Edgio Inc (EGIO) 2018 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Limelight Networks 2018 Third Quarter Financial Results Conference Call. (Operator Instructions) Please note, this event is being recorded.

I will now turn the call over to Dan Boncel, Limelight's Chief Accounting Officer. Please go ahead.

Good afternoon, and thank you for joining the Limelight Networks Third Quarter 2018 Financial Results Conference Call. This call is being recorded on October 18, 2018, and will be archived on our website for approximately 10 days.

Let me start by quickly covering the safe harbor. We would like to remind everyone that we will be making forward-looking statements on this call. Forward-looking statements are all statements that are not strictly statements of historical fact, such as our outlook for 2018 and beyond, our priorities, our expectations, our operational plans, business strategies, secular trends and feature functionality announcements. Actual results could differ materially from those contemplated by our forward-looking statements, and reported results should not be considered as an indication of future performance. For more information, please refer to the risk factors discussed in our periodic filings, including our most recent annual report on Form 10-K. The forward-looking statements on this call are based on information available to us as of today's date, and we disclaim any obligation to update any forward-looking statements except as required by law.

Joining me on the call today are Bob Lento, our Chief Executive Officer; and Sajid Malhotra, our Chief Financial Officer. We will be available during the Q&A session at the end of prepared remarks from Bob and Sajid.

I would now like to turn the call over to Bob Lento.

Thanks, Dan, and good afternoon. Today, we announced our third quarter results. This was a solid quarter, building on the strong momentum of our record performance in the first half of 2018.

Revenue in the third quarter was up 7% year-over-year. GAAP gross margin was up 30 basis points over the year-ago quarter. Our GAAP net loss improved over 80% year-over-year. And our non-GAAP earnings were up over 40% from the prior year quarter.

Adjusted EBITDA grew as well, up 5% from the third quarter 2017.

In addition, during the quarter, our total cash balance, including marketable securities, increased over $7 million. We're pleased with these results compared to a very strong third quarter in 2017, which included sizable customer events.

This year's third quarter performance reflects the impact of the expected price compression, resulting from the successful contract renegotiations from 6 of our top 10 customers as discussed on the last quarter's earnings call. Additionally, we made the conscious decision to deprioritize certain traffic that we do not find attractive for a variety of reasons, including its quality, price point and type of content. As a result, 2 customers, who were previously in our top 10, are no longer on our network, which we believe is a positive outcome, allowing us to serve more customers' traffic in industries that align better with our overall business objectives.

Even with these headwinds, our third quarter revenue and gross margin were better than our strongest quarter in 2017, which was a record year for us. I'm proud of the momentum we continue to generate, even as we look at ways to both improve our margins and reserve our capacity for our best customers. I continue to believe that we've established a solid foundation for future financial performance that positions us well to achieve our financial goals in 2018 and beyond.

During the third quarter, we made progress on multiple priorities. We continue to see evidence of improving customer satisfaction as traffic volume grew to record level, while incident tickets continued to decline year-over-year. We're pleased that our efforts to meet customer needs have resulted in more traffic as the quality and performance of our network continues to improve.

Another indication of improved customer satisfaction is the result of our recently completed annual customer survey that generates our Net Promoter Score, which we use as a benchmark for customer satisfaction.

Last year, we achieved the score that we believe was best-in-class based on all we know about NPS in our industry. The good news in this year's feedback is that scores were up slightly again. On an equally weighted average, we're especially pleased that our NPS for our largest customers improved at even -- at an even faster rate. We're proud of our strong customer satisfaction levels, which we believe will continue to drive top line growth going forward.

Earlier today, we announced an important strategic partnership with Ericsson. As part of this agreement, we will deploy and manage our content delivery capabilities on the Ericsson Unified Delivery Network or UDN, which is an edge cloud platform with over 40 points of presence within service provider networks. We plan to install our software technology on UDN's existing and future PoPs, which will expand our footprint -- our delivery capabilities globally and increase our delivery capacity in strategic areas for our customers. As a result, we'll be able to expand our revenue by using the UDN sites for delivery to our customers, and Ericsson will be able to expand its revenue by selling more CDN and edge services to its customers.

We see this partnership as a great revenue opportunity for both companies starting next year, as we implement our technology on the Ericsson UDN network. This deal has further strategic importance as it places us inside the carrier networks and aligns us with the industry leader in the deployment of 5G, which we expect to support the acceleration of our edge services strategy.

I'm excited Ericsson chose Limelight for this partnership due to our superior performance. This will allow both companies to provide even better performance. Overall, it's a huge win for Limelight, Ericsson, the service provider partners on the UDN platform, and most importantly, all of our customers.

In the third quarter, we continued our effort to improve customer experience and better position us as the market leader through the development and launch of key functionality across our product portfolio. For example, we recently announced our breakthrough in interactive real time video streaming, which was demonstrated at IBC 2018 last month and well received by both existing and prospective customers.

This new offering, Limelight Realtime Streaming, is the industry's first global, scalable subsecond live video streaming solution that is natively supported by major browsers and devices. This new service also supports integrated real time data, making it possible to create interactive live online experiences. This groundbreaking technology opens up a world of possibilities for how viewers can interact with each other and with content distributors.

We believe many industries can benefit from real time streaming, including sports broadcasting, gaming, live auctions, gambling, surveillance and more. This new offering establish -- establishes us as the industry leader in subsecond global video delivery. We expect to see a new generation of immersive, low-latency, content-rich applications driven by Limelight Realtime Streaming.

As demonstrated by this new offering, we continue to invest heavily in video to become the leading provider at scale. We believe in the strong future of the OTT market and its potential to drive significant growth. Based on industry research, we expect over 80% of Internet traffic to be video-based, and this traffic is likely to grow fourfold between 2016 and 2021 at a compound annual growth rate of about 30%.

Additionally, the subsector of live video traffic is expected to grow fifteen-fold in the same period, with a CAGR of around 75%. These compelling statistics are especially exciting to us as we launch Limelight Realtime Streaming. With this new offering as well as other video-related product and network enhancements, we see great potential to increase our market share with our existing customers as well as expand beyond our current customer base. We're confident in our strong video delivery capabilities.

In the third quarter, we continue to make progress in edge services, which leverages our infrastructure to address our customers' needs at the edge for low-latency and connectivity. In addition to our strategic partnership with Ericsson discussed earlier, we recently announced the partnership with Linode, which combines our global private network and Linode's compute capabilities to provide unmatched flexibility for edge compute deployments.

Customers can now provision, compute, storage and load-balancing capabilities on demand across Limelight's globally distributed edge locations to quickly deploy and scale their workflows. We're pleased with the traction we're gaining in the edge services markets, and expect to gain momentum as we head into 2019.

In summary, we generated a solid third quarter and executed well on our long-term strategic priorities to create customers for life, grow profitable revenue while generating cash, deliver significant and innovative features and capabilities, and improve our position as employer of choice.

We continue to have great confidence in the industry we're in, the business we've built and the quality we're providing, the customers we are serving and the execution we're delivering quarter-after-quarter. At the end of 3 quarters, we're pleased that our financial performance is ahead of plan.

We're happy with the progress we've made this year on many other fronts, particularly our strategic partnerships, additional feature functionality and expanded capacity. We are on track to meet our financial goals for the year and expect to end the year strong by continuing to drive momentum in our core business while focusing on strategic initiatives in edge services and video. We believe this will translate into sustainable, above-market shareholder returns.

This was all made possible by the hard work of our global team, and I'd like to express my gratitude to each and every one of them for doing a job well done in the quarter.

With that, I'll turn the call over to Sajid to discuss the third quarter's financial performance in greater detail and our guidance for the remainder of 2018. Sajid?

Thanks, Bob, and good afternoon. I'm happy to discuss another great quarter for Limelight following our very strong first half at 2018, a quarter in which we generally met or beat expectations and stayed on course to deliver the best full year performance in Limelight's history.

We are very excited about our new products, initiatives and relationships, including Limelight Realtime Streaming, our new agreement with Ericsson, the developing opportunities with Tencent, and the possibility of Limelight contributing meaningfully to the development and deployment of global edge services. These developments should forever move Limelight to a faster growth and financially stronger enterprise, but first let's discuss the third quarter.

Starting with revenue. In the third quarter, we reported revenue of $49.3 million, up 7% from Q3 last year. This is the eighth consecutive quarter with year-over-year revenue growth and highest third quarter revenue in company history. While it is seasonally the weakest quarter, at $49.3 million, it is better than the best quarter from last year. Foreign exchange headwinds in the quarter amounted to approximately $200,000, or less than 0.5%.

International customers accounted for 42% of total revenue in Q3 compared to 36% a year ago. Approximately 18% of our third quarter revenue was in non-U. S. dollar denominated currencies.

We believe, we provide CDN services to some of the largest customers in the world and have one of the large -- and have one of the highest ARPUs in the industry at $73,800. Our top 20 customers account for approximately 72% of our total revenue, consistent with last quarter. At the same time, we remain disciplined on the type of revenue we pursue and protect.

In the quarter, we shed some large legacy revenue with poor price points, given the type of content and our overhead to deliver it. The loss of this revenue should be generally positive to the margin and profitability profile of the company. This is very hard to do, and equally hard to communicate. When we've come across revenue where the incremental cost of carrying it translates into negative margins, we have tried to say disciplined and turn it down. Our competitors have picked up this revenue to the best of my knowledge. It may serve them and their shareholders well, but we believe it does not do the same for us.

Moving on. GAAP gross margin was 48.7% in the third quarter, an increase of 30 basis points over the same quarter last year. We continue to expand our gross margins, even with the contract renegotiations we mentioned last quarter. With the pricing in place, with our bigger customers we expect continuing efficiency gains and improving margins in 2019. These cost improvements come about from improved throughput and network efficiency.

Also, we believe the continued adoption of OTT products as well as new revenue streams from Limelight Realtime Streaming and our edge services offerings, will continue to contribute to overall margin expansion for years to come.

Total GAAP operating expenses were $24.1 million, which is an increase of $150,000 from the third quarter of 2017. G&A expense decreased $200,000 due to lower litigation-related expenses; and R&D decreased $600,000 due to lower employee compensation. Sales and marketing increased $900,000 as expected, due to our expanded sales and marketing teams.

Interest income, expense and other income and expense netted to a loss of $100,000 in Q3 this year compared to income of $100,000 last year. On a GAAP basis, we broke even this quarter compared to a $0.02 per share loss last year. Non-GAAP EPS was $0.03 per basic share compared to $0.02 in Q3 2017. Adjusted EBITDA was $7.7 million, up from

$7.4 million in the third quarter of 2017.

Moving to balance sheet and cash flow. We had cash

and marketable securities of $52.6 million (sic) [$52.7 million] at the end of third quarter, up from $45.6 million at the end of Q2. This was our strongest quarter ever for cash generation as we reported over $11 million in cash provided by operations.

DSO as of September 2018 was 49 days, 6 days lower than Q2. We had really strong cash collection performance in Q3. We normally expect DSO to be in the range of 50 to 55 days.

As of September 30, we had approximately 113.2 million shares outstanding. As a reminder, we would have an additional 8 million shares in our fully diluted share calculation. Total employee count at the end of the quarter was 551, up 2 from the end of last quarter and up 16 from third quarter of last year. The increase relates to additional R&D and sales personnel.

Moving to guidance. We believe revenue for the full year will be between $200 million and $203 million. We continue to expect gross margins to increase by more than 150 basis points over 2017. GAAP EPS is expected to be between $0.12 and $0.14 per share, and includes the positive impact of approximately $0.12 per share from the litigation settlement with Akamai that concluded in April. We expect non-GAAP EPS to be between $0.14 and $0.17.

Adjusted EBITDA is expected to be between $35 million and $37 million, and we now expect capital expenditures to be under $18 million, down from our previous guidance of around $20 million. We expect to issue our preliminary 2019 guidance in December as we near completion of our budgeting process.

Let me just say a word about the evolution of our guidance and my observation of the analyst estimates. Early on, with wide business variability, we naturally gave conservative guidance. Now with increasing predictability and confidence, we are giving guidance closer and closer to actual expected performance. On the other hand, based on the history of meet, beat and raise, the analysts are closer to the higher end and sometimes outside the range of our guidance.

This is my observation and I just wanted to take a moment to share this with the broader investment community. Last quarter, we updated our long-term goals to reflect the positive business momentum we are seeing in our core delivery business and the strategic initiatives we have been focusing on as well. We raised our long-term annual revenue growth rates to 15% and guided gross margins to be in excess of 55%. Let's take a moment to talk about the initiatives we believe will drive our long-term revenue goals.

The earliest contributions are also from the smallest opportunity pools. We have talked about our agreement with Tencent. We continue to have a constructive two-way business relationship with them. It's healthy and it's growing. The next opportunity is around the Limelight Realtime Streaming offering we just announced.

We believe the market for these services is bigger with a financial impact visible in the medium term and should be more consequential. Then the Ericsson announcement from earlier today has the potential to be even bigger, but will take time to fully realize.

Lastly, the market for edge services' effort is clearly the biggest opportunity but also the furthest from realizing its full potential. Together, we expect these efforts to contribute million of dollars of revenue next year and even more in the years to come, all at healthy margins, with longer-term deals and lesser price compression. We believe almost 1/2 of the future growth target of 15% could come from these initiatives.

It is also refreshing to be talking about these growth catalysts and their anticipated contribution to growth in improving profit profile, a sea-change from the conversations from just 12 to 24 months ago.

To that, add the long-term secular trends and the growth in Internet traffic. We have very little FX exposure and are somewhat insulated from trade wars. The risk of contracting economy or rising interest rates is minimal, and new markets beyond the developed world are opening up.

This is a good industry and our position in it is getting better. The trends look intact and appear insulated from popular concerns. There is a lot to like as Limelight continues to expand and improve.

Layer on these other strategic initiatives to our core business and we feel very good about the growth and value-creating opportunities at Limelight.

None of this happens without a singular focus on customers, the meaningful contribution of our employees, and the support of our shareholders. And for that, we are very grateful.

With that, we'd like to open the call up for your questions. Operator?

(Operator Instructions) Our first question comes from Mark Kelleher with D.A. Davidson.

To start, just a numbers' question. Were there any 10% customers in the quarter?

Just one. I think we've been saying all along for a few quarters now that Amazon has been a 10% customer and it was again this quarter.

Okay, great. And then you made a comment about reserving capacity for your best customers. Where does your capacity utilization stand right now? And can you kind of map that into this new opportunity at Ericsson? Is that going to require more capacity, more CapEx? I know you just took your CapEx expectation down a couple of million, can you kind of tie in your CapEx expectations with your need for capacity?

Mark, it's Bob. I'm going to make a couple of comments and then Sajid can add his thoughts. And so from a capacity standpoint, we are continually adding capacity, including this quarter, despite those customers -- moving those customers away. And we do that both through the purchase of new equipment, but also through software optimization, which is something we're now probably in our third year of doing. And so we're adding capacity all the time, and we're spending that CapEx principally in 3 ways: grow existing locations, although a lot of that comes from software today; refreshing technology, taking out old, putting in new; and then adding new locations, which we've done more of in the last, say, 12 months than in the previous 3 years combined. And so we're continuing to do that. The Ericsson opportunity is a really interesting one for us, because relative to our largest competitor, we've always been in independent data centers, which has its advantages and its disadvantages but they've been heavily embedded inside the carrier network. With the Ericsson opportunity and partnership, we will now be in 40 places to start and it will take us some months to ramp into that. But we'll be inside a carrier network in 40 locations, many of which are places today where we don't have a lot of capacity, and so it's really going to help us from that perspective. So always adding capacity to our existing sites, always looking to add new -- Ericsson basically leapfrogs us into many locations that are -- or have been harder for us to get to. And at the same time, creates an opportunity for us to be inside the carrier network. And obviously, saves us money because those are locations that Ericsson already has built out and we just need to put our software in to take advantage of the hardware and the capacity that already exists.

I think Bob you covered it all, except for the fact that I think you had a question around utilization. And while we don't give utilization numbers, Mark, it's because it's not like a factory where it's running and then you're selling the goods the next day or 2 days or 3 days later and you're looking at factory utilization. We have a time of day we build for the peaks. And at peak, you kind of look at what the capacity is at and how close to the peak you're getting. But at 2 in the morning, there may be very little usage of the network, and you follow the sun and where -- and what time of day it is. So capacity utilization goes down a lot in Asia-Pac when it goes up here in North America. And so there are just all kinds of dimensions around that, and we try to maximize utilization. That's what I'd tell you, hopefully that answers your question.

Okay. That does, I appreciate that. And then just one more on the Ericsson situation. When does that -- what's the timing of deploying that and starting to ramp revenue there?

Yes, we'll probably convert the first PoP in the next couple of months. And my best guess is it will take us until mid next year to get the initial project completed, meaning, taking all of their existing PoPs and then they want to aggressively add in more locations. So I think in the second half of next year, we'll see meaningful revenue from that partnership.

The next question comes from Jon Charbonneau with Cowen and Company.

In terms of traffic, can you talk about the trends you saw from the fixed customers that saw repricing largely go into effect in July 1, and was that in line with what you're expecting when you did reprice them?

Yes, so all of those customers are growing customers and have been. Obviously what changes from one day to the next is the price. But the traffic, it's not like we went in and bought traffic where we signed a contract and it artificially went up because we were being held back by price or something like that. So traffic continues to grow with all of those customers. Pricing, the contracts were renewed because we want and the customer wants a long-term relationship to be kept in place. And so while pricing changes as you sign a contract, and it takes a while for traffic to build up to the point where you're revenue neutral and then can start growing again, all of those customers are growing traffic with us.

The next question comes from Sameet Sinha with B. Riley FBR.

A couple of questions. First, going back to the Ericsson question, can you give us a little more detail, are you going to be the exclusive CDN there? Can you talk about how the sale is going to be happening? I mean, is it going to be Ericsson's going to be marketing your services and such? And the reason I ask is because Akamai has had a couple of hardware relationships in the past between Riverbed and Cisco and they haven't gone too far. So I just wanted to understand how this one is different. And secondly, if you can talk about those 2 customers that you decided to let go, can you talk to us about what kind of traffic, what industry? And if you can also tell us when exactly in the quarter did those -- did that traffic go away? I'd appreciate it.

Sure. Let me talk to the Ericsson because I think it's very different than the examples you cite on the Akamai side. So the way this is going to work is it is exclusive. They are replacing their current software with ours exclusively. So as soon as a PoP converts, we are going to be able to push traffic into that PoP, which will provide a cost and a performance advantage to us. So that's good immediately for us. Based on having that quality and that capability and reach, it should help us on the revenue side. But it will also enable Ericsson who has been marketing and has customers on the UDN platform, to not only continue -- the plan is for those customers to stay on the UDN platform but now on our software and enable them, based on the capabilities that we're bringing to the table, to more rapidly grow the customer base that they're selling CDN and edge services to. So I think this is not just a -- it's not just a, we think this may happen, it's going to happen, right? We will have that capacity as soon as we convert it and as soon as we do, it has immediate benefit to us. So we're very pleased that Ericsson chose to partner with us on that. Does that answer your question on Ericsson, Sameet?

Yes, fair enough.

On the 2 customers, as you know, we don't really talk about individual customers, but they were -- there are certain types of traffic that, because people are less interested in quality, tend to be much lower priced. So it was video traffic, but very low-priced and content that quite frankly we're not that interested in carrying. There's just a lot of risk in what can happen with that traffic in terms of some of it's user-generated and there are far less controls over the quality of the content. So -- but mainly it was driven by financial issues. We're very good at measuring down to the customer level -- within the customer level, location. We're very good at modeling and measuring gross margins and these particular customers were negative, almost every location where they delivered -- where they deliver traffic. So when you're not that happy with the type of content that is -- your network's being used for, and you're not taking any margin on top of that, other than the revenue hit, it sort of makes it an easy decision. One of the 2 customers was gone coming into the third quarter, and one is gone coming into the fourth quarter, which I think was also part of your question.

The next question comes from Tim Horan with Oppenheimer.

You've got a lot of opportunity in front of you here, with these 4 different major initiatives. Can you give us a sense, is this kind of doubling maybe the size of the market that you're going after, tripling? Or is it 50%? It's just so hard for us to know kind of from the outside. And I guess, related to that, this 15% long-term revenue growth, is that something we should see in 2020, 2021? And can we start to see it in the second half of next year with all these initiatives? Just a little bit more color behind the growth.

Let me talk about the initiatives themselves and then Sajid can comment on the financial side of it. So some of these initiatives are near in. So for example, we signed an agreement with Tencent, earlier this year. We've started to utilize their network in China for our customers midway through the year. They have started to use our network for some of their customers as of just very recently. So as we go into 2019, I think that is going to be a contributor for us. So bad news is it'll take us 4 to 6 quarters before it's meaningful. But the good news is it's not years. I think Ericsson is going to be similar. I think as we look at -- maybe even quicker, when we look at the second half of '19, we're going to start to see some meaningful opportunities from that relationship. Neither one of those really changed the market size that dramatically because it's helping us deliver higher quality in more locations to our existing customers, and customers that we would be targeting regardless of whether we have that capability or not. So the market size doesn't change. Different story in edge services, edge compute, where the -- it's longer-term. So it reminds me a little bit, in 2014 and '15, some people were talking about OTT like it was going to happen the next day, and as we know, really starting to build measurable momentum in 2017 and '18. So it was a couple of years longer. I think with edge services, it's going to take a little bit of time for that to emerge. All the analysts that are looking at it and reporting on it, project it to be a huge market. So what portion of that, that we can reasonably achieve, remains to be seen. But I view that as opening us to -- up to a market that, in a few short years, is going to be very large. But I think partnerships like Ericsson, Linode and the capabilities that we have within our own infrastructure today, could add to material growth. We're not -- in order to get to that 15%, we're not counting on having a large share of edge compute, for example. But we are counting on Tencent continuing to grow, Ericsson continuing to grow and be meaningful. And so that's how we kind of see it from a market-size perspective. Sajid, do you want to comment on the financials?

Yes. I mean, clearly the market opportunity is much bigger. If we think about the addressable market or TAM in the base business, around $2 billion, this is many times the size. So collectively these opportunities open up a lot of room for us. When I look at the revenue itself, these opportunities, new initiatives or, call it, the place where we've invested, a couple of things to keep in mind: one, very little CapEx; two, almost accretive and profitable right out of the gate. And if I think about it in very, very simple terms, this is the business that was $1 million or $2 million last year; grows to $4 million or $5 million this year; grows to $8 million to $10 million next year; and grows to -- something, $16 million, $20 million the year after. That's kind of the progression that would make sense, kind of a doubling of the business again and again and again for the foreseeable future. And what it does for the growth rates is it takes us from 7%, 8% to 9%, 10% this year and then all the way steadily moving up to a 15% growth company. So I don't think that we are talking about getting to a 15% growth rate next year. But we are talking about a 15% sustained growth rate in about 3 years and moving up to that number over that course of time. And rest assured, everybody inside the company is going to be working very hard to get there as quickly as possible. But I think if you think about the progression, that would be logical. And again contributing organic, profitable with a very different price point, with a very different turnover, with more sticky revenue, with less price compression, with reuse of infrastructure, like, there's a lot to like about each one of these opportunities.

No, that's really impressive. And just to be clear, those numbers you cited were for all 4 services, right? Not just for edge?

Correct. I'm just kind of collectively pulling them together and just giving you a sense of what we think that this business -- I think we would want all of them to do very, very well. If that happens against schedule, we'll get ahead. If all of them do very poorly, we'll fall behind. But if you just kind of risk adjust where this thing lies, I think that should be the right progression.

And a just sort of last question, real time video, can you just describe what you've done from an engineering perspective to improve that? And at this point, do you have a lower market share in real time video, and is this meaningful enough, do you think, to gain pretty good share because you're talking about a hugely growing market in real time video?

Yes. So let me comment on that from an engineering perspective, with the caveat that I'm not an engineer. Today at best, live is about 30 to 40 seconds behind broadcast. And there are a lot of reasons why that is, there are 10-second chunks of video that get transmitted. And most players, like Roku or whatever, will wait to get 3 chunks, so that's 30 seconds, and then start to play back. And so that's the -- kind of the reasons why it is where it is today. But there's a growing frustration amongst consumers. We've seen stuff around US Open Tennis and the World Cup, and, sort of, it's no longer good enough to just deliver at good quality. People want it delivered at good quality near real time. And so what we did was we've taken a fairly well built out technology called WebRTC and we've worked to integrate that into our CDN, so that we can deliver anywhere in the world from anywhere in the world, at less than 1 second of latency, but -- and so that's really important. But I think even bigger is because WebRTC allows you to utilize a video channel, an audio channel and a data channel, you can do things like have video and audio on separate channels. So for some reason, you're on your mobile phone and you're in an elevator and you drop the video, you'd still have the audio. But also to be able to have a chat room right on the same screen, to be able to place bets real time on whether the next play is going to be a run or a pass. So it gives us a lot of flexibility and a lot of things we can do with the real estate on the screen. And we demonstrated that at IBC. So the people that are -- when we talk to people, the people that are, "Hey, when can we have this implemented?", tend to be in the gambling space. The next higher sense of urgency is in gaming. And then from sports broadcasters, that's really interesting, but we need to do some testing. We need to understand how we would use those other channels. And so everybody finds it interesting and exciting. Sense of urgency depends on which markets you're in. And so we're excited about it. We hope to go into production with the first couple of customers by the end of the year, not that we hope -- we plan to go into production by the end of the year. And I think once we get a few people into production, as you said, the market for live video, especially live events as opposed to live linear, is very, very big and fast growing. And you look at companies like DZone out there and Amazon Prime Video and others that are streaming live events that 2 years ago didn't do any of that, right? So we're seeing more and more of the viewing move to Internet-enabled devices by companies that only utilize the Internet like DZone and Amazon Prime Video, as opposed to traditional broadcasting. So we're excited about it.

The next question comes from Mike Turits with Raymond James.

Bob and Sajid, 2 questions. One, on edge computing, there was a conference about it this week, I believe you guys participated. Can you walk through what you're doing to upgrade your capabilities so that you can provide compute on the edge, including, I think, probably going from just kind of basic infrastructure as a service up to something that enables more application logic? That's the first question, and then I have questions about the loss or termination of those 2 customers.

Okay. So right now from edge services, edge compute standpoint, a big part of our strategy is having virtual machine capability, which is why we did the partnership with Linode, it's -- the partnership we did with Ericsson to get inside the carriers' network and as they build out 5G to have assets that are inside the network. But from a software perspective -- because having infrastructure is nice, but from a software perspective, we're working on something we call function as a service, which will enable our customers to develop, sort of mini applications that they can push to the edge and have the workload be done at the edge as opposed to back in the cloud data center. So it's evolving. It's new. We're learning as we go, as the industry is. And there are a lot of different definitions for this, as you know. We will hear this week, I'm in New York City today with Sajid at the conference talking to many customers. So it's taking shape, and all of these things that we're doing are meant to help build out the ecosystem around this. But from a software perspective, function as a service is our way of bringing to life the customers' capability in terms of having the application reside in a different place than it does today. I don't know, Sajid, is there anything different?

I think that's right on.

Yes.

And then a question on the other -- on the customers that you lost. So you said that they were not -- I think you said that they were not profitable customers. Is that a change? Because I mean I can't imagine -- I mean, you've done this before. You've gone through periods where you've gotten rid of customers that were not so profitable. Is it that the pricing went down? Or did they suddenly become not profitable? Also I think you said that they -- you got rid of one going into the third quarter, so I would've thought that would've been visible from the last guidance. And then is that what accounts for the less than seasonal guide into the fourth quarter?

Well, we met the last guidance. We obviously knew that was happening and gave the guidance going in that we did. And so these customers have never been that profitable. But as of late, we have seen unusual price compression in that area, as best I can tell, as our competition looks to fill up idle capacity. I don't know why else you would do that. But we were not going to go -- we would have rather have lost the business than have handled it at negative margin. And it just makes room for us to expand with our customers that -- where quality matters, and where quality matters we feel much more comfortable competing.

Just to add a little bit here. Like, the guidance that we gave anticipates some of this going on. But we are always looking for the most efficient way to grow our revenue, but not also have to borrow from the bank to grow the revenue. And that's what this puts us in the position of having to do. So, I mean, it's a customer loss, for sure, but this revenue and these customers are available. They're just available at a price point where the gross margins are -- just don't even make sense. And we do not want to build out capacity and take our precious dollars to go build that capacity to serve customers at the margins where we lose money. So those opportunities exist in the marketplace. They may be beneficial to some that are trying to do something with their business, or others that have excess capacity. But we don't find ourself in the same position. And we see good opportunities with our base of customers and other customers that are coming on board, that it makes sense at the right time. It's always a hard decision, right? At some point, you have to go make these decisions, and there's never a good time to do it. But at this juncture, where you have enough good revenue coming, on, on the other side to offset this, seem to make the most amount of sense, so that you're still on the trajectory of the revenue. At the same time, you continue to improve your margins. You manage your business accordingly and you go ahead and make these choices. So factored into the start of the year that here is a kind of churn we might expect, here's the kind of progression we might expect. And when we gave guidance at the start of the year, from that, we raised it twice. And the guidance we gave last quarter, I mean, we are executing against it. So no real surprise to us.

No, we definitely appreciate the discipline, for sure. So just a clarification, so the guide to the midpoint of the full year guidance at this point -- the implied guidance for the fourth quarter, based on going to the midpoint of the full year guide, gets you to about 1% sequential which is below seasonal. So should I understand that below seasonality as a function of having gotten rid of the second of these 2 customers?

Yes, a couple of that and then the renegotiation for these...

Yes, the renegotiation -- the contract that I said, when we talked about that last quarter that they hit -- they'll hit throughout the third quarter and they had different effective dates. Some hit 2 of the 3 months, some hit 1 of the 3, some hit all 3. So fourth quarter is the first full quarter, where the full effect of that pricing is in place.

But there's multiple ways to look at this. I mean keep in mind, we used to have seasonally dips in the third quarter, right? So you have Q1, Q2 pretty strong, Q3 drops a lot from that, Q4 jumps up a lot. Now we are reporting 3 quarters that are roughly equal to each other. So it doesn't look like, okay, the fourth quarter isn't moving up, but the third quarter is much higher than what it normally would be, right? We've dropped significantly. So looking at what you're comparing against is equally important as looking at the outcome at the back end of that process.

(Operator Instructions) The next question comes from Jeff Van Rhee with Craig-Hallum.

Bob, Sajid, several from me. If we could just go back to the real time video for a second, I think last quarter to another quarter, you were commenting to, I think, a half dozen or so proof of concept customers. And I believe based on their indications that a lot of them wanted to go live quickly. I think you commented today that you're thinking you'll maybe have a couple live by the end of the year. Is that in line with what you expected in terms of the time line of go lives on the real time video when you commented last quarter? And then maybe just if you could take a second to expand on what's happened to the pipeline, obviously coming out of the launch and the publicity related to it.

So we -- to be clear, the software is going to go into general availability, GA, in about a week. So customers are still working with [beta] versions. And so the pipeline is pretty strong. But even the 2 customers that have signed on to go into production, they're just getting access to the GA versions either this week or next week. And so there's a healthy pipeline, healthy level of demand. We feel good about where we are, anxious to get it into production and get some proof points in the marketplace. But -- so that's where we are.

Okay. And then just with respect to the 2 customers that you took a pass on the business, did you see similar -- the competitor that in that case chose to take that business to essentially fill the pipes, did you see other instances of that behavior in, maybe greater frequency? Or any other suggestions elsewhere in head-to-head business that, that was a much broader theme that they might have accelerated?

Not really. I mean, look, there's always price compression. There's always aggressive pricing to win new business. It's a tough competitive marketplace. But I don't -- I wouldn't view that as a broad-based phenomenon within the industry.

And I think just being -- I know you're not new to the story, but new with the coverage. But we've talked about on our calls for some time that there were times, in the absence of quality, where we have competed on price and price alone. As we moved our quality up, we've kind of warranted a higher price, and gotten it from our customers. And they appreciate the quality and all. So -- but there's always a low price provider in the marketplace. What has been surprising is who that has become, because for the market leader who has traditionally commanded a premium, to be kind of at parity, that is surprising. On the other hand, at some level or the other, you always expect somebody to be available, capacity, fighting for that last customer, last dollar. And we feel pretty comfortable about the position that we are in and the price that we're commanding and the feedback that we're getting from our customers. Here at the conference this week in New York, in the conversations we had at IDP -- IBD -- there's been some -- IBC, there has been good conversations around all of this stuff and just real interest in what we have to bring to the market, not just from the standpoint of price, but from the standpoint of feature functionality and quality and availability and footprint.

Got it. Then just 2 other quick ones for me. Expand just very briefly the reduced CapEx, what improved there? And then secondly, Sajid, on the guide, at the end of your prepared remarks, you gave a bit of a qualifier, a bit of your sort of perspective, if you will, on The Street versus the guide, of even some outside the expectation. Just revisit that again just to make sure I understand that.

Sure, a couple of things. Let me do the guide first. We were guiding at closer to like a 70%, 60% confidence interval, right, because: one, having taken on that position; two, the business being fairly variable, the outcomes being very wide and not just having enough confidence in our ability to kind of really hit the numbers, which you really don't -- I mean, we know going in that the penalty for not making it is very high. And so the way you offset that is you guide conservatively. And we were doing that for a period of time, and we were doing that leaving ourselves some room where you want to be sure that you're going to hit the numbers. I think good business becomes more predictable over time. And that's what's happened to our business. Our business has become better and more predictable as time has gone on. The variability we see in our G&A lines is within hundreds of thousands of dollars. The variability we see on our FX is fractional. What we can predict about our business going into the future, it's become more reliable. And so as it becomes more reliable, you move the confidence interval up from 60 to 70 to 80, maybe even get closer to 90, because you feel going into the quarter that you -- here's what you know about the business and this is what it should be able to do. So that's what we've done in terms of our guidance. I think, early on, because we've never ever met a number that we gave to The Street, The Street would kind of believe us in what we said and stay pretty close to the guidance. And now with a series of meet and beats, I think some people now reach to the outer extremes of where we guide, some are even outside the guidance. So I just, kind of, want to remind everybody that as we think about the business, we want to execute against the guidance. And we feel pretty good about what we've done there, not just this quarter, not just this year, but over the course of the last 8, 10 quarters. So that's on guidance. Your other question was?

CapEx.

It was around CapEx. So future CapEx, I think, we kind of suggest 10% to 12%. A deal like Ericsson clearly takes pressure off CapEx, so that helps. Current CapEx, when you get rid of -- I mean, you have a couple of choices. You're going to go ahead and build more capacity to serve customers where you're going to lose money, right? And instead of reporting a $50-plus million cash balance at the end of the quarter, you report something in the 40s because one, you lost money with the customers; two, you deployed capital trying to get there. Or you run the business like it should be run, you show up with some money at the back end with -- in the bank, for everything that you do, and relieving these customers freed up some capacity. So we have software implementations and gains that make capacity available and the availability of capacity because of the churn in the customers, all of that helps us. And as a result, we want to take that to the bottom line. So we show up with some very strong numbers in terms of cash in the bank, as well as the continuing improvements in the profitability profile. I mean, we are talking about 150 basis point improvement in gross margin this year after in excess of 500 basis points last year, in excess of 200 basis points the year before. And I think we are on track for a few years of 100-plus basis point improvements if we're going to get from where we are today to our stated goal of 55% gross margin.

This concludes our question-and-answer session. I would like to turn the conference back over to Sajid Malhotra for any closing remarks.

All right. Well, once again here, I mean, we are pleased with the quarter that we delivered. It was another good quarter. I think the progress against our long-term goals and strategic objectives is evidenced in the recent news flow. We are excited about the future. And with that, I just have to thank you again for joining us for this call. We are always available to answer any additional questions, and thank you, operator. And with that, we'll conclude the call.

The conference is now concluded. Thank you for attending today's presentation. You may now disconnect.