprietary voice as a service or synthetic content platform, which is so exciting. We're seeing such high demand for that solution. But again, I think what's exciting about this to summarize is that all of our advancements -- I mean, all of the big businesses that we're servicing across all these different business segments are all running on aiWARE, a common code-base. And for many of you in the software space knows, that's a big thing, and it allows us to really focus on scale fast, helps us minimize, always continue looking at technical debt or doing third-party integrations on legacy code basis. But that passion of that has really been a strategic opportunity for us. And we'll continue to look to partner not just with end customers looking to raise money, but there are big players out there that are doing great things. We've announced partnerships and workings with NVIDIA and others in the past. We will continue to focus on building and extending our scale and leverage by partnering and integrating with other market leaders in this space, again, such as Microsoft, such as Google, such as NVIDIA. Some of you may have read about it, but we did -- just as we've seen our business mature and we sort of evaluate the market opportunity, we really started to look at the business and have formalized that effort into really looking at holistically across the company as enterprise AI 2 major segments: Commercial Enterprise and Government and Regulated Industries. Let me point out again that all of these solutions, again, are going -- are all being run on aiWARE, the same common AI-based platform and code base. But the commercial enterprise business, the go-to-market, the sales and everything, as you can appreciate, is very different than how we're servicing regulated industries such as energy grids, utilities, state and local law enforcement and government agencies. And we've built the infrastructure underneath that as it relates to general managers, product managers, sales leadership, marketing, product marketing, to really be subject matter experts to hone their skills to make sure that we are communicating aiWARE and presenting aiWARE in the best opportunity, in the best fashion to service the needs and frankly service the required processes that some of these business segments produce. So -- and just to add one more point to that, Historically, our commercial enterprise business has been almost 100% client direct sales. Our field sales team is going direct. In the government regulated industry, it's almost out of the gate. It's been a much more partnership for, I'll say, proxy-based type of relationship. Working with leading prime contractors such as BAE Systems and others, working with Deloitte, who has obviously strong penetration or a [ mega-product ] contract owner and really licensing and being, at times, a subcontractor to a lot of these different partners groups. Therefore, that requires, obviously, a different management structure in a different way of organization for us to service those. So I think we've got a -- I mean, we didn't even call this a reorg, it was just a realignment that I think it's been performing very well for us just over the last several months. And we really think that this is going to help continue to accelerate our growth and our -- frankly, our ability to service more effectively our customers through last year and 2022. Commercial Enterprise Solutions. As we talked about before, we are definitively a leader in the space of working with a multitude of different customers to help them make sense of their data silos, but then also allow them to act upon it and really turn it into value. And there's so many different use cases, but if you just -- if I kind of describe one macro customer and just how they have advanced over the years is let's take an Odyssey. And Odyssey for those is kind of the -- one of the top major audio-based content companies here in North America. They are the second-largest owner of radio stations, a major streaming company, and we've been working with them for years. And we've been ingesting hundreds of real-time streams every day. All their podcasting, their over-the-air programming, et cetera, just a tonnage of content. We've also been ingesting other forms of structured data like spot rate data and other information that is critical to how they run their business. So in effect, we've now been able to harness all those different data sets, ingest those and create a common intelligent data lake. And this is, by the way, all done on aiWARE. And then working with them, we've enabled them to build and deploy solutions that their sales team can use, that their programming departments can use. Actually, that their -- the monetization in their traffic groups who are trying to find the right mix of what types of advertisement should be delivered and what type of programming should be delivered and when, that is almost turnkey to them now. And so that's a relationship that is truly started with the land and expand model. We basically employed years ago on a few different data sets and a few radio stations, and today it's a group-wide effort. And now we're actually doing completely new and novel things. One thing I'll kind of break here is we're excited to state that we recently have expanded the relationship with Odyssey to -- with MARVEL.ai to help them and empower them to start using synthetic content to start building truly dynamic and programmatic audio-based content using MARVEL.ai. That will be distributed both off over the air, traditional radio, but also online through radio.com and their podcasting network. So again, we're thrilled and that just is a good example of how groups have worked with Veritone to sort of tip their toes in the water with one small use case that could scale in its own right but also really allow them to expand with one partner, just Veritone, to service radically new different market opportunities such as MARVEL.ai. But again, we, again, enable customers to interface us -- interface with Veritone either through an API perspective, meaning they really are showing up with enhanced data science teams, and they have a good idea of systems control and they want interface through us through our API layer, absolutely. We have others that we empower them that they may have subject matter experts, but they don't quite have the engineering capacity, and they can actually build a lot of these solutions right on aiWARE using Automate Studio, our low-code environ. And then ultimately, we -- sorry, at the very end is we have our applications. We have dozens of very high valuable, hardened turnkey applications that we've been selling for years that actually can be used and trained and deployed in near real time. And as I mentioned earlier, we will service IT departments and data science groups, but we're also actually having end sales reps, advertising sales reps and content analysts, who are using end -- the end final built application from Veritone Discovery, redact attribute, hardened applications that literally with 2 few hours of training, they're literally off and running, and they have all the power of AI and aiWARE underneath the covers. So again, it really affords us the ability to service really any type of business relationship with commercial enterprises. Government-regulated industry. I can also -- I can just change a lot of the words that I just said on the previous slide and reiterate those. But I think the important thing to impress here is, number one, same technology stack. So all the solutions and applications that we've been selling and deploying with and through partners or directly to government and law enforcement groups has been built on in the aiWARE. So today, again, we have groups like Deloitte, who are building business practices on aiWARE, right, where they're taking -- we're also licensing and selling [redacted and hardened] end application to thousands of sworn officers across the United States, right, across hundreds of different police agencies in different states. I mean, that's a turnkey application. So again, very similar is we will always be ready and prepared to support our customers, who do invest, like a Deloitte, with the sophistication to build a lower level on our stack through our API level, but we will always remain passionate, core to our DNA, to have hardened end solutions, turnkey applications built on our stack that we can sell in a much more transactional efficient basis. And so government-related industries is no different. And with that vision, I think the biggest change between government and regulated industries and our commercial business is, number one, the proxies by which we have to sell through or with contract vehicles. Again, Deloitte is a good example, we've talked about them a lot, is we're usually the prime on many elements of working with Deloitte or it's a combination in a teaming agreement where we're both in effect of the prime. So partnerships. In the partnership ecosystem at least today is much more prevalent and, frankly, important in our government-regulated industries today as compared to our commercial business. And then lastly, which is not probably the most -- not the greatest, is the visibility of the pipeline and the timing of getting these contracts done is different, right? We don't -- we have less insight to when -- again, even when we win an award or a contract, when will that definitively turn into post-award and granted contract to actually a revenue opportunity. So there are differences. So the life cycle of these -- of the deal flows in regulated industries is radically different, I'll say it that way, than the commercial business. But again, very big opportunity. You'll continue to hear great things and advancements from Veritone about our government-regulated industry initiatives. And then lastly, Energy Solutions. Again, which is under the umbrella of government regulated industries is we have really developed something very powerful and impactful here. Our proprietary technologies and models through aiWARE are really showing incredible results in terms of driving optimization and performance, working with energy partners. Again, we've been talking about for a while now our efforts working with major utilities about real-time forecasting for energy supply, which obviously will drive decisions on distribution, but also solar smoothing and battery optimization. And a lot of these are -- we've put out press releases on that, so a lot of this is actually proprietary models that we've developed ourselves and in-house. So we're very excited about this space. We do think it's going to be a very material portion of our business in the future. But also, I'll put this in also, a very important in the category of doing AI for good, right? This is really going to make the investments that we're making in wind, in solar, in upgrading our -- and our energy grade consumption opportunities even that much more efficient. It's very hard to take advantage of all this IoT and these new elements of data of energy creation without having world-class intelligence software to manage that. And I think Veritone is very clearly in a advantageous position to not only be the leader in [ deploying ] some of that software, but also we're internally involved in sort of helping drive the narrative of making decisions. Maybe they should look at expanding too quickly into that new solar farm without having the appropriate battery -- battery storage facility to be coordinated with that effort. So again, we're viewed as a partner, not just a vendor for a lot of our efforts here with these energy customers. As I mentioned earlier, this is, in some instance, the family business now, but it's a founder-led business. We started again this whole business to solve business problems that we had internally through other businesses that we owned. And we bring a very rich background on a data-centric team. I think it's so important is everything in AI, machine learning is very predominantly based upon data, right? I mean training data, I mean data to act upon. And even in process automation, data is critical. Everything we've done has been built upon data, right? And not just data, but data at scale and speed to make decisions, right? And we come from an advertising [mark tech] background that goes back to the mid '90s. We've built some of the largest groups that have made, let's say, billions of decisions in a few seconds. And just to be -- sort of do the parallel, that means that we have to literally have to adjust and analyze millions and millions of input data sets and very quickly try to decide, at that time in these previous companies, what was the right advertisement to serve the right individual at the right time. So the data sets, and that was kind of our organic -- thankfully they were successful businesses, but that was really a setup in our PhD program to prepare us for what we feel is such a huge strategic advantage of ours is that not only do we understand large scalable SaaS-based environments in distributed systems, and we've been very successful about that building those solutions in the past, but they've all been based upon data, right? And so now it's really set us up to -- the proof is in the pudding. We've been successful, and I think we have the right strategy and right vision to really be the de facto leader moving forward. With that, I'll turn it over to Mike, if you can have the opportunity to speak, and you can walk through the more detailed financial data.

Sure. Great. We can go to the next slide. I'm going to go super quick because I think we're passed the time. Listen, we've got a track record here of driving sustaining growth. This year, we're going to be projected to do upwards of 80% top-line growth. More importantly, we are targeting on the upwards of our guidance to be breakeven for the year from the non-GAAP net income, net loss perspective. Just comparably, last year, we earned north of $21 million in revenue. This is really driven by sort of pro forma revenue growth. As a reminder, we acquired PandoLogic on September 14, so towards the latter part of our quarter. Assuming we acquired them at the beginning of the quarter, our revenue growth would have been 34%. That's really driven by software product and services growth year-over-year. This is comparing Q3 in the prior year of 54%. Gross margins of 82%, they're basically flat year-over-year. And then net income gross margin of roughly about 11%. This top-line growth is really driven from customer. Our customer count on a pro forma basis increased 35% year-over-year. But more importantly, our average annualized revenue, which is basically taking our revenue over the past 12 months and average the unit over the number of customers, increased north of 90%. On top of this is our gross revenue retention, which is really our proxy for churn, north of 90%, and I can tell you with the likes of Amazon, et cetera, it's well in excess of 90%. So one thing to point out, we didn't mention before, PandoLogic. On a pro forma basis, they're projected to do over $50 million of revenue, and they have a plus 50% in operating margin. So the total acquisition cost is roughly about 3x revenue. As we go to the next slide, really talking about trailing 7 quarters, the CAGRs here on the revenue side are very impressive. North of [indiscernible] And as we mentioned, both quarter in Q4 facing non-GAAP net income of roughly $8 million to $10 million. If you look at Q3 overall on a GAAP basis, our software products revenues grew 169% year-over-year and 62% sequentially. And overall revenue was 44% largely driven from our managed services. Our managed services, which include the advertising and content licensing, those are going to grow consistent with industry standards, more 10%, 15%. And that's it. On a pro forma basis, in terms of balance sheet, where, as I mentioned before, we're north of $250 million in terms of cash, and we're super excited about prospects for next year.

And then lastly, we're just looking at the balance sheet. As we discussed earlier, at the end of Q3, sitting on cash, cash flow have been around $72 million. As we mentioned earlier, post the recent capital raise, we're sitting right now with north of $250 million, with about right around $200 million in unencumbered cash, less media dollars held by some of our media and advertising clients. So again, from a balance sheet perspective, by far, the healthiest position that we've been in the company and its history that, again, really allows us to focus on fueling and accelerating both organic growth and providing the opportunity to be strategic with corporate M&A when the opportunities present themselves.

Yes. I think that concludes our presentation today.

Yes. Thank you, everybody. And again, just hopefully everybody has subscribed to our newsletter. We will continue to update you on exciting new dealings and happenings at Veritone, and we expect a very exciting balance of the year and a very exciting and financially beneficial 2022 and beyond. So thank you for your time today.

Thank you.