Appen Limited (APX) Earnings Call Transcript

Good morning, ladies and gentlemen. My name is Richard Freudenstein, and it's my pleasure, as Chair of Appen Limited, to welcome everyone joining us today for our Annual General Meeting. I'd like to begin today by acknowledging the traditional custodians of the land on which we gather, the Gadigal people of the Eora Nation. I also acknowledge the traditional custodians of the various lands which you are joining us from and the First Nations people participating in our meeting. I pay my respects to elders past and present. I wish to advise that today's AGM is being recorded and will be made available on our website after the meeting. It is now just past 10 a.m., the nominated time for the meeting, and I've been informed that a quorum is present. I note the meeting has been validly constituted, and I'm pleased to declare the meeting open. I would like to begin by introducing my fellow directors that are present with us today: Robin Low, Nonexecutive Director and Chair of the Audit and Risk Management Committee; Stuart Davis, Nonexecutive Director; and our Chief Executive Officer and Managing Director, Ryan Kolln. Our 4 other Nonexecutive Directors are attending by phone as they are based in the U.S.: Steve Hasker, who is also Chair of our People and Culture Committee; Vanessa Liu; Lynn Mickleburgh; and Mini Peiris. We also have various Appen executives present, representative of the company's auditors, KPMG, and representatives from the company's share register, Link Market Services. There are 4 components to today's meeting. First, I will provide you with an update on Appen's business. This will be followed by a detailed overview from our CEO and Managing Director, Ryan. We will open the meeting for questions on general business after my address and Ryan's presentation. Following the general business questions, we will then move to the formal business of the meeting where the items set out in the Notice of Meeting will be put to shareholders. We will also allocate time for questions on each of the items of business when they are considered. Thank you to those shareholders who have submitted questions prior to the meeting. We have endeavored to answer those questions during the upcoming presentations. I want to acknowledge that 2023 was a disappointing year for our shareholders. While the rise of generative AI has created growth expectations, difficult macro conditions and a slowdown of tech spending has significantly impacted Appen's performance. I will start with a discussion of the key areas the Board has focused on, including improving Appen's financial performance and resetting the company. I will then provide some commentary about the various items of business. I will then ask Ryan, as the CEO, to focus on the performance of the business, strategy and trading update in detail. Note that values referred to are U.S. dollars unless otherwise stated. Turning to financials. The company recorded a statutory loss of $118.1 million, which included a noncash impairment of $69.2 million related to its global services business. It's important to note that the impairment is a nonoperating and noncash item and in no way impacted Appen's liquidity. Total operating revenue declined 29.7% to $273 million, primarily due to the slowdown in spending by our largest customer and general slowdown in tech spending. In turn, this led to a significant reduction in the company's profitability as Appen recorded an underlying EBITDA loss before foreign exchange of $20.4 million compared to $13.6 million in the prior year. The Board made a decision not to declare an interim or final dividend in 2023. This decision was considered appropriate due to Appen's financial performance and to ensure an appropriate allocation of capital. In response to the challenging external conditions, we committed to decisive action of resetting the business. The immediate focus was to refresh leadership, remove costs and position Appen to participate in the generative AI services market. In support of our strategy reset, the company raised AUD 60 million of equity in June 2023 and a further AUD 30 million of equity in December 2023. We also implemented a $60 million cost reduction program, made operational improvements to the business, for example, in project delivery and crowd management. Generative AI was an important focus of the industry in 2023. While there were some early benefits from our generative AI initiatives, these were insufficient to offset the revenue and earnings decline in our core business. Earlier this year, Google advised us of its decision to end its global service contract, with all work ceasing on 19 March. In 2023, Appen's revenue from Google was approximately $83 or 30% of total revenue. In line with Google's decision, we have taken action to reduce our cost base by a further $13.5 million. To date, approximately 80% has been implemented with the remaining cost reductions expected to be achieved by 30 June 2024. We are very focused on profitability. As such, we are committed to managing our costs in line with revenue and will implement further cost reduction initiatives, if required. As part of our reset, we refreshed our leadership and management team. On 5 February 2024, Ryan Kolln was appointed as our new CEO and Managing Director. Ryan joined Appen in 2018 and has made an extraordinary contribution during his time with the company. He has a deep understanding of Appen, coupled with strong technical understanding of the AI market and a proven strategy background. Ryan is the ideal leader to guide Appen through its next phase and navigate the evolving generative AI market. He is based in North America. We've also assembled a refreshed management team with deep operational experience and a clear structure to maximize opportunities going forward. In addition to our financial metrics, we are also focused on good, social and governance outcomes. Despite the challenges faced by our business, our commitment to the crowd and customers remains as strong as ever. In 2023, we invested in systems and processes to improve the crowd and customer experience. While we experienced a significant uplift in customer NPS, crowd NPS declined due to a variety of factors, including lower project availability. An integral part of our strategy is our commitment to responsible AI and to ensure that AI performs correctly. Fostering diversity among the crowd and continued high ethical treatment of the crowd is key to building responsible AI that reflects the real world. In support of our commitment, our team conducted further research programs to understand representation across the crowd and address any gaps. Diversity across our organization also remains a priority, with female representation amongst our employees at 55% at the end of the year. An unintended consequence of our organizational change has seen a reduction in female representation of senior leadership team from 30% to 22%. Despite this, our target of 30% female representation in senior management positions remains. Our social impact work is an important priority. Last year, we established 5 partnerships with global nonprofits to foster diversity and offer work opportunities within our crowd to underrepresent individuals, including refugees. Maintaining a strong focus on governance is a key priority of the Board. Our policies and practices remain consistent with the latest ASX corporate governance principles and recommendations. Last year, there were no changes to the Board structure. The Board comprises 8 directors, including the Managing Director. Appen has 3 nonexecutive directors based in Australia: Robin Low, Stuart Davis and myself. The 4 remaining nonexecutive directors, Vanessa Liu, Steve Hasker, Mini Peiris and Lynn Mickleburgh, are based in the U.S. The composition of Appen's Board maintains the right mix of skills and includes relevant diverse backgrounds. Board gender diversity also continues to meet our target, with female representation of nonexecutive directors at 50%. Today, Robin and Vanessa are standing for reelection and will address shareholders on their candidacies later in the meeting. I will now make some comments on remuneration, and in particular, item 5 on the agenda, the grant of long-term incentive performance rights to the CEO and Managing Director; and item 8 on the agenda, the approval of Appen's long-term incentive plan, terms and conditions, and the subsequent issue of securities under that plan. Steve Hasker, Chair of the Board's People and Culture Committee, will speak to the remuneration report in more detail. We devote significant time to strike a balance between setting remuneration for executives at levels that reflect company performance and shareholder expectations while also being sufficient to attract and retain talent in the highly competitive technology markets in North America and Australia. In keeping with this approach, we made significant changes to our remuneration framework in FY '22 and further updates in FY '23. Steve will address these. I will keep my comments brief with respect to 2023 remuneration outcomes. As revenue and earnings performance targets were not met, there was no STI payable for the company's financial performance. In terms of our nonfinancial performance, customer NPS was 140% of target, and the employee engagement was 91.9% of target, which was above the 90% payout threshold. In recognition of these results, a partial STI, which equates to 10.6% of the maximum amount payable, was paid. There was no STI paid to the former CEO. Turning to the CEO and Managing Director's remuneration arrangements and item 5 on the agenda. Ryan's package includes a base salary of $600,000. His STI is equivalent of $600,000, per annum, with a maximum opportunity of 150% of fixed remuneration. The STI will be delivered 75% in cash and 25% in deferred equity, which vests 12 months -- after 12 months subject to continued service. The LTI component of Ryan's remuneration is valued at $1.5 million. If at the end of a 3-year period, Appen's share price reaches $1.60, 50% of the LTI will vest. For 100% to vest, the share price needs to be over $2.30 to trigger full vesting. These targets are aligned with the annual STI plan for other executives that participate in the plan. As you can see, for any vesting to occur, the share price must be significantly higher than it is today. Therefore, Ryan will only be rewarded for delivering significant shareholder value. The Board considers his package to be fully aligned with shareholders' interest. Turning to item 8 on the agenda, which seeks approval for Appen's long-term incentive plan and subsequent issue of securities under that plan. Appen wishes to exclude issues of securities under the long-term incentive plan from the 15% limit in ASX Listing Rules 7.1. ASX Listing Rules 7.2 provides -- exception 13B provides that ASX Listing Rule 7.1 does not apply in respect of the issue of securities by the company under an employee incentive plan, if within 3 years before the issue date of the relevant securities, the shareholders have approved the issue of securities under that employee incentive plan as an exception to ASX Listing Rule 7.1. Shareholder approval was last obtained on 18 May 2018 and was valid until 2021. Since 2021, the issue of securities under Appen's long-term incentive plan have been under ASX Listing Rule 7.1 and not as an exception under ASX Listing Rule 7.2. The directors consider the securities issued under the incentive plan as a cost-effective, well-aligned and efficient incentive to retain key employees when compared with other forms of incentives such as cash bonuses or increased remuneration. The maximum number of securities proposed to be issued under item 8, subject to shareholder approval today, is 33 million securities. Outside of remuneration related to these items, we're also seeking to ratify the prior issue of shares under the equity raise completed in December '23 as per item 6 and the prior issue of shares and warrants to vendors of Quadrant under item 7. The issue of these shares effectively uses up part of the 15% limit under ASX Listing Rule 7.1. If approved, the issue of these shares related to item 6 and 7 will be excluded from calculating the company's 15% limit in ASX Listing Rule 7.1. In closing, I would like to reiterate our commitment to improving Appen's performance and delivering better return for shareholders. While good progress has been made to reset the business, there remains much to be done. On behalf of the Board, I would like to thank Appen's shareholders for your ongoing support and patience. I also thank all Appen employees for their contribution and commitment. I now welcome Ryan to give his first AGM address as CEO and Managing Director.

Thank you, Richard. Good morning, everyone. It's a pleasure to meet with you this morning and provide an update on the progress we've made to reset Appen's business. I'll start by providing an overview of Appen's FY '23 performance and an update on Appen's strategy. I'll also share an update on our year-to-date trading performance and provide an updated outlook statement. Turning to our FY '23 revenue performance. Revenue declined approximately 30% to $273 million in FY '23. As Richard mentioned, this is a disappointing result for the business and for our shareholders. The reduction in revenue was driven by reduced spending from a large customer, along with the general slowdown in tech spending and the uncertainty related to generative AI investments. In response to the revenue slowdown, we reduced our costs by $60 million throughout FY '23. These reductions focused on removing some of the direct costs related to revenue, streamlining existing operations, minimizing organizational layers and optimizing our growth investments. Because of the cost-out timing, the first full year benefit was only realized in FY '24 and therefore, negatively impacted FY '23 profitability. Excluding the impact of foreign exchange, we recorded an underlying EBITDA loss of negative $20.4 million compared to $13.6 million in FY '22. In December, we realized some of the benefit of our cost-out reduction measures and exited the year cash EBITDA-positive. While Appen's full year revenue declined, we did see some positive trends in Q4 FY '23. China produced a record Q4 revenue as we gained significant traction across multiple generative AI projects. Outside of China, the new markets business also experienced growth in Q4 when compared to Q3. Our business typically experiences Q4 seasonality, especially in the Global Services business. While Q4 performance for Global Services was significantly down on the prior year, we did see improvement compared to Q3. In 2023, many customers were testing different approaches to building large language models, which led to a lot of pilot projects and some customers scaling their data needs. In turn, we saw a 410% half-on-half revenue increase from large language model-related work. These are very pleasing trends and have carried through to our year-to-date performance in FY '24, which I will share with you in a moment. Appen has been supporting the development of AI since the company was founded in 1996. The industry has evolved significantly since then. And now with the introduction of generative AI, the potential for AI is at an all-time high. As a proof point, Bloomberg and IDC forecast the generative AI market to reach $1.3 trillion by 2032, growing at a 42% compounded annual growth rate. We are very bullish on the impact of generative AI, and our strategy is strongly focused on capturing value from the market. The impact of generative AI has a significant impact on Appen's total addressable market, or TAM. Since inception, we have been providing data services to technology companies that build AI deep learning models. Analysts estimate this market to be between $12 billion and $17 billion by 2030. The introduction of generative AI expands our TAM in 2 areas. The first relates to generative AI model builders who are creating the foundation models. This group includes both very large technology companies and many well-funded start-ups. We estimate that Appen is already working with approximately 80% of these companies, mostly providing data that is used to improve their models. The second opportunity to increase our TAM relates to the enterprises who adopt generative AI models into their product and services. We estimate that new generative AI opportunities will expand our TAM by $4 billion to $8 billion by 2030. Turning to our strategy, which is aligned to both the deep learning and generative AI opportunities. There are 5 pillars to our strategy. The first element of our strategy is to replatform our core internal system for crowd and project management. We are using best-of-breed technology to embed AI as a core capability into our platform, which will add significant automation to our internal operations. This is well underway and on track to be implemented early in the second half of FY '24. Secondly, we're employing a greater focus on using automation in how we create data for our customers. We have recently launched a function in our data annotation platform, called ADAP, that enables large language bottles to be deeply integrated into the data annotation process for greater efficiency and quality controls. The third element is a SaaS platform offering to support enterprises who are customizing large language models for internal use. The focus is to provide a software suite that enables data science teams to connect with internal experts for model training and data feedback. This solution uses Appen's existing data annotation platform. The fourth pillar is a modernized approach to sales and marketing. We have strengthened the technical abilities of our go-to-market team, including the addition of a field CTO to support more technical discussions with our customers. We also now have a dedicated account management program in place for our large customers that oversee existing projects, new expansion areas and overall account health. The final change I have made is to flatten our sales and marketing function. And as part of this change, the CRO and CMO have departed Appen. The final pillar is to implement tight financial controls. We remain laser-focused on managing costs. And as highlighted by the Chair, we are committed to proactively managing costs in line with business performance. Overall, our strategy seeks to deliver a higher quality data for our customers, a better experience for our crowd and improved internal operational efficiencies. I'll now provide a trading update. Before I start, I would like to mention that we will continue to provide monthly financials as we get the business back to cash EBITDA positive. Also, the numbers on the slide are based on our unaudited management accounts. Starting with revenue. You will see that at a group level, our revenue has declined throughout 2024. This decline is driven by the termination of the Google contract, which ended on 19th of March. Revenue excluding Google shows a continuation of the stabilization that we saw in the second half of 2023. We are pleased to see the revenue levels in March and April that are well above the non-Google revenue in Q3 2023. Moving to EBITDA. As mentioned earlier, we achieved our commitment of exiting 2023 cash EBITDA positive in December. However, the revenue loss from Google in Q1 has impacted our profitability. To offset the loss, we took swift action to reduce our cost base and have been able to largely control our losses. Now moving to our outlook statement. As discussed earlier, revenue stabilization has continued through the first 4 months of the year when you account for the loss of Google revenue. We are seeing positive signals on large language model-related growth in 2024, including from our global customers. Tight cost controls remain in place in keeping with the company's focus on managing costs in line with revenue. And in FY '24, we will see the first full year benefit of our $60 million FY '23 cost reduction program. Following the loss of the Google contract, we announced $13.5 million of cost-out initiatives, which was incremental to the $60 million cost-out in FY '23. We have achieved 80% of this benefit in March this year and remain on track to be complete by the end of June FY '24. Our cash balance at 30th of April FY '24 was $36.4 million, and we are confident in our cash position. We remain highly focused on ongoing cash positivity, and our target is to reach cash EBITDA positive on a run rate basis in the early part of -- early half -- early second half of FY '24. In conclusion, I'd like to thank all of you for your continued support of Appen. Appen has played a major role in the development of the industry-leading AI, and I strongly believe that this will continue as generative AI matures. It's a privilege to lead Appen, and I'm confident that we are well positioned to return to growth. I look forward to speaking with you after this meeting. I'll now hand back to Richard.

Thanks, Ryan. And I think that last slide certainly starts to show some positive movement. I would now like to open the meeting to general business questions. You have the opportunity to ask questions pertaining to each resolution when we get to the formal business of the meeting. Before we begin, visitors are reminded, this is a shareholder meeting, and only shareholders, proxy holders, body corporate representatives or attorneys are able to ask questions at this meeting. Visitors have been issued a red card. If shareholders holding a blue or yellow card wish to make comments or ask questions, you should raise your admission card. When called upon, please state your name, or if you're acting as a proxy, identify for whom you are appointed to represent prior to making a comment or asking a question. In the interest of all shareholders, I would also ask you, you be concise in your question or comment, and we'll endeavor to respond in the same way. I will now take business -- general business questions.

Thanks. Peter?

Good morning. I'm Peter Gregory from the Australian Shareholders Association. We're an organization that's largely volunteer operated, independent and not for profit, and we represent the interest of individual shareholders. Today, I have proxies from 17 shareholders for 81,000 shares, which I know is considerably less than we've held in the past. I also note that 70% of the shareholders of Appen own less than 1,000 shares, and probably a large proportion of those now have unmarketable parcels. So that reflects on the very disappointing situation that has been in place last year. And I guess when you think back to last year's AGM, we heard a CEO who was extremely optimistic about the future of the organization, both at the AGM and the investor meeting that followed. We now have, I guess, heard some good indications of the pathway going forward. But I'd like to understand from both a Board level and also a management level, how the culture of the organization is changing to enable it to move from unbridled optimism to realism.

Thanks, Peter. And to be clear that the Board and management also share the disappointment in where the share price has gone in particular. As I think Ryan mentioned, there were some big things happened last year. As generative AI starts to take off there, there was a bit of a reduction in our -- in spend from some of our customers in traditional AI, and that would result in a reduced revenue. And I think the team did a very good job to put in $60 million of cost reductions, get back to EBITDA cash positive by December. And then unfortunately, we had the shock of Google surprise of terminating their contract. And I think Ryan and the team have taken very good steps to find further cost reductions, and we are targeting, as he said, cash positive in the first half of this year -- second -- or the second half of this year. I think the culture at Appen now is very much back to that mindset of entrepreneurial, no excess costs, really focused on driving operational excellence, really focused on our customers, building strong relations with our customers. We had built up quite a big cost base reflective of our revenue a few years ago. And now I think Ryan and the team have really cut that back. So Ryan is very close to the business in both the operational side and the sales and marketing side. And the people running those areas have been with the business for a while, know what they're doing. So it's very much back to the tradition of Appen from many years ago: smart, hungry, entrepreneurial and realistic -- but also realistic. And I think you won't hear from the Board or from Ryan any overpromises. We know the share price will start to improve when we deliver results, and that's what we're aiming to do.

Okay. Can I also ask about Google? The loss in January, as we have all know, has had a pretty major impact on the results. Can I ask a bit more about why Google left? Was it because they no longer needed the services that Appen delivered? Was it because they decided they could do it better internally? Have they gone to a competitor? Has Google decided that the IP that's associated with their data is so important that they need to retain it in house and can't trust an outside organization to be involved? And given the answer to those questions, can you comment on how the other big players, the Metas and Microsoft, whether they're sort of kind of thinking down the same track and whether there's a risk to those businesses leaving for the same reason and how you're mitigating that?

That's a very good question. Thank you, Peter. The Google loss took us all completely by surprise. And the reason it took us by surprise was the working relationship with Google was very, very good. There have been some hiccups in the past, but the working relationship for the previous 12 months have been very, very good, and we've seen an increase in Google revenue coming back. And the people we work with day-to-day, were as, I think, as surprised as we were with the change. We weren't given any reasons for the change other than it was a strategic review at a higher level. So we are unclear about why they did what they did. We believe the work has probably gone to some of our competitors. Potentially, there was a bigger deal done at a higher level within Google with other large organizations, but it wasn't because of our working relationship with Google. That's -- but that does -- is one of the reasons why we think that customer Net Promoter Score is a very, very important thing for us to measure and to reflect in executives' STI. Our relationship with our customers is absolutely vital. And one of the things that Ryan has been particularly doing is building -- rebuilding those relationships. And at an operational level, I think the relationships are very strong. And hopefully, you're starting to see those revenues are starting to increase. So the reason our revenues dropped a couple of years ago was because one big customer reduced their spend. They were spending a lot of money with us, they still spend a lot of money with us, and that relationship is good. So sitting here today, we can't see that happening with any other customers, but it's a very dynamic market.

Okay. Thanks.

Okay. One more from you, Peter, and then -- yes, sorry. Go on.

Maybe you want other people to have a go.

One more from you then.

Okay. Appen has -- I'm oversimplifying, I know here, but in my mind, Appen has 2 product categories and sources of revenue: the services provided by the Crowd; and secondly, the software and technology-based businesses. They're the 2 categories of product, if you like. You also have, broadly, 2 groups of customers: the large global businesses; and then the new markets, enterprise, government, Asia. You have in the Annual Report, on Page 97, what I would describe as a product customer grid. As I look at that, it looks to me like there's no technology products sold to the global customers, and that there's low Crowd involvement in the new, what do you call it, new markets customers. Is that a reflection of the situation? And if it is, can you tell me why? And if it's not, can you look to make that table more meaningful in the future so it gives a fairly simple understanding by shareholders of the product market segmentation that's in play.

Thanks, Peter. That -- so that's the way, from an accounting perspective, our business is divided up. The Global Services is services that we provide to the big global customers where we do work, but it's done on their platforms. The new markets work is work services and software we provide on our platforms. So the vast majority of that work is to enterprise customers. But you'll see from there that there was $9 million of work that was provided to the global customers on our platform. And that's sort of the difference. So the new markets' work is generally enterprise customers and work that we do on our platform. Global Services is work that's done on other platforms. I think we'll certainly take on notice your perspective that maybe it could be a little bit clearer...

Can I just clarify? Does that mean that the Crowd is not being used in the new enterprise market?

No, the Crowd's being used in both of those. It's where -- how the Crowd is managed and how the software works. Sometimes we do it on big customers platforms and often we do it on our platform.

[ Brian Olson ], shareholder. Listening to commentators when they talk about Appen, they say it's very labor-intensive with data generated by hand basically and that sort of thing, and they're casting the aspersion that maybe this won't be necessarily all in the future. So I'd just like to have your comments on that and how you see it.

Yes. Thank you. I mean, certainly, one of the big advantages we have is our Crowd, and we strongly believe that going forward, as generative AI obviously gets bigger and bigger and bigger, there's a real role for humans in, what we call humans in the loop, making sure that the information that's being generated, the data that's being generated has someone checking it. I don't know, Ryan, did you want to maybe expand on that a little bit?

Yes, sure, Richard. It is very important for the human alignment component to be -- to have in generative AI. In deep learning, a lot of the historical work that we've done is to really bring that human-generated data, which was critical to the development of speech recognitions, search systems and other systems that were replicating human-type behaviors. In generative AI, because it's creating content that is human-like, it's very important to have the alignment between the output and what humans expect. So we spend a lot of time with leading researchers and understanding where the market is headed. And there is almost unanimous alignment that the human involvement in the future for generative AI is going to be ongoing and significant.

Any other general business questions at this stage?

Thank you. [ Mehmet Subasic ], shareholder. I want to just give a little context to my question initially and then ask 2 questions. I look to the performance of the business over the last 5 years, including pre-COVID period. We have seen the business being really successful in terms of its overall performance, its scale, market cap, share price and everything. And I personally believe we should have asked the difficult questions when the days were good. Unfortunately, this did not happen. And I think what I'm saying right now goes to everybody who has been with the management for the last 5 years, or with the Board. So please take it personally, I'm not sparing punches here. I think lots of complacency crept in, and we can ask those questions. There has been -- I think there's a bit of falling asleep at the steering wheel and really watching the high-speed train crashing in slow motion. Now this is a big statement because this is a public company that has been really, really successful, and we have to take these things really, really seriously. I think there has been remarkable incompetence in terms of managing revenues and costs responsibly. That did not happen, and there has been massive value destruction. So this is the background to my question. What -- and looking to Ryan's presentation, I see some slight hints of positivity going forward. I appreciate all these things. Obviously, we are sort of at the bottom right now. But we're looking at the business that is almost $300 million in size and still losing money, just too fat in terms of P&L. And cost-saving initiatives are good. $73 million is not easy to deliver. That's easier said than done on an annualized basis, but we cannot cost-save our future to prosperity. That's not going to happen. So I think it becomes more about rightsizing the business. So my questions are, what else can you tell us in terms of revenue diversification going forward, especially for the global customers? I appreciate what's happening in the nonglobal customers and market expansion and so on because we've seen through Ryan the global total addressable market is increasing considerably. So under the light of this and based on what happened in the past, let's judge the business not on a quarter or in a year, but let's have a long-term perspective. What do we do in revenue diversification, global business? And do you see any more opportunities in terms of rightsizing the business? Thank you.

Thank you for your question, [ Mehmet ]. And look, I do appreciate how disappointing our share price performance has been. I would just maybe just disagree with you slightly in terms of your analogy of crash happening in slow motion. When I joined this Board 3 years ago, the company had just grown its revenue by 20%, and that was off the back of many, many years of rapid growth. What happened wasn't in slow motion. It happened very, very quickly. Tech downturn in the U.S., one big customer cut their spend significantly very quickly. So arguably management should have been thinking further ahead. But at the time, they were doing everything they could to manage the rapid growth they were having. When that happened, I think management moved relatively quickly and comprehensively to cut cost base to what it should be. So I would slightly disagree with you in saying that it all happened in slow motion and the management teams out there watching it. I don't think that was the case. Maybe I'll just ask Ryan to comment on your questions about the business.

Yes. Thank you. We are very focused on revenue diversification. We are also very focused on having a longer-term view of where the market is headed and what new opportunities created. The reality is that the market is moving very, very quickly. Our customers are exploring different approaches. They're investing very rapidly in different areas. So our focus is serving our large customers and continue to serve them very well, and we want to be the best vendor that we can for their data needs, but also breaking into new areas in existing customers that aren't as large because we're seeing the signals that they're going to be investing a lot in AI, that is public. There is a huge amount of investment from large technology companies, and we think a good component of that will be on data, and we want to make sure that we are very well placed to capture revenue from a more diverse set of customers, which includes global and many large technology companies that sit within our enterprise segment.

And the only other comment I'd make is that the cost base is large because a lot of the work we do is variable costs. So you've actually got pay Crowd people to do the work. It's not as though our overheads are particularly large now. Any other general questions? Yes.

I'm a shareholder, but I think they gave me a card which is -- I believe because they couldn't find my name. I come from Singapore, and I have been an investor in Appen. And as he rightly pointed out, I have seen the share price fall from $35. Now today, we are at 2%, 2%, I repeat, less than 2% of the value in 2020. So we cannot attribute our problems to only a loss of business in Google. Something has drastically gone wrong somewhere in terms of perception of the company because the revenues are not as bad as they turned out to be. Yes, EBITDA is bad, but the perception somewhere is lacking. And market is only about perception. In the end, all the effort that you're making is for a market cap. If you cannot achieve a good market cap, all these efforts are in vain. In my opinion, and sitting in Singapore, the perception of the company is bad. I don't know where it is lacking, but some effort has to be definitely put in to improve the perception of the company. Now for example, as you earlier pointed out, could be a regular monthly communication in terms of new businesses achieved, in terms of new developments in the company. See, we report our results every 6 months. So as a shareholder, I cannot wait for 6 months waiting for your results. So you have to do because you have already lost it. You have come to 1.7% of the value of 2020. I give a benefit, okay, 2020 was not the right price. Okay. It was an exceptional price. Still, 1.7% is a pathetic performance in the marketplace. And it is not as bad in reality. So something is wrong somewhere, where the CEO has to regularly now meet the investor, meet investment bankers, talk to them, make presentations and sell the company. Effort has to be made on that front also apart from what businesses you -- or the effort that you're making on the business front. But in the end, it is all about market cap.

And look, I would -- in response to that, I again repeat that we're all incredibly disappointed with the share price. And clearly, there has been a huge change in -- there's been a change -- to your point, there's been a change in the performance of the business, but there has been a huge change in how the market perceives the business. I think as I said earlier, Ryan does meet with an investment bank as he meets with investors. He spends a lot of time, more time investing -- meeting with customers, which I think is very important. Where we are now, the way we'll change perception in this business is to deliver results. And that is what the team is incredibly focused on delivering results. And I'm hoping that as you see results start to go the right way, that will be reflected in the share price over time. Here, and then back to Peter.

My name is [ Chris ], and I am a shareholder. I'm almost holding 150,000 shares with Appen. I've been witnessing all the highs and downs. First of all, I would like to thank to the entire team members of our team, like congratulation and doing good. Like I know it has been paying hard going all the way down and coming up. And I have a big faith in you guys, which we will be doing absolutely good. And especially Ryan, like congratulations, I would say. My question is -- just to give a bit of 2, like one is like, are we going back like try and convince Google and maybe see some possibility to keep them back in and sell our products? One. Secondly is regarding telco, I think there has been lots of rumors and gossips going on. If you can just please a little bit about the telco would be appreciated. And once again, thank you very much for everything.