At the beginning of this year, it was reported that Appen's stock price soared from $0.26 to $1.08 within just five weeks. However, I remember the ticker symbol APX, once a part of the WAAAX stock portfolio, with prices once exceeding $20. While other stocks in WAAAX have seen impressive gains in recent years, why has Appen performed so poorly?
I believe many investors in the ASX, especially those interested in tech stocks, might share my curiosity. Hence, I've decided to organise and share my research materials. However, please note, these are not meant to serve as buy or sell recommendations; do your own research (DYOR).
Appen is a company specialising in data annotation, categorised under the AI industry because whether it's traditional data categorisation or the currently popular large language models (LLMs) like OpenAI's ChatGPT, Meta's Llama, Google's Gemini, or xAI's Grok, they all rely on human-annotated data for training.
Appen's core business is providing human data annotation services. These annotators, often referred to as the "crowd" or crowd-workers, are spread globally. The aforementioned large tech companies frequently assign tasks to Appen, which then locates these workers to complete the data labelling tasks. From financial reports, it can be seen that Appen's gross margin has decreased from around 50% in earlier years to about 30% now. This decline is mainly attributed to increased benefits for the crowd, including better local work environments and activities beneficial to their physical and mental health.
Appen's IPO took place in December 2014, with the company issuing 30 million shares at $0.50 each, reaching a total of 94.6 million shares post-IPO and a market capitalisation of $44.9 million. In FY13, prior to going public, Appen recorded revenues of $60.5 million and a net profit after tax (NPAT) of $4.0 million.
From 2014 to 2021, with the rise of artificial intelligence and machine learning technologies, the global demand for high-quality data annotation significantly increased. Particularly during the COVID-19 pandemic, due to increased demand in advertising services, Appen experienced explosive growth. In 2018, revenue grew by 112% year-over-year, and in 2019, revenue further increased to $535.999 million, a growth of 47.1%. During this period, Appen not only solidified its position as a leading provider of data annotation solutions but also acquired Leapforce in 2017, Figure Eight in 2019, and Quadrant in 2021. Its stock price also saw significant increases, making the company a star on the capital market.
The company also demonstrated strong profitability, with dividends as follows:
2015: $0.042 per share
2016: $0.05 per share
2017: $0.06 per share
2018: $0.08 per share
2019: $0.09 per share
2020: $0.10 per share
2021: $0.10 per share
Total dividends: $0.512 per share
On October 28, 2021, Mr. Vonwiller, Chair of the Board at Appen, retired, marking a turning point for the company. In 2022, revenues declined by 13%, and the company reported a loss for the first time, resulting in no dividend payout. In January 2023, Mark Brayan, CEO and Managing Director who led Appen to its heights, announced his departure, with Mr. Armughan Ahmad succeeding him.
In 2023, with the success of ChatGPT and the development of General AI (Gen-AI) technologies, the demand for manual data annotation decreased. Against this backdrop, the newly appointed CEO, Mr. Armughan Ahmad, took proactive measures within his first year. The company invested GBP 2 million in Mindtech and USD 0.5 million in Reka AI, beginning a transition towards a LLM platform. Despite these efforts, the situation did not fully turn around. Under his leadership, the company had to raise capital twice and lost Google, a major client contributing $82.8 million in revenue.
In February 2024, Mr. Ryan Kolln, who joined Appen in 2018, replaced the underperforming Mr. Armughan Ahmad as the new CEO and Managing Director. At the recent Annual General Meeting (AGM), Mr. Ryan Kolln reported improved performance. Despite the loss of a major client like Google, the company's revenue has been stabilising, maintaining over $15 million per month. Additionally, the cash balance increased from $32.15 million as of December 31, 2023, to $36.4 million by April 30, 2024.
In exploring the demand for data annotation in the era of LLMs, I purchased a second-hand graphics card and began using the llma3-8b model. Through this experiment, I discovered that custom datasets are still necessary for specific environmental optimisations. These datasets are transitioning from previously low-cost, non-professional levels to more professional areas, such as the legal services industry. This shift not only reflects the increasing depth and complexity of data annotation work but also its ongoing demand in professional fields.
I am grateful for all the rational discussions in the HC, especially to
@Calvo and
@Chrisinvest, whose continuous focus and analysis on Appen have greatly aided my understanding of the company. Seeing Appen's stock price fall from its peak, I deeply empathise with the loyal shareholders' disappointment and frustration.
As
@Phillipho pointed out, while some participants do add to the liveliness, their contributions often lack substantial usefulness. Based on this, I have chosen to ignore such noise.
Finally, according to the annual report released in February 2023, the number of shares held by Mr. & Mrs. Vonwiller has not changed since the company's listing. Although the dividends they have received now exceed the share price at the time of listing, this is a well-deserved reward for their efforts, isn't it? They have consistently supported the company, and I am deeply grateful for their commitment.
Thank you to every participant in the HC, whether bullish or bearish on APX. After all, investing is a personal decision, and finding the right approach that suits oneself is key. I wish everyone good luck in their investments.
(20min delay)