Thank you. Hello everybody. Welcome to our full year result conference call for 2019. Thanks for taking the time to join us today, and thanks for your interest and support of Appen. My name's Mark Brayan. I'm the company's CEO, and I'm joined today by our CFO, Kevin Levine. We'll step through the presentation that was provided to the ASX this morning, and then we'll allow some time for questions later in the call. If you could please turn to page three. We're very pleased to report another high-performance result for shareholders. A total revenue of AUD 536.0 million is up 47% on last year. This is comprised of AUD 67.7 million of speech and image revenue, which is up 32%. AUD 430.0 million of relevance revenue, which is 37% higher than last year. Year-end Figure Eight ARR of AUD 33.7 million.
This returns Figure Eight to high growth and lifts its 56% CAGR over the last four years. Underlying EBITDA for the year was AUD 101.0 million, an increase of 42% on last year. Underlying EBITDA margins of 18.8% compare favorably to last year's 19.6%. That's considering investments made in 2019, most notably in Figure Eight, but also in technology in general. EBITDA margins were consistent with our commitment to keep margins in the mid to high teens concurrent with investments for scale and strength. Stripping out Figure Eight reveals very strong organic growth, highlighted on the right-hand side of the page. Organic revenue, that is revenue excluding Figure Eight, at AUD 497.6 million, was up 37% on last year. Organic underlying EBITDA of AUD 107.3 million was up 51%, outpacing the growth in organic revenue.
Organic underlying EBITDA margins at 21.5% are well up on last year's margins of 19.6%. The organic results show the operating leverage of the business derived from our investments in technology and economies of scale. Turning to page four, we can see that each of our core divisions contributed to the result. Speech and Image, formerly Language Resources, includes the collection and annotation of audio speech recordings, text-based natural language data, as well as image and video data. Revenue growth of 32% to AUD 67.7 million is a standout result, considering a historic revenue CAGR of 17.9% since the company listed in 2015. The result came mainly from multiple projects across our major customers. Margins were impacted by a few large projects peaking in the first half and trade-offs in gross margins for sales expansion.
To the right-hand side of the page, relevance continues to be the company's growth engine. Revenue of AUD 430.0 million was up 37%, its EBITDA up 66% to AUD 110.5 million, delivering very pleasing year-on-year margin expansion from 21.3% to 25.7%. It's also noteworthy that one of our largest customers recently updated their agreement with us to include a clause to hold prices for five years, mitigating margin pressure. Please turn to page five. We're very pleased to report that Figure Eight has rebounded to high growth. Year-end ARR, that is annual recurring revenue, of $24.95 million or AUD 33.7 million, lifts the business to 56% revenue CAGR from 2015 to 2019. Sorry, that should be ARR CAGR from 2015 to 2019. The result benefited from a strong second half that included new customer wins, new expansion deals, and improved churn management.
We also saw an increase in larger customer wins. This is important because large customers are less likely to churn than smaller ones. The path to profitability for Figure Eight is ahead of plan due to prudent cost management of non-essential spend that has no impact on sales or growth. We estimate an earn-out payment of AUD 36.8 million, bringing the total price of the acquisition to AUD 287.6 million or 5.8x revenue. This compares very favorably to similar acquisitions that are valued at double-digit revenue multiples. To page six now, we have commenced the integration of Figure Eight in earnest now that the earn-out period is complete. We are moving to a simpler, unified product set, updated messaging, and a fresh visual identity. We're going to market with a single sales and marketing team, consolidating our crowds and shifting work to Appen's crowd and facilities.
We're also combining teams, culture, and back-office functions. Synergies discussed at the time of the acquisition are on track, and we expect the major components of the integration to be completed by the end of this year. Our customer cohort chart on page seven benefits from the addition of Figure Eight's customers and continues to show the strength in our customer relationships year-on-year. For those unfamiliar with the chart, it shows revenue by cohort in year of origination and each successive year thereafter. We're also making good progress in our new growth markets. Please turn to page eight. We're investing in two growth market sectors, government and China, to add customers and strength to our business. Governments worldwide are investing in AI through their defense intelligence and civilian agencies for a variety of applications.
We added a new government customer in 2019 that's using AI to assist disaster relief, for example. We've participated in the government market for some years now, but deeper involvement requires us to be set up with appropriate clearances and capabilities. We're midway through this process and have leased an office in Washington, D.C. We expect the process to be complete soon, and it will give us a significant differentiated capability. To page nine now. Our work in China is progressing at speed. With 53 people in our corporate office and 154 data labelers in our facility in Wuxi. Our pipeline is growing predominantly with business with technology companies, and we're delivering growing revenue, albeit off a small base. Thus far, we've seen negligible impact on our business from the coronavirus outbreak.
All of our staff are safe and accounted for, and we've implemented a work-from-home policy that will remain in place until conditions improve. We've also suspended all travel in and out of China. Our projects with customers in China are currently ongoing, and we don't foresee a letup in that work, but we are monitoring that situation carefully. Our 2020 domestic targets for China are modest, and hence any impact will have a negligible impact on group revenue and earnings. Of our major U.S. customers, combined, they rely on China for less than 10% of their revenue. We don't expect their business with us to be materially impacted by any loss of their revenue from China. Some of our U.S. customers rely on China for their hardware supply chain and have signaled supply challenges.
We're not seeing any impact of this on our business, nor do we expect temporary hardware shortfalls to reduce their need to invest in AI. It's fairly early in the situation and we're monitoring it closely. Overall, we maintain a view that the coronavirus will have a negligible impact on our 2020 group revenue and earnings based on currently available information. Please turn to page 10. We continue to invest in technology to improve the interface with our customers, manage and communicate more effectively with our crowd workers, and improve worker productivity through better annotation tools.
We've released a number of new features during the year, including a secure workspace solution that improves security for at-home workers, our automated checking of crowd applicants that reduces the time taken to validate a worker from up to 20 hours to a few minutes, and machine learning-assisted tools for transcription, data collection, and other applications that are showing productivity improvements in pilot projects. On page 11, we're also pleased to announce a significant enhancement to our capabilities in the autonomous vehicle market with the launch of our lidar annotation tool. Lidar is similar to radar but far more accurate. It generates 3D models that are critical for AI applications that move in space, such as autonomous vehicles. Autonomous vehicles and other AI applications using lidar form a rapidly growing market that relies on vast data sets.
Working with and labeling data in 3D is complex and requires sophisticated tooling and a trained workforce. We have a number of pilots ongoing in this space and are investing in business development to win more. Page 12 now, and an update on our ESG work. As a people-based business, we have a low environmental footprint. Our work-from-home model enhances this. That said, we aim to improve and are buying carbon credits to offset our 2019 travel and will continue to do so in the future. We've made solid strides in our commitment to the ethical treatment of our crowd and social impact through our Crowd Code of Ethics, membership of the Global Impact Sourcing Coalition, and our work with Translators without Borders. The latter is a not-for-profit organization that assists people in need at borders with interpretation and translation support.
We've also implemented an internal development program to improve gender balance at the senior levels of the company. On governance, we are adding a U.S.-based independent non-executive director to our board to improve board independence and further our knowledge of U.S. markets. I'd now like to hand it over to Kevin to take you through our financial results commencing on page 13.
Thank you, Mark. We are pleased to report that we continue to translate strong revenue growth into efficient and meaningful bottom-line growth. Our revenue increase of 47%, including Figure Eight, has been driven largely by continued increase in demand from current projects with existing customers. Growth has also come from adding Figure Eight and new projects with existing customers, as well as sales to new customers. Underlying EBITDA, including Figure Eight losses, is up 42%, driven by delivering increased volumes and improved margins in relevance and achieving the benefits of scale, process, and technology. Excluding Figure Eight, organic performance is strong, with underlying EBITDA growth of 51% on organic revenue growth of 37%. Underlying NPAT is up 32% on the prior year, impacted by the investment in engineering and the result in higher levels of amortization.
The effective tax rate for 2019 is slightly better than last year, it's impacted by tax deductions of employee performance rights. These can vary from year to year. If you can please turn over to slide 14, we'll talk through on the balance sheet. The balance sheet is strong, with increased activity translating to higher cash and receivables balances. Intangibles have increased significantly with the acquisition of Figure Eight. Impairment testing revealed there is significant headroom while using conservative inputs. Appen is debt-free at year-end, the robust balance sheet contributed AUD 35 million out of the total debt repayment of AUD 57 million during the year.
An estimated earn-out liability of AUD 37 million has been booked in respect of Figure Eight, payable next month. Appen has declared a final dividend of AUD 0.05, partially franked to 50%. This is up 25% from AUD 0.04 in the prior year. Over the page to slide 15. Cash at year-end has increased by AUD 35 million to AUD 75 million. In fact, the cash balance would have been higher if not impacted by some timing issues with receipts at year-end. These issues were resolved in January. Strong performance this year has seen cash flow from operations increase by 47% to AUD 83 million. Cash has been effectively deployed during the year to pay debt, dividends, CapEx, and transaction costs for the Figure Eight acquisition. Even with the receipt timing issues, cash conversion from EBITDA was strong at 88%. Over the page.
We call out that we have currency translation exposure given that most of our transactions are in USD and we report in AUD. In 2019, we benefited in a meaningful way from a weak Australian dollar. This contributed to 10% of the revenue growth and 11% of the underlying EBITDA growth. I'll now hand you back to Mark. Thank you.
Thanks, Kevin. On page 17, we highlight some important market trends. Firstly, that AI continues to rely on large volumes of high-quality training data. Quite simply, the more data, the better the AI performance. The availability of quality data continues to challenge AI developers. Data supply is a large problem for those starting their AI journey and the largest problem for those in production. This mirrors our own research. It's only when developers and organizations move to production and deployment in the real world do they fully realize the need for high data volumes and data quality. The data supply challenge is exacerbated by the growth in AI, and with it, the growing demand for data. Organizations must invest in AI lest they fall behind their competitors. Over the page, we're very well positioned to benefit from these trends and the need for large volumes of high-quality training data.
We are a proven provider. This is essential because our research also shows that companies that want to deploy AI in the real world want to partner with reliable and proven data providers. We have the two essential assets to deliver training data, the global and diverse crowd, and the enabling technology that allows the sourcing and delivery of training data at scale. Page 19. To our outlook statements. Our work and investments in recent years have strengthened our position in the high-growth AI training data market. 2019 saw substantial investments in technology that are now paying off. 2020 will be a year of investment in sales and marketing to grow our customer base and further cement our strong position. This investment will soften our first half margins to the mid-teens. We expect margins to return to the high teens by the year-end.
We reiterate that we expect a negligible impact from the coronavirus on 2020 group revenue and earnings based on currently available information. Our order book, inclusive of year-to-date revenue and orders in hand, is currently approximately AUD 210 million. Our full-year underlying EBITDA is expected to be in the range of AUD 125 million-AUD 130 million, that's at $0.70 to the Australian dollar. The outlook is, of course, susceptible to upside or downside from factors including timing of work from major customers and Australian dollar fluctuations. That ends the formal presentation. Thank you once again for your interest in and support of our company. I will now hand it back to the moderator to manage questions. Thank you.
Thank you. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press star two. If you are on a speakerphone, please pick up the handset to ask your question. Your first question today comes from Garry Sherriff with Royal Bank of Canada. Please go ahead.
Yeah, a few questions, Mark and Kevin. Firstly, thank you for going through expectations on coronavirus for calendar year 2020. I get everything you're talking about in terms of your tech customers only having about 10% of their revenue out of China. I guess I'm just trying to get a sense of those hidden indirect impacts. By that I mean, have any of your larger tech customers in the last couple of weeks been in touch? Maybe just talk around their tech budgets should consumer or business sentiment become more negative? That's the first question.
Hi, Garry. We've canvassed all of our customers, and there's been no indication on current projects currently. As you suggest, if they suffer a material shortfall in hardware supply, for example, that may prompt some cost control in general. At this point, we're not seeing any indications of that, but we're looking at it closely.
Another way of looking at it, calendar year 2020, your expectations, and I know you haven't disclosed revenue per say, but what portion of your calendar year 2020 is underwritten or under contract?
We disclosed the order book of AUD 210 million and you could look at order book numbers for prior periods to get a sense for what percentage that may indicate for the full year.
Okay. EBITDA margins for Speech and Relevance going forward. I noticed in your guidance you've talked about first half being a little lower. I'm just trying to segregate that out by segment, if we can just get a little bit of color on what we should be thinking for margins for each of those major segments going forward?
Yeah, Garry. Hi. Yeah, Speech and Image, we would guide in terms of for 2020 to be similar lines to 2019. I think for both of those divisions, basically similar, but obviously overlaying impact for investment in sales and marketing and a bit of engineering. Flow that through in terms of from the total overall margin guidance.
Okay. The last two questions, just on Figure Eight. Looked a really big improvement in Q3 and Q4. Two questions. One, is there any form of seasonality that we should be aware of in Figure Eight, and if so, what drives that? Secondly, did you have to sacrifice price or margin to land some of those larger deals?
No. We've looked back over the last few years. There doesn't appear to be any meaningful seasonality or consistent seasonality. Clearly, some quarters do better than others, but there doesn't seem to be any pattern to it. No, we didn't sacrifice price to deliver the Q3 and Q4 results. It was a combination of better sales leadership. Our head of sales, Jon Kondo, took a hands-on approach with the Figure Eight sales team in Q3 and Q4, and it was also due to much better management of existing customers and churn. We saw far less churn in Q3 and Q4 compared to Q2.
Thank you very much. Thank you.
Thank you. Your next question comes from Quinn Pierson with Credit Suisse. Please go ahead.
Hi, good morning. I guess just firstly on the sales and marketing investment you flagged that will pull down the first half margins. Could you maybe just talk us through the timing of that investment deployment? Is the way to think about it that there is a step-up occurring basically now, which will pull down the first half margins, but then the return on that investment can drive sales in the second half, which will, I guess, kind of bring the margins back up. Is that the timing to think about? Thanks.
Yeah, pretty much, Quinn. We're staggering the hiring throughout the year, but there's a lag obviously between hiring and return that plays out as you suggest. By staggering the hiring, that does give us the opportunity to throttle the hiring if the returns aren't coming. In general, what you suggest is correct. We hire early in the year, and that delivers sales and a return on those hires in the second half of the year.
Yeah, thanks. Then secondly, on Figure Eight, could you just talk us through how much of the revenue growth is coming from existing customers versus new customers and what the pipeline of new customers looks like for Figure Eight? In particular, I'm just trying to understand how well Figure Eight is being used as a deployment to, I guess, sell the existing crowd to a new set of customers?
To the first question, it was pretty well-balanced in terms of new customer wins and expansion deals in the second half. Expansion being upsell of existing customers. There was good flow of new customers, and then good balance with upsells to existing customers. In terms of the platform as a platform for further sales growth, it opens up our addressable market quite considerably. Our largest customers have their own data labeling platforms, and we work on those platforms. The majority of companies worldwide do not have a data labeling platform, and they need one. This is what we have now with the acquisition of Figure Eight. We have a much greater addressable market. What you suggest is entirely correct. We'll now be able to access more customers, sell the platform, and that'll have pull-through to our crowd resources.
Is that thesis already playing out, or is it kind of the sales and marketing staff being hired now that can kind of take that to the next level?
It's playing out, but it's early days. We've had some joint customer wins. Sale of the platform and sale of Appen Crowd. Going forward, that's the focus of our sales and development efforts.
That's helpful. Then just lastly from me, just what you're seeing in terms of the competitive environment. We've seen a fair amount of, I guess, investment in various kind of data labeling solutions, especially more on the platform or kind of tech-heavy side of things. I guess, are you seeing some of these competitors as direct competitors to you, or do you see them, I guess, playing in different segments of the market? Any views on how you're seeing the competitive landscape would be helpful, please.
We see them variably in our sales campaigns, but we don't see them in the very largest accounts. We don't see them in our global customers. We're told by our global customers that they rely on our size, our strength, our scale, and our reliability to service them. Yeah, we do see various competitors from time to time, but not in our biggest customers.
That's helpful. Thank you very much.
Thanks, Quinn.
Thank you. Your next question comes from Siraj Ahmed with Citigroup. Please go ahead.
Thanks. It's Siraj. Just a few questions. Just first thing on sales and marketing investment. Kevin, if you could just talk to the quantum of the investment and maybe, Mark, and where are you targeting this as well? Is it the sales and marketing investment, is it Figure Eight? Is it lidar?
We're not disclosing the quantum of the investment other to say that it'll have an impact in margins, which we've illustrated in the pack. In terms of where we're deploying that investment, we currently have good focus on our large customers, we want to strengthen that focus more broadly across the technology market in the U.S. We also want to tackle other market verticals in the U.S., autonomous vehicles being one, financial services being another, health and pharmaceutical being another. We're also able to target mid-size companies very effectively with the Figure Eight platform. We're setting up sales teams to go after each of those broad segments. We're also investing in sales in the U.K. and Europe, and furthermore, investing into sales presence in Asia Pacific outside of China. It's fairly broad brush, Siraj.
We've balanced the resource with where we believe the opportunities are in terms of market size and market stage of maturity.
Got it. Just thinking through the margin, as you said, first half margins at mid-teens, full-year margins at high teens. Is the expectation that margins will be up year-on-year at the full year level? Two, just thinking, should the second half margins be the base going forward? That sort of implies 22%-23% margins in the second half. Should that be the base going forward in 2020?
Where do you get margins in the 20s when we say high teens?
Oh. No, I'm saying for the second half.
Got it. Sorry. No. We maintain the practice, Siraj, of keeping margins in the mid to high teens, while also investing in growth. We're just calling out specifically that early in the year, in the first half, we'll have some sales resource that's going to cost money that will take a few months to deliver revenue. There'll be a softening in the first half. For the full year, we'll return to those mid to high teens.
Mark, just confirming, or just clarifying, do you expect margins to be up year-on-year for the full year? Is it too early to say?
Too early to say.
Sure.
Yeah. Siraj, if you look at the guidance and you look at our numbers, I think you get the answer there. Obviously, I think it's consistent with our theming, certainly, where we basically strive to get efficiency in our margins and then redeploy it back into the business. In 2019, we did that with engineering. We continue to do that in 2020, but to a lesser degree, because we've established that base. 2020 now is reestablishing our base in sales and marketing. We're doing that and obviously going to the future, we'll identify other areas of growth where needed and obviously look to redeploy where appropriate.
Sure. Very clear. Just lastly, on the renegotiated deal, pretty positive keeping pricing flat for five years. Could you just talk to any commitments and volumes? Have they consolidated the providers, as you said, for the large customers Appen's seen as a preferred supplier? Have there any commitment on volumes in that deal, please?
No, there's no commitments to volumes. The only commitment is to keep prices flat over the next five years. As you say, that is positive because it mitigates some margin pressure, but there's no commitment to volumes.
All right. Great. Thanks for the update.
Thank you.
Thanks.
Thank you. Your next question comes from Josh Kannourakis from UBS. Please go ahead.
Hi, Mark and Kevin. Can you hear me okay?
Hey. Hey, Josh.
Cool. First question, just to follow up on that five-year deal. Are you able to say whether that is your largest single customer, that deal, or just a large customer?
No, we can't. We've made the statement, and that's it.
Yep. Got it. Just to clarify, so is that holding prices from FY 2019 levels moving forward?
Yes.
Great. Perfect. Just following on from the margin commentary before, so just to be 100% clear, I think it was pretty clear in the outlook statement, but the high teens margin is for the full year 2020 result, not necessarily just the second half.
Yes.
Got it. Then just in terms of, I guess, moving forward, do you expect sort of that cadence in investment in sales and marketing to continue into 2021, et cetera, or is this a sort of more one-off in nature?
Probably more one-off. We'll continue to invest in sales for growth, but like last year, we made a big investment in technology. This year, we're making a big investment in sales and marketing to build that base. There'll be ongoing investments that'll be more incremental in nature going forward.
Perfect. Then just government contracts. Historically, we've talked to that being a pretty significant opportunity, and I think you mentioned some help around disaster relief there. How are you going in the progress of getting a sort of prime contractor license in? Do you think you'll have any progress on other departments like Defense, for example, within this calendar year?
We're close to getting things all set up and complete. We expect that to be finalized in Q2, and that'll provide us with the opportunity to bid on business as a prime contractor. We have a solid pipeline of opportunity, and we're optimistic that that'll play out and deliver meaningful revenue this year. Of course, we've got to win the business.
Yep. Of course. Are there many other prime contractors providing annotation services that you know about in the market at the moment?
Not specialists like us. We're a sub to somebody at the moment. You could argue that they're providing labeling services, but we're actually doing it.
Yep
we're unaware of any other specialists.
Okay, perfect. Just one really quick final one for Kevin, just around the AASB 16 impacts, Kevin. I think in the second half for the right of use depreciation and interest, it implies about sort of AUD 3 million for the second half and AUD 5 million was the full year. Is that sort of a run rate we should expect into the future? Or can you guide us to any changes you expect from that?
Just to kind of link it through, when we released our 2018 report, we talked about an expected EBITDA impact of one and a half million for 2019. What happened, we were pretty close to that, but we actually brought on some new premises in 2019. Sorry, that impact we've factored into our guidance. In addition to that, we had, I guess, additional rental, which obviously didn't flow into EBITDA, and that was for new premises. Obviously, it was new. It wasn't factored into the guidance and obviously it wasn't factored into the numbers. Essentially, your numbers are in the ballpark.
Yep
Yes, would be a good proxy going forwards.
Just to confirm, the second half run rate is a reasonable proxy for going forward?
Yeah.
Great. Thanks very much, guys.
Thanks Josh.
Thank you. Your next question comes from Stella Wang, a private investor. Please go ahead.
Hi, guys. Thanks for taking my question, just a couple of them. Firstly, following on the contract towards a big customer, are you looking to do similar agreements with your other top five customers? Has that pricing now set flat, has it been going up or down in the last two, three years?
Hi, Stella. We have regular conversations with customers to try and get future assurance on a variety of levels. We constantly try that. In this instance, what we were able to achieve is an agreement that held prices flat. The prices, we had, you may recall, I think 2017 or 2016, we had a lot of pricing pressure. Since then, prices across our customers have been more or less at a constant level. We always are concerned about price pressure, so we were quite happy to strike this deal, because as volumes go up, people are demanding better prices. In general, prices have been around this level, and this mitigates any future pricing pressure.
Thanks for that. Second question. During the lockdown in China recently, did any of your global scale and technology advantage place you in a better operational form compared to your Chinese competitors? I'm asking this because I want to see if any of your competitive advantage shines through that might benefit you going forward to get a stronger traction coming out of the virus lockdown.
Our sort of global capabilities have put us in a really strong position in China, because we bring those capabilities such as different languages, et cetera, to bear in the Chinese market. We bring our knowhow into the Chinese market. We're not solely reliant on the Chinese market, so we can invest into China, whereas other local companies may be impacted much more by the virus situation. In general, our global capabilities are to our advantage. As I said earlier, domestically, I think we'll be able to weather the impact of the virus better than our local competitors.
Glad to hear that. Nearly lastly, what's the retention rate for Figure Eight? Growth is going up. How is retention, though?
Pretty good. It's no worse than history. We did flag at the half that retention was an issue that we were focusing on. We see now through Q3 and Q4, much stronger engagement with staff. The integration plan is ongoing, and that's combining teams and cultures and all of those other things. We're pretty pleased with how that's going at the moment. We don't see any retention issues other than what the business has experienced through the course of its history.
Great. Last question from me. Can you give us some sense about the R&D CapEx going forward? You previously, before F8, you guided around AUD 12 million, and then that's impacted, or helped by Figure Eight acquisition. Going forward, how should we see this line?
Stella, hi. The best guidance to give you is to use the FY 2019 number as a proxy. FY 2019 was around 2.9%, call it 3%. Around that range is a good guide for 2020.
Great. Thank you. That's all from me.
Thank you.
Thank you. Your next question comes from Johnny Huynh with Evans and Partners. Please go ahead.
Hey, guys. Thank you. I just wanted to ask, it looks like you guys have called out some relevant contract wins or customer wins in the financials. Can you just talk a little bit more about that?
Sorry, Johnny. Your voice is quite low. Could you speak up, sorry?
Hey, sorry. Can you hear me now?
Yes, I can. Thank you.
It looks like you guys have called out some relevant customer wins in the financials. Can you just talk a little bit more about that?
Some relevance wins?
Yeah, Johnny.
Yes, some relevance customer wins.
Yeah, no, it's coming mainly from existing customers and e xisting projects with those existing customers. There has been growth in new customers as well, but obviously the size of the existing business is so large that it dominates.
Okay. That's all from me. Thank you.
Thank you.
Thank you. Your next question comes from Tony Mitchell with Ord Minnett. Please go ahead.
Hi. Good morning. Kevin, can you tell us what was the dollar amount of those timing differences that came through in January?
Well, Tony, maybe probably the best way to answer this is that basically most of the analysts have cash balances of circa AUD 100 million, and we reported AUD 75 million. Essentially, if you think about that as a range, that should help you with that answer.
All right. Thank you for that. Also, just by segment, can you tell us what percentage your largest customers represent of revenue and EBITDA?
We can tell you about revenue. We disclose that. In terms of the top five, the top five make up 88%.
Top five, 88%. Okay.
Yeah.
Thanks very much. Thank you.
Okay.
Thank you. Your next question comes from Chris Savage with Bell Potter. Please go ahead.
Thank you. Good morning, Mark. Good morning, Kevin.
Good morning, Chris.
Just that comment in the pack, Figure Eight path to profitability ahead of plan. Does that just refer to the lower loss Figure Eight generated in 2019? Do you expect to get to positive EBITDA quicker than you anticipated, like when you flagged it at the acquisition last year?
Both.
Last year, you said you expected EBITDA positive contribution in the second half of 2020. Should we now think about it in the first half?
Yes. Having said that, the integration of Figure Eight into Appen will make that harder to see going forward.
Is it one of the reasons why you expect a stronger margin in the second half? It gets to EBITDA breakeven and positive in this half, but then next half, we get a better positive EBITDA contribution.
One off, I think the bigger reason is the investment in sales and marketing in the first half.
Sure. Sure. Just that new U.S. government customer you called out in the pack, what area of the business was that in? Was that Figure Eight?
Yes.
You made that comment at the half that the Figure Eight platform was being implemented behind government firewalls. Was that the preliminary work done and then that converted into a government contract?
When we acquired the business, they had government customers and the platform was implemented on government or behind government firewalls. This win is another installation of the platform behind government firewalls. Once you get behind the firewall, as it were, it makes it easier to win the next deal.
Cool. All right. Thanks very much.
Thank you.
Thanks.
Thank you. Once again, if you wish to ask a question, please press star one on your telephone and wait for your name to be announced. Your next question is a follow-up question from Siraj Ahmed. Please go ahead.
Thanks. Just on the government secure facility market, can you just clarify that? I mean, I thought Figure Eight, as you said, is a platform. Are you investing in the secure facility ahead of potential work that could come in? Just trying to understand that, please.
Yeah. We are investing in a commercial secure facility in the U.S. If you look, I think on page 18, we make mention of a facility in Dallas that we signed a lease for and we'll be staffing up. That gives us the opportunity to work with certain data types that can't be exported out of the U.S., private and personal PII type information. That's on the commercial side. We are looking at a government secure facility in the U.S., but it will be contingent upon growth from that market sector. We're not actually doing anything at the moment, but it's a possibility.
The Washington, D.C. office is more from sales and marketing then?
Yes. Sorry, I misunderstood. Yeah, that's a fairly small office and it's to support the customers. Yeah. Sales and marketing/corporate.
Got it. Secondly, you did mention, I mean, but just the work in hand to the full year revenue conversion. Last year I think it was around 30%. Any reason why it should be different? I understand Figure Eight is a bit more recurring revenue, et cetera. Just trying to understand should that be different this year around or just assume similar conversion?
I think it's safe to assume similar. Also just be aware that the more customers we win, the billing cycle. It's never going to play out exactly the same year-on-year, but it's a safe starting point.
Got it. Just on the speech and image revenue. Kevin, you did say that you expect margins to be flat year-on-year. Looks like speech and image revenue declined in the second half year-on-year. How should we think about that going forward? Have you won more work?
Yeah, no, there's obviously sensitivity to either increasing or reducing revenue, right. We've got a stable base. We're investing in that base for delivery. Obviously, as you know, the life cycle of speech and image projects is a little bit different to relevance, where it's much more susceptible to client investment cycles and product life cycles. Depending on when there's a concentration or lack of concentration of revenue activity in a given half, that could impact the margins. If you obviously blend it over a period of time, it kind of evens out.
Got it. No, that's it from me. Thanks a lot. Thanks, Kevin.
Thank you.
Thank you.
Thank you. Your next question comes from Curtis Larson with Norse Capital. Please go ahead.
Hi, guys. Just a couple of questions on your outlook. You highlighted on your last point on your guidance page that you review capital management and dividend policy. Is there any change in thoughts, given you've bedded down two transactions in the recent past? Have you put that there because you're thinking differently on M&A or?
Hi, Curtis. We've sort of had that statement there the last few releases.
That's because as the business gets more complex and demanding and there's M&A and debt and other things that might use the cash, we're just flagging to investors that we may choose to use that cash in other ways rather than paying it all out or paying out large amounts in dividends. We want to keep a bit of the cash or keep our powder dry to help with the growth of the business.
Okay, thanks. One more question. Is there any numbers from China within your guidance?
There's-
Are you factoring it in, I guess?
It's bundled in. I mean, per the statement on the coronavirus, it's negligible in the scheme of things. It's early days in China for us, but it's included in the guidance.
Okay. Thank you.
Thank you.
Thank you. Your next question is a follow-up question from Quinn Pierson. Please go ahead.
Yeah. Hi, guys. Just quickly one more. On Figure Eight, one of the thesis of that acquisition was the synergies. One synergy component was cost savings by speeding up the annotation by using the Figure Eight platform. Is that part playing out? Are you, I guess, reducing your crowd costs by using some technology in the annotation? Are you able to actually bank those savings? Thanks.
It's early days, Quinn. It is playing out. For example, we've introduced or we're piloting machine learning assisted transcription using speech recognition, and we're seeing efficiency gains, I think in the early 30%, 33% or 35%. Mind you, these are in pilot projects. If we can play that out over time into production work, it'll be meaningful. Overall, we are seeing the benefits, but it's early days.
Helpful. Thank you.
Thank you. There are no further questions at this time. I'll now hand back to Mr. Brayan for closing remarks.
Thank you. Thank you everybody for joining us this morning. We look forward to meeting you or seeing you as we conduct our roadshow. Once again, we appreciate your interest in our company and your ongoing support. Thank you very much. I hope you all have a great day.
Thank you.