Good morning, everybody, and thanks for joining us on this webcast. Just going through our slides for those who are new to Pro Medicus. We are a healthcare IT company specializing in enterprise imaging, mainly for the radiology or diagnostic imaging industry. We have three jurisdictions we work in: headquarters here in Melbourne, Berlin, which is our largest R&D center, and our largest market is in the U.S. Importantly, over half our staff are engineering-based. We have two product sets. The first one is what we call Visage RIS, which was effectively practice management, billing, scheduling, interface to payers that we sell largely here in Australia, and on the right side of the screen, the Visage 7 product, which is a clinical product that radiologists use to call up images, manipulate them, and make a diagnosis, and it's the product that we sell largely in North America.
The Visage RIS, just to start off with that, this was redeveloped, as many of you would know, about four or five years ago. It's a very modern platform designed to scale. We have two key clients here in Australia, the largest and second-largest radiology providers in I-MED and Healius, which was previously known as Primary Health Care. The good news is that the rollout of these are progressing, and we'll talk a bit about that later. It has added over AUD 1 million to our revenue going forward, and once fully deployed, will continue to add up to AUD 4.4 million a year. In terms of the Visage product itself, we believe it is number one in the three key areas. They are speed, functionality, and scalability. It's the combination of those three things that we think make it a leading-edge platform.
The thing that is driving changes in the market, the biggest factor nowadays, is the massive escalation of the size of data sets. They're not going up in a linear fashion. They're going up in an exponential fashion, and some examples here on the slide, things such as high-density CT, breast tomosynthesis, which most probably is the biggest driver. Mammograms used to be about 300-500 megabytes. Breast tomosynthesis, particularly the high-density version or high-definition version, between 6 and 10 gigabytes of study. So what is happening in the market is legacy technology, which takes an image, compresses it, sends it down the network, uncompresses it on a workstation, and uses the power of the workstation to view the images. That technology is starting to crack. The images are just getting too big, and our differentiator is we don't do that.
We don't actually compress, send, and move the image. We use server-side rendering, which we've been doing since Visage was originally developed. We are one of the leaders in this technology, if not the leader, and we actually stream the pixels to the end user, hence the speed and the ability to read any exam, regardless of size, pretty much anywhere, even on limited bandwidth. In terms of the half-year review, there were a few key points. One of our sales, Ohio State University, which we announced in November, was a $9 million five-year deal. It is a state university but has a medical center attached to it. So it's a bit of a hybrid between state-owned and academic.
And then we announced the deal I'll talk a little bit more about later on with Nines, who is an AI startup based out of Palo Alto that's looking to combine AI with teleradiology services. That's a $6 million five-year deal. Also important in the half, we were able to complete phase I of Partners, which occurred at the end of July. We believe in record time, and I'll talk about that. We have released and have clients in what we call Visage in the Cloud, which is our cloud-based offering that uses Google Cloud. And then we also believe that in the half, we've seen the pipeline that was strong before increase even further, particularly after the recent RSNA conference in December 2019. In terms of the financial metrics, you would have seen our results released this morning.
We think all of our key numbers headed in the right direction. Revenue was up, as was profit. And importantly, our margins still stayed above 50%, whereas usually in the first half, they dip a bit because our biggest cost in the RSNA is fully expensed in the half. We're also very pleased that our cash reserves increased significantly to AUD 38.8 million. And hence, the board announced an increased interim dividend of AUD 0.06 per share fully franked. And of course, we remain debt-free, so feel that we have a good claim to state that our balance sheet is strong. In terms of the split, this is a telling slide. It's half to half. The salmon color at the bottom is transaction revenue, and that grew very strongly, so roughly 30% from the previous six months.
We did preempt the market that we believed it would grow because of partners coming online in the July timeframe. Pleasingly, so did the blue sector, which you see, which is service contracts and OEMs, grew through that period as well. So effectively, the important part for us is the blue and the pink areas of the graph, pretty much largely recurring revenue, and so we'll set the base going forward for the second half of the year and, of course, into the future. The operational transaction model, I think most of you are familiar with. It is based on a committed number of minimums, and with the recent contracts we've won, the forward revenue, assuming those contracts that are going to renew within the next five years, renew at the same dollar value, is now increased to about AUD 195 million over the next five years.
It is an annuity-style stream. Clearly, any new contracts that we get going forward will add to this figure and sit on top of those transaction revenues that we already have. As I mentioned before, exam revenue is recurring. It increased 30% half on half. It was due, as we forecast the Partners. We will get some additional revenue coming from Duke, which goes live in another week or two. OSU will come late in the piece and contribute some. Then clearly, going forward into next year, both Duke, OSU, and Partners will contribute a full 12 months' worth. We are also seeing some existing organic growth from clients as they add on new practices, new hospitals, etc. And all of that goes into the mix. Professional services, this was pretty much in line with what we expected.
The new accounting standards came into effect in the last half. And effectively, they're roughly 10% of the value of the contract that we announced, and they're spread over the life of the contract. So we believe we have good claims to state that we have a very scalable offering. There's no hardware involved. So all of the fees we charge or all the revenue comes directly to us. We don't pay out to third parties because we own the software in its entirety. We have a contained cost base, which as a percentage of revenue continued to go down, even though we invested in new staff and will continue to do so. So we have high operating leverage, and margin continues to grow as our footprint increases, particularly in North America.
Of the two sales, the first one was in Ohio State, as mentioned, was a $9 million deal over five years. It's our standard transaction-based model. It is a large regional hospital, and as I said, it is one of the key medical training and teaching facilities in the state of Ohio. That implementation will now occur in the fourth quarter, so towards the end of this financial year. It is largely one significant campus, so the bulk of it will go live in one go, and we believe the main impact of revenue will start early in financial year 2021. Nines is an unusual one because it's a very different sort of client. It's a five-year, $6 million deal. It is transaction-based, but because they are more of a startup, the volumes increase, the minimums increase year on year.
It is a pure software-as-a-service offering based in Google Cloud or Visage in the Cloud. Nines team come from a mix of very highly respected people in Palo Alto that have experience in machine learning AI and some key radiologists in both Mount Sinai, which is a tier-one academic in New York, and Stanford. So their plan is to leverage some AI algorithms that they've produced and provide a more efficient teleradiology service using their radiology manpower. All of it will be based on Visage in the Cloud. So it opens up a whole new segment for us in the market, and they will go live 1st of April in 2020. We still think one of our strengths is not just the technology, but our ability to fast-track the implementations. We have reaffirmed to the market this morning that we are ahead on or ahead of schedule.
We were able to deliver the Partners' project, we believe, in record time for phase I, which many of you will remember were the two key hospital systems of Massachusetts General Hospital and Brigham and Women's Hospital, a large hunk of Boston. We believe that by being able to do it, it delivers huge savings for the clients, and it frees up our staff so that we're able to use the same teams on other implementations going forward. The other big thing that we've noticed is it reduces the barrier to change because implementation issues, time, and cost were a huge barrier for people thinking about change. The fact that we've been able to prove to the market that we can do this very efficiently and quickly has been a big plus for us.
In terms of the proven ROI, we are still known as the most expensive in the market, but we think we justify that by our return on investment, not only from a financial point of view, but also from a clinical point of view. There is a lot of talk around the industry about radiologist burnout, and we think that with our application making it quicker, easier for them, and the fact that they can read from anywhere has really had a significant impact in helping reduce that burnout and make them feel they can get through more in the same amount of time. In terms of our growth strategy, it's multi-pronged, clearly expanding our footprint, and we added two new clients, and as we've mentioned, we'll talk a little bit more about our pipeline continues to grow. We are getting transaction growth from all of our clients.
No one has done less than 100% of what they told us they were going to do in the beginning, and some have grown by more than that, which has been good for us because that organic growth we're able to harness. We have new product offerings, and over the last 24 months, we've introduced the archive. We have introduced a new worklist product, the enterprise imaging product we are starting to put out in various sites, and clearly, AI, which is emerging technology, we believe we're well-suited for when that becomes mainstream, and I'll talk about that a little later. Certainly, expanding to other markets, the main one being Europe, and we're trying to leverage our R&D capability, so we are, particularly with our academic clients, we are looking at partnerships that will help us commercialize some of their product and some joint development.
And the first product that came out of that was the breast density algorithm that we've written with medical input from Yale. Going forward, the pipeline, as I said, we believe it is stronger than it's ever been. RSNA kicked it along even further. Not only were we able to progress those existing in the pipeline, but a number of new opportunities from all areas of the radiological spectrum, private practice, some cloud-based, newer startups following Nines and similar, and also some of the larger institutions that are now telling us they'll be coming to market within the next 12 to 18 months. So all in all, the pipeline has grown not only in volume but also quality. And as I said, RSNA kicked it along in a significant way.
New products, the archive. Just going through it, this is the ability to actually have the software that stores the images. Many of you will remember Mercy became one of our first clients, and it was one of the biggest archives with over 30 million exams in it. They're now not only using it for radiology exams but for other ologies. So they're storing their cardiology and other ologies exams in the archive. So the number of transactions we do are actually even greater than the number of exams they do in radiology. And interestingly, a fair percentage of the opportunities in the pipeline are for both viewer and archive. And clearly, as we offer a fully managed service in the cloud, it will involve our archive rather than third party. So the archive has been instrumental in a number of different areas in helping us in terms of opportunity.
The enterprise imaging, this means different things to different companies. For us, enterprise imaging means one viewer across all diagnostic disciplines plus one viewer across all the clinical review, which means a non-radiologist looking at radiology. Our two key markets that we're looking at are cardiology and ophthalmology. We have some proof of concepts in both working, and clearly, we're looking to extend that. So the bigger our footprint, the more the opportunity to extend that going forward, and then a new thing that we released at RSNA, which is a worklist. It adds to the now the third component of a stack of technology, which is worklist viewer archive. We've been doing worklist here in Australia through our Visage RIS for over 30 years, so we had a lot of domain expertise.
We felt it was important that we were able to offer a single vendor solution for those in the market that required it. Now, not everybody does. Some people already have worklist and archive solutions in place, and we sell the viewer. But there are a number of opportunities that we're coming across that don't have any of these modules and want to buy all from one vendor. So we now offer that as an option. It is also an integral part of Visage in the Cloud, and Nines will be using the Visage Worklist as well as viewer. And we've designed it specifically so that it can interface with modern AI algorithms because a number of algorithms try and prioritize worklist, particularly for emergency room reading. And it was important that this was capable of accepting algorithms, either our own or third party.
That product has now been made officially available in the U.S. And as I said, will also be part of our software-as-a-service Visage in the Cloud strategy, which brings me to my next slide. We do have a partnership with Google Cloud. We are certified for Azure. Importantly, Visage 7 was always designed to be cloud deployable and is highly optimized for cloud, where most of the other systems are not. You get the same ultra-fast performances on-premise. You get the full functionality. The beauty is it's very rapid to deploy because you don't have to purchase and configure hardware. It's pretty much on demand. So it's suitable for all size implementations, small practices all the way through to some of the largest that we would deal with.
And we believe it'll open up market opportunities for us not only in the mid-market, but those institutions who have already committed to put everything in the cloud. Being able to offer this as a solution, we see as a plus. The Visage 7 AI, we did announce the AI accelerator platform just before RSNA. We did showcase it. It is unique, and it's a complete solution that supports both research and production environments, which is particularly important for some of our larger-scale academic clients. It's based on the same market-leading technology. And its whole reason for being is that it can allow our clients to fast-track AI within their organizations, which is something all of them want to do but have struggled with because of all the disparate bits of data and systems that they have. And the accelerator platform allows them to unify it as one.
It has multiple components, but I think clearly the important thing is it works from research to bedside, as we say. And it'll allow third-party algorithms as well as our own and as well as those our clients produce to be run natively on the Visage platform rather than having an additional third-party platform, which radiologists don't want. They want it all integrated as part of their normal desktop. The people that are running it, Malte Westerhoff and Detlef Stalling, are the two co-developers of the Visage platform. Malte's our Global CTO. Detlef is our head of operations and software development. And Ming De Lin, who we hired out of Yale, is a PhD medical scientist. So we have someone on the ground in the U.S. working with clients to work out which opportunities are suitable for us to be involved in.
And that team will continue to expand as more and more opportunities come to light. Then to prove that it all works, we use the AI accelerator ourselves. We developed a breast density algorithm with medical input from the breast imaging team at Yale, which is one of the most highly regarded in the country, rather. The beauty of the product is it works natively on Visage hardware, so you don't have to send images to the cloud and back. It works in virtually real-time. We previewed it at the RSNA as works in progress. Because of the response we've got, it is now in for regulatory approval for FDA 510(k). We will use that as a model for future collaboration with other academic institutions.
Being able to prove that we've done something and it's worked far quicker than most people expected, I think, was a positive. Just going through it in summary, we do have a growing American footprint. We were very pleased that our transaction revenue went up 30% compared to the previous half. It puts us in great shape and gives us a great leg up for year-on-year growth in transaction revenue. We still believe we have market-leading technologies. We didn't see anything at the RSNA that even remotely came close to what we believe we have. Having partners Mayo, Yale, and others under our belt, we think we've proven quite clearly to the market our implementation and support capability. We are also seeing very strong return on investment from all of our clients that tell us about clinical return on investment, financial return on investment.
The pipeline continues to grow. And we think with the Visage Accelerator that we're strategically positioned to leverage AI as it becomes more mainstream in the market. So 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're on a speakerphone, please pick up the handset to ask your question. If you wish to ask a question via the webcast, please enter it on the ask a question box at the bottom right corner of your screen. Your first phone question comes from Chris Cooper of Goldman Sachs. Please go ahead.
Hi. Morning. Thank you for taking my questions. I've got a few. I'll ask them one by one if that's okay. I guess first one for Clayton. Just on the math of the sales growth, I mean, transaction volumes were up 30%. PACS revenue looks to be up 20%. Is this just an issue of, I guess, implementation revenue or service revenue, or is there something else I'm missing there?
Thanks, Chris. It's a mixture of both. So there is some exam revenue growth. We obviously spoke about that when Partners HealthCare came on board. There is some organic growth from some of our existing customers. There is some growth from exam volume from Mercy Archive that was a quarterly amount in the last six-month period. This period, they had six months or six months. And also, there was some, t he support revenue. The German government contract that we licensed that we put on for the German government, we did get additional support for that and additional OEM revenue. So it was in a few different areas. So there was an increase in exam revenue, but there was obviously an increase in the support.
Got it. Okay. Thank you. Second question, just on the contract wins. I mean, I appreciate the last couple of years were particularly strong, particularly fiscal 2019, but you are currently tracking at a slower run rate, I guess, in terms of both number of wins and also contract size. Just focusing on the transaction size part of that equation, I mean, given the success you've had to date in some of the really large institutions, is it reasonable to assume at this stage that more of the incremental growth coming forward is likely to be smaller on average, or is that a trend you expect to recover? I mean, obviously, this is a lumpy business, but clearly, we have very little visibility, and I guess any comments you could offer would be helpful.
Sure. No. The answer is it's not a trend. It's just how it falls. We had our highest state. It's been in the pipeline for over four years. Some of the others in the pipeline are large. There's a mixture. And they just come in. They just come at their own time. So there's quite the opposite. We don't want the market to think that we've done all the big ones and they're only small ones. It's a complete mix in the pipeline with a good smattering of all sizes, including a number of very large ones.
Okay. Thanks. And just last one, pricing. You are pricing your product at a premium to the competition. You've shown a good trajectory in that regard for some time now. We are hearing more and more instances of competitors being more competitive on price. I just wanted to kind of get your thoughts on whether you're seeing that in the market and also what impact that is having on you as a company. Is that changing your conversations at all, either with existing or potential customers in any way?
There's always been someone that wants to price cheaply, so there's nothing new. And for some of our competitors, particularly those with the more aging compression-based technology, it's most probably the only leverage they have. So we're not really seeing any change in the dynamics. It might be different parties doing it, but realistically, no. There's always someone that will be under half or a quarter of what we offer. What we are seeing is some of the ones that have offered it cheaply in the past have not been able to deliver. So that sent a bit of a warning bell to the market. But clearly, look, we've always expected that. We expect it from the majors. We've seen it before. We're seeing it now. We don't think it's really changed.
Thank you. Your next question comes from Garry Sherriff of Royal Bank of Canada. Please go ahead.
Yeah. Hi, Sam and Clayton. Just a few questions in relation to, firstly, contract renewals over the next 12 months. Are there any that you can call out? And if so, have negotiations started on those potential contracts?
We have two. And yes, negotiations have started on both of them. And we expected that. So we'll have two coming up within the next, I would say, six months, roughly. And they've been in contact with us, and we have started those negotiations, as you would imagine we would.
Yep. That makes sense and if for some unusual reason they were to go elsewhere, which I doubt, but let's say they did, what would the transition period be? How long would it take them to actually transition to somebody else?
First, there are a few things. First, they'd have to find someone else. It just depends on who that someone else is. Look, it could take anywhere from a year to two years, depending. Again, it's more who they might transition to. We've thankfully not had that issue before, but it's not an insignificant job because, remember, you have to retrain all your radiologists, and you have to give them a platform that is as good, if not better, than ours. Otherwise, they're not going to be happy. Look, we haven't come across that, thankfully, but it would take, depending on size of organization, clearly, it would take anywhere from 12-24 months or possibly even longer.
I also think that the fact that they're engaging with us is an encouraging sign. It doesn't guarantee it, but at this stage, they're already customers, and they're engaging with us. So we think that we're confident that they'll renew with us.
Yeah. That's clear. And I should have probably clarified. So they haven't actually put an open tender out for other players to bid for the contract that you're aware of?
No.
No? Okay. And in terms of price negotiations, given that you guys historically have been able to push the price lever over time, I imagine that you are expecting or hoping for a material price increase on these legacy contracts if they are renewed?
Yeah. The thing is this. Look, if someone started with us five years ago, clearly, they should have a benefit if they're an early adopter. So we won't expect them to come to current market price. I think that wouldn't be right. But we are hoping that there is a negotiation, and price is part of that.
Thank you. Question on the Partners' Implementation. So you've talked through phase I, which has completed in record time. Could you give me maybe just a rough split on that contract value? So if we look at the contract value for Partners, phase I, roughly what is the percentage of that contract? Because I'm just trying to get a sense for phase II, which will take, I think you said, nine months to implement. I just want to get a rough split, I guess, on the transaction value between one and two, just in percentage terms.
Yeah. So the amount that we announced to Partners was purely phase I. So that amount was just for the Brigham and Women's, well, it's for Brigham and Women's and Massachusetts General. And they're roughly 70% combined of Partners' total volume. Now, the reason phase II will take longer is the hospitals aren't as concentrated, and the numbers are smaller hospitals. And for various logistic reasons, nothing to do with us, they'll be done as sort of much smaller implementation. So it won't be that big bang that we had when we did Partners phase I. There'll be much smaller teams. They'll just be done, I won't say in background, but it's not like the size of team we would use for phase I. So the timing at the moment, based on their timing and their schedule, finishes off in nine months.
So it fits and starts, which i s what we expected. It'll be done sort of almost in background, finished by end of year.
Thank you. Your next question comes from Sarah Mann of Moelis Australia. Please go ahead.
Hey, guys. Just wanted to ask quickly on the U.S. part of the business. So you flagged some of the, you benefited from organic growth from some of your sites, I guess, kind of expanding their operations. Are there any kind of notable M&A among your customers or talk of M&A that we should be aware of that will kind of help your organic growth going forward?
Yeah. Well, the answer is, look, we hear rumors there. The organic growth that we had was more hospitals taking over the radiology of smaller hospitals. Some of our bigger clients, more regional hospitals, and that just led to the mix. We did have one M&A event which was significant but smaller when Mercy took over St. Anthony's in St. Louis that happened last year. And look, there are rumors in markets all the time about M&A between hospital systems. And we live in hope for what we call a free kick where a client takes over someone, 60% or 80% of their size. That hasn't happened yet, but clearly, if it did, would be of significant benefit to us.
Great. And then just in terms of, I guess, your enterprise imaging product and expanding into other departments, you mentioned clearly you've got some proof of concepts that are happening that seem to be working well. What do you think needs to happen from here before they can kind of get converted into contract wins that you can announce?
I think what we need to do is get someone that uses all the departments. As I said, I know a lot of people who talk about Enterprise Imaging. It's just a loose term for us. They use it for the diagnostic component as well as the viewing component. Most people talk about it purely from the viewing component, and that's a far, far different price point. It's much, much, much cheaper. We want to do it for both. I think, look, it is a mindset thing because in some of these hospitals, they have different IT departments for different modalities. Look, we are seeing that change. It's not going to change on a dime or just tomorrow.
But we do see that changing, and we do think we have a compelling offering in as much as it's the same software, same infrastructure, same management, same remote access, things that are significant for a big organization. But clearly, we need to have one group use it across all the modalities, use that as a showcase, and I think that will open it up further. But we are seeing more and more. And as I said, some of it's happening in background that they're archiving into our archive, the other modalities. So we're already getting money from it, but that's not the viewing layer quite yet. But we think that will come.
Excellent. And then just the last one from me. Clearly, you signed the Nines contract, and that's purely cloud-based. And you've, I think, mentioned in the past, you've got other contracts in the tender pipeline that are looking for cloud-based delivery system too. I mean, just interested in how you think that's going to evolve over kind of the next three to five years in terms of, I guess, roughly what kind of % of contracts do you think eventually will kind of move to the cloud in that period?
I think, look, over five years, I think this is just a guess, but 50% of our work would be cloud-based. There are a number of hospitals now. Health is a big market for cloud providers. And a number of big things have happened, particularly with Mayo Clinic will be putting everything into the cloud. So we are seeing a good percentage of the opportunities in the pipeline saying, "If we go, we'll go with cloud." Now, clearly, two things have to happen. They have to go with us, and then they have to push the button on cloud. But that's been their preference upfront. So I think we'll see 50% over five years, maybe even more.
Excellent. Thanks, guys.
Thank you. Your next question comes from Julian Mulcahy of Evans & Partners. Please go ahead.
Hi, Sam and Clayton. Just a bit more color around the revenue sort of surprises sort of first, so it's a little bit stronger. So the new contributions that weren't in the previous half, am I right in saying? So you've got Mercy Archive, Carle, Partners phase I. And then there seems to be quite a reasonable contribution from OEM sales. Is that right?
Yeah. There's a contribution from OEM sales, but also the implementations for Primary Health Care and I-MED. They did kick up a lot in this first half, which we'd expected. So we.
No, I'm just talking about the exam revenue, 30% increase.
Primary Health Care is exam revenue. And OEMs is not part of that exam revenue.
OEMs is in the blue section.
Right. Okay. And so in the second half, we're going to see Duke, Ohio, Nines, and.
Yeah. The end of the six-month period. So it'll be a small flush of that revenue.
Yeah. But Duke is going live. Yeah. So Duke is going live next week. And it's a bit like Yale. It's largely focused around the main teaching hospital campus. So 80% or 90% of Duke will go live in one go. So come end of February, we'll have March, April, May, June, fair percentage of revenue. So it'll be four months of nearly about 80+% revenue from Duke. But the rest will be completed fairly quickly. So when we go into the first half of 2021, we should have pretty much all of Duke and most of the other done, and they will then contribute a full six months. And this coming half, we'll also get a full six months of Partners because we completed it towards the end of July.
So it was about five months' worth. And then anything else that we implement, there are some other things where people have acquired or growing or starting new hospitals. They'll come into this half, depending on where they go live, and then contribute a full six months next half.
Right, and would you expect many OEM sales in the second half?
We always have OEM sales, but we've had that since acquisition. They kicked up a little bit in this half, but not materially different from other periods. So whilst that was in the blue section, so not in the exam revenue, the majority of that was through increased support contracts for I-MED and the German government sales.
Right. So the first half numbers for Primary and I-MED, are they now the full run rate, or is there still a bit more in the second half?
No, they're not full run rate. We're still implementing. So the five-year period starts at the end of those rollouts or implementation periods. So we still have some implementations in the second half.
Okay. Thanks, guys.
Thank you. Your next question comes from Peter Meichelboeck of Select Equities. Please go ahead.
Hi, guys. Just in relation to the, there's obviously a series of growth opportunities that you're continually talking about in terms of the pipeline and etc., etc. And I'm just thinking in terms of the U.S., what sort of resourcing do you have there at the moment to actually, I suppose, take full advantage of that opportunity? Just trying to get a sense of your staffing and resources for that. And is there an easy ability to expand that at a reasonable cost if you have to do that?
Yeah. Well, good news is we increased our sales team by 50%. So we went from two to three. New guys just started on Monday. Look, it's hard to find really good people. It's a different mindset to those that have worked in big companies. Clearly, if we felt we were missing opportunities, we would put on more. We have put on some more technical people, some on the East Coast, because it means that we have people that sort of 8:00 A.M. in the morning their time would be 5:00 A.M. in the morning West Coast time. So we've had a new support person join us. We've had some new people in Berlin. So we have been actively recruiting throughout the half. And you'll see that reflected in our cost base. But as we've told the market, we're doing it in ones and twos rather than tens and twenties.
Look, at this point, clearly, we don't feel that we need to double our staff or increase it by 30%. Otherwise, we would. We're comfortable that we're well positioned. Having put our implementations, getting them done is really, really important because, as I said this morning, we've got Duke and OSU ahead of us. Then we still have plenty of runway if we get new clients to be able to put them in with the teams that we've got. Yeah, look, we've been selectively hiring and ferreting out the good people, doing a lot more on the research side. That may take some more people, but more in ones and twos incremental than a whole 20%-30% increase in staff at this point.
Right. Okay. And just on the, I think, in the open briefing, you talked about how you view the cash as sort of the three buckets, the investment in the business first, and then the dividends, and then the M&A. I suppose I had two questions on that. I mean, I don't think too many people would disagree with that approach. Just on the M&A, has the environment at all changed in terms of opportunities for you guys at all? Do you see any change in that? And the second part of that question is also, I mean, should we be thinking what should we be thinking in terms of the buyback? Obviously, only a very, very small amount has been utilized. Where does that sort of fit into that picture, I guess?
Well, look, we're always looking for opportunities. And as you can imagine, there are a number that come around. One of our competitors, which was owned by private equity, recently got sold to another private equity. Clearly, that was not our main focus. We're looking at something where we can have adjacent technology rather than taking our market share. Look, they're hard to come by. In the AI space, there are a lot of companies, but will they ever see the light of day? Will they ever make revenue? These things are hard to judge. But look, we're always looking. I will say it has changed a bit maybe in the last eight to 12 months. Some of those companies that thought they'd get more traction early on seemingly have not been able to. So maybe they'll find it a little tougher in terms of funding that may suit us.
But look, we're looking, but there's nothing just on the radar at the moment. But that doesn't mean there won't be tomorrow or the week after.
In terms of the buyback, that just sits there, or?
Yeah, I think so. I think at the time we felt that the shares were undervalued. But look, we make our own decision. People make their own. And so at the moment, yes, I think it's there, but.
Look, and just the last one from me. Sam, I think it was well, there's two parts to this as well, just in terms of the share register. I just wanted to get a sense of how does the share register compare now to, say, where it was sort of 12 or 18 months ago, particularly around sort of international holding. And just on the share register, the second part I wanted to ask was, I think it was two years ago that you guys were encouraged, yourself and Anthony were encouraged, I think, to sell 3 million shares each over 12 months. And I think from memory, you've done 2 million each. I just wanted to also find out if the board was still providing that encouragement for another million shares each or something to help with the liquidity.
Look, on the register, we believe we've got more overseas institutions only because they tell us. We don't know that empirically, but we're pretty sure there's been a greater uptake than maybe a year or two years ago. Look, the board talked to us about two years ago and a month ago, and we said up to three million shares. You're right. We've done two million each, which is just a very small percentage of our shareholding. But that doesn't necessarily mean it will be three million. It just means up to. So I wouldn't read too much into it.
Okay. Thank you.
Thank you. Your next question comes from Josh Kannourakis of UBS. Please go ahead.
Hi, Sam and Clayton. Can you hear me okay?
Yep.
Yep.
Great. Firstly, I know you just mentioned before on some of the incremental growth coming through in the second half. Can we maybe just reference the second half in terms of cost and profitability, maybe talk through your cost and spend of SG&A in the first half and any other costs that we should be aware of in the second half?
Look, nothing major. The one thing is we will be having those staff that came on partway through the first half. Clearly, if they're only there for three months, three months of cost, if they're there through the second month or the second half at six months. So there's a little bit of that. Again, we plan that we will have some additional staff sometime in the second half, depending on who we find and when. But we don't think those costs will be material compared to the growth in revenue. We have conferences that we go to. In fact, we're off to one in Orlando, HIMSS, the big health informatics one. But the cost is just a fraction of RSNA. So effectively, we think the cost base will most probably go down because there's no RSNA.
There's no major cost in the second half that wasn't there in the first. And as I said, maybe the cost base in terms of staff a little bit, but that'll be more than offset by not having RSNA.
Yeah. Got it. And then, so just on the RSNA spend, so it's sort of similar to historical years that you've called out in terms of sort of the scope of.
Oh, the bigger thing. The bigger last year, rather. Simply, we had a bigger stand. We had more display areas. We had more staff. That was intentional, and I think it was money well spent.
Yep. Got it. Cool. Just on the enterprise customer side, so you did mention that you have a few customers that you are rolling that out with, even though it is in the early stages. Are you able to give any context of, if not who the customers are, just what type of customers and whether they've got any other what their sort of rollout strategy is for enterprise across their businesses?
Yeah. Well, it's different for each one. But the two areas currently are cardiology and ophthalmology, and each hospital is different. Ophthalmology is an interesting one because there's new forms of imaging, like ocular CT, that's come in, and they can create some pretty big files, Retinal FlowMaps, all this sort of stuff. And so we're well suited to that. So as those ologies create bigger data sets, it's better for us. So yeah, we've got some in ophthalmology, some in cardiology, but not in the same institution at this point. But clearly, we're working towards it.
Got it. And just in terms of where you're at in terms of the development process for the clinical components of those, are they like if we sort of looked at them, do you feel they're 80% of the way there from what you feel like you need to be to be in a similar position to you are in radiology, or where do you sort of think you are on that development curve?
Yeah. North of 80%, absolutely. It really depends. Some people require more sophisticated functionality than others, even in the same discipline, so cardiology in one place is not the same as cardiology in another, but look, well north of 80%.
Yeah. Got it. And just with the U.S. election coming up, Sam, is there anything in terms of any market chatter around changes to reimbursement that would result in any sort of slowdown on deals happening, or are you pretty confident that the outlook in that regard is not going to have an impact?
Look, no one knows what happens in U.S. politics. None of the politicians themselves. But look, at this point, we haven't had any indication that there would be a change that something would slow down. So nothing has happened globally in terms of impact of what's happening in China and other bits. There doesn't seem to be any impact on us.
Yeah. Okay. Cool. And then just margins moving forward. I'm not sure I came onto the call a little bit late. So I'm not sure if you mentioned it, but just how we should look at that margin. I know you've obviously said this sort of holding the 50% EBIT margin, but whether or not we should still expect some proportional operating leverage to come through on what your views are on investment over the next few years.
Look, margin is a good metric, and we're pleased it's about 50%, but we think spend on the business first. We're lucky, and we've been working that we can do both, and the margins have increased. Usually, first-half margins are lower because of RSNA, so we think we're in good shape for the full year in terms of margins, but clearly, we'll have to wait till that all plays out.
And just on that, Josh, this time last year, margins were around 48% for the half. This year, about 50%. Over the full year, we got them to 52.5%. So we do expect second half to drag it up. But again, we'll see how we go with investment. But first half has generally been lower.
Thank you. Your next question comes from John Hester from Bell Potter. Please go ahead.
Good morning, Sam and Clayton. Just on slide nine, sorry, slide 18, which is referring to the Nines contract. Sam, can you just tell us how are these guys using artificial intelligence now, and in what capacity do they use it, and how is that going to really influence your relationship with them over the next or over the term of this contract? What is it going to mean for you?
Look, first of all, they're a little different because they don't have an existing practice. So their numbers are going to build from zero up. The other interesting thing about them is if you look up the two founders, one is a really well-credentialed sort of AI entrepreneur out of Silicon Valley. I think he was one of the head people at Stanford behind the first driverless car that became Waymo. So he's been around a while, David Stavens. And then if you look at the medical input, he's vice chair of radiology at Mount Sinai. And then if you look at all their advisory boards for who's who of radiologists, so I think they're saying they've got their own engineers who have written algorithms. They're in FDA approval at the moment.
They believe between their algorithms and our platform, they'll be able to offer a far more tailored and better service in terms of speed and quality of reads of diagnostics. So clearly, we'll see more of it when it goes live in the first of April this year. But there's a lot of buzz around them in the industry, and I'll be interested to see how it all goes. But for us, this could grow significantly if they get a foothold in the remote reading market. It's currently done more by corporates on a cost basis. So if they can improve the quality and get market share, it can only be good for us.
So these guys are like a hired gun. They're going to go and target hospitals who are underserviced in remote areas.
I'm not sure exactly their business plan, but yes, I don't think it'll just be remote areas. It could be T1 academics that they do their after-hours, all sorts of things. Bear in mind they have a lot of very high-end subspecialist radiologists. So it might be that they're specialized reads. There are all sorts of things that they're targeting. But as I said, we'll find out more come first April.
Okay. All right. Thank you. That's all from me.
Thank you. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. And if you wish to ask a question via the webcast, please enter it in the ask a question box at the bottom right corner of your screen. Your first webcast question comes from Stella Wang, a private investor, with thanks, Dr. Hupert, regarding the teleradiology space, which may be a larger market for the company than many people realize. Could you please give examples to help us understand that potential?
Yeah. Look, teleradiology at the moment is largely done, as I just mentioned before, by corporates. There are companies like vRad, Envision that do a few million exams where it's purely teleradiology they read for others. And I think that there's some new entrants that are saying maybe we need a higher quality read, a higher quality diagnostic process, and we could do it more clinically accurate and better if we use AI. And that's the theory behind Nines. So clearly, they would be targeting the, at least in the beginning, the teleradiology work of these large corporates now. How successful they'll be, one will have to wait and see. But they're approaching it from a slightly different angle than the radiologists that they're using. A very, very highly qualified one who has recently come on board from Nines was head of emergency radiology, emergency room radiology at Yale.
So very well-credentialed, has left Yale and then become part of Nines now. What that means longer term, as I said, we'll start to see from the first of April onwards. But look, it's a great possibility for us. And if they grow, we grow. If others enter the market, see that we've been the platform for someone successful, at least opens a door for us. So we think it can only be positive.
Thank you. Your next question comes from Claude Walker from Ethical Equities with question, With Nines, based on the expected ramp-up, when would you expect to have your first $1 million a year from it? Furthermore, is there a secondary value to PME with their AI algorithms? Will they be making their algorithms available to other customers?
Look, they're all questions that's a bit hard to answer. We did say that they would be starting with step minimum. So the minimum second year is bigger than the first, third year is bigger than the second. But really depends on how quickly they're adopted by the market or how quickly they penetrate the market as to when we get our first 1 million. Look, it could happen at any time. Just really, it depends. So if they come out of the gate very strongly, get a lot of clients, then it could occur sooner rather than later. But purely up to them. In terms of the algorithms that they're theirs, they're their own that the FDA are approving. We haven't even broached the fact or the option of providing it to others.
We will be seeing them when we're in the U.S. in March, and then they'll let us know what their plans are, but we don't know the answer to that question.
Thank you. Your next question comes from Lewis Abboud Hogben from LHC Capital with, "Hi Sam and Clayton. Great to be chatting again. Congrats on all the progress over the last two years. When you say transaction revenue is up 30%, how much of that growth is the impact of volume growth above minimums versus the impact of more sites coming online, partners with their minimums being recognized for the first time? And how many dollars are you investing into AI algorithms developed in-house?
In answer to the first question, Partners certainly was a big chunk of it because they're a big organization, and we got them up on board fairly quickly. From August onwards, they would be a reasonable percentage of that increase. I think we told the market we were expecting that. Look, there were some other good contributors. As I said, no one was under their volumes. Some were materially over because they'd taken over hospitals or they'd built new sites. It was a whole mix, but certainly Partners was a big part of it, no question. In terms of AI, we have multiple projects going. Some of them are developing our own algorithms. Some are joint research with clients. It's clearly a growing field, and it depends client to client, project to project.
But currently, we're funding that out of all our current cash flow. So there's no need to set aside significant additional funding. It's all being done through our normal expenses.
Thank you. Your next question comes from Ian Lee of Allianz Global Investors with question, "Nature journal reported that AI has more accurately diagnosed breast cancer from mammograms. Can you please tell us how far along you are with the relationship with Yale? Whose AI system was used in the Nature journal?
I'm not sure about whose it was in the Nature journal, but there are a lot of people writing algorithms to detect breast cancer anywhere from Holland to Korea to our actual clients themselves. So our view is we need to have an open platform. We're agnostic. They don't have to just use our algorithms. We did the breast density one because the opportunity arose, and it's a proof of concept for the accelerator platform. We're able to take concept to regulatory FDA in pretty much record time using the platform, which proves that all the bits work. So look, there will be a large number of groups, and there are an increasing number getting FDA approval that diagnose breast cancer. But it's not one thing. There's breast density. There's breast cysts. There's breast calcification. So breast cancer is not just one diagnosis. It's a whole spectrum of algorithms.
Look, they're only just starting to get FDA approval and being used. As I said, we're looking to tap into that, whether it's ours or someone else's. We're agnostic.
Thank you. Your next question comes from Peter Richardson of TI with question, "How much of the increased revenue is attributable to exchange rate changes?
Yeah. I think the increase in the revenue was fairly minimal from exchange rates. The period ending 30th of June and the period ending 31st December, the exchange rate actually came back to be around the same amount. So while it would have had a small incremental increase to the revenues, it also would have affected expenses in terms of both our U.S. and in terms of the euro exchange rates for our development staff. So yes, it did have an impact, but it was small. It would have been less than AUD 100,000.
Thank you. Your next question comes from Anthony and, sorry, from Rania with question, "Are you still maintaining your competitive distance from others in the marketplace with an imaging solution? Do you have to change the software much to create an ophthalmology or cardiology offer and relative to radiology? How big is the opportunity?
In answer to the first one, yes, we still think we're adding to 24 months ahead. We know some have tried to put out a server-side rendering streaming platform and have had issues. As I said, we're not standing still either. Quite the opposite. We've released more new product and more enhancements to existing products at the RSNA than any other year in the past. So we still think we've got a strong lead, not just in the technology, but also in the ability to implement it, which is equally as important. With ophthalmology and cardiology, we already did a large number of cardiology-type tests, cardiac CT, cardiac echo, or ultrasound. So the changes, as I said, would be less than 15% additional to what we do. Ophthalmology, pretty much not much, if anything.
The way we handle reflective light with photos, videos, and the specialized OCT was all native to the product. So it's more a mindset of people getting used to one product for everything rather than the technology itself, at least with our product.
Thank you. Your next question comes from Enterprise Super with question, "Sam, what is the major source of new sales, word of mouth, presentation at trade conferences, or sales staff on the ground contacting hospitals?
Really, the first two because we tend not to contact hospitals. We find that ineffective. It's word of mouth. The radiologists speak to each other all the time. They're on panels together, etc., and we call that the network effect. Once we got Mayo, that was one part, then we got Yale that added to it, then we got Partners that added to it again, so each group that we get and successfully implement helps that word of mouth, and conference is the big one, RSNA, but we will attend a few other much smaller but more targeted conferences like the breast imaging one, etc., and they've usually heard of us and then come up and see us, so word of mouth and conference, but certainly not cold calling.
Second question from Enterprise Super with: "You mentioned the capacity to build AI algorithms within the product. Can you expand on the benefits of this compared with other platforms?
Yes. Most platforms are third-party. They're an additional platform that you have to buy support, and it runs independent of the radiologist's desktop. A window pops up or something showing, something pops up, and radiologists don't want that because they need it integrated into their desktop so that if the AI algorithm points out an area of interest or something where they think there's a nodule or a tumor, then clearly they don't want to see that in a separate window, so there's a huge advantage of being able to show the output of that natively within the product, and we've written this API that does that, so that's one thing.
The second thing is because of our association with some of our large academic clients who all have data sciences labs, we may jointly develop algorithms with them, or we may have, in particular, agreements, data use agreements where we get certain anonymized data sets that we can train the machine on. It's a spectrum of anywhere, but just knowing that it exists and putting it on the platform to being involved in the development from the beginning or partway through. It really depends project to project. It gives us a pretty wide source of available possibilities for algorithms going forward.
Thank you. Your next question comes from Nigel Beale of Beale Superannuation Fund with question, "Is there any prospect of expanding beyond the current markets into new geographical areas, particularly Asia?
The answer is yes. We don't have any salespeople there at this point. Certainly, if an opportunity came up, Singapore or Malaysia or Thailand or wherever there's a large tier-one hospital, there'd be absolutely no reason we couldn't service that. It would be very straightforward, then there's the whole black box of China, and at the moment, we know little about that. Clearly, it may be of interest in the future, but we don't have people there at this point, so really, we're focusing on what is the largest market, which is the U.S., and since we're sort of making a significant imprint there, we think that's where we should focus most of our resources, but then we're also looking at parts of Europe, and as I said, if something more opportunistic came out of the Middle East or Asia, there's no reason we couldn't do it.
Thank you. Your next question comes from Paul Markham, shareholder, with question, "Have you outsourced annotation of your image data sets specifically to companies such as Appen?
No, and we couldn't. The only annotation that's done has to be done by radiological experts. They have to codify, tell you it's a grade 2B tumor, etc., so you couldn't use a third party, so really, what we call curation or what you call annotation is all done by trained radiologists, and the advantage that some of our large academic clients have is they have some of the best radiologists in the world, so therefore some of the best curators, and they also have the data sets. The thing that they were missing was the platform, and that's where our research platform comes in, but in answer to your question, no, that's not outsourced to third-party companies.
Thank you. And your final question is a follow-up question from Garry Sherriff of Royal Bank of Canada. Please go ahead.
Yeah. Two more questions, guys. Just wanted to clarify. In terms of your existing customer base with Visage, my understanding is that everybody is still on-prem. There is no deployment into the cloud. I just wanted to A, clarify that. And secondly, could there or are you considering pricing differently for cloud deployment just because the customer wouldn't have to have new or additional hardware configurations?
The answer is we do have some clients in the cloud. They're small-ish, so we haven't made a material sort of announcement about it, but we do have some, and we think there will be some bigger ones going forward. Even some of our existing clients are looking at, do they migrate part or all of their implementation to the cloud? The difference would be two things. First, implementation costs because it's much quicker for us. So for them, in terms of professional services, cloud deployment is much cheaper because we don't have to spin up all this hardware and then go through their network and deal with all their security people and everything else, so for them, implementation costs would be less. For us, it'd be quicker and easier to manage, so there would be some savings there.
Look, their cloud costs, there could be some pass-through, but that we would deem to be small revenue. It's easier to spin up. It's easier to maintain. And the other thing is for someone like Nines, for instance, as they grow, you can just dial up more resources in the cloud pretty much instantly. You don't have to wait six months for purchasing decision of hardware and other things. So it's ideal for people like them.
So just to clarify, so you wouldn't have differential pricing on a cloud-based product relative to on-prem?
It's pretty much the same. The only thing that would also be, as I said, professional services. And the only other thing is in cloud, they would use more modules because that would be our archive and usually our worklist. So the bundle of applications would be more of our applications than in some on-prem. So it just depends. If they've got nothing and they want worklist, archive, and viewer on-prem, then other than the professional services, it would be relatively similar.
But like for like, for the viewer rate would be the same.
Same. Yeah.
Yeah. Okay. That's clear. And just final question. Again, I know people have talked about staff numbers given it is your largest OpEx item. So it did grow about AUD 1.5 million in first half 2020 on first half 2019. And I know we've talked about RSNA not going to be in the second half, but there will be some incremental flow-on effects from staff that you have hired cycling through. So should we be thinking I guess what I'm trying to get to, second half staff costs, do you foresee them being slightly higher or flat-ish?
They could be flat-ish, but there's also some variable compensation in there. So it just depends whether that's repeated. So some of it is the staffing costs just purely the salaries. Others are variable amount.
Yeah. Some of the bonuses, STI, all that sort of stuff. So it varies depending on what targets we hit. And clearly, they're paid out more in the previous period than the next. So there's a little bit of seasonality in there.
Thank you. There are no further questions at this time. I'll now hand back to Dr. Hupert for closing remarks.
Thank you. I just wanted to say thanks, everybody, for joining us and appreciate all the questions, and we'll look forward to doing this again at the end of the full year results, so thank you.
Thanks.