The most intriguing aspect of the recent August reporting season was the stellar returns of a small group of technology-related companies. In the month of August alone, this group returned an average of 37%. For the twelve months to 31 August, their individual returns were: Wisetech Global (WTC) +168%; Altium (ALU) +184%; Appen (APX) +228% and Afterpay (APT) +381%.
However, in the recent equity market sell-off, technology stocks have been the hardest hit, with Wisetech down 18%, Altium down 19%, Appen down 25% and Afterpay down 31% since 31 August. Arguably, the sector was over-valued before the recent sell-off, but these price moves provide an opportunity to re-examine these stocks. Among them, Appen (APX) is the only company on a price to earnings multiple below 30x. Partly for this reason, we have undertaken an analysis of the stock, summarised below.
What does Appen do?
Following the acquisition of Leapforce for $105 million in November 2017, Appen is the global leader in the provision of human annotated data sets to inform artificial intelligence algorithms. Essentially, Appen has access to a global crowd of over one million people and is able to use that crowd to, for example, test the relevance of search engine results. Individuals recruited to a project might review specific search engine results for queries such as “What will the weather be in Melbourne tomorrow?” and rank the relevance of different results. That test data is used to improve the algorithms which guide computer systems, particularly in the areas of search, social media and e-commerce.
As the CEO, Mark Brayan, described it at the time of the Leapforce acquisition, “Both businesses are effectively labour companies, selling highly specialised datasets that computers are able to interpret and learn from, building a machine’s understanding of written and spoken speech, as well as image recognition.”
Central to Appen’s customer proposition is the quality of data provided. Appen ensures data quality through initial testing of the crowd, providing extensive guidelines and managing the output through technology overlays. Quality of data is the number one factor considered by potential customers in training data procurement.
What makes Appen potentially interesting as an investment?
Improving technology is opening up ever-increasing applications for artificial intelligence in almost every industry. A McKinsey Global Institute study estimated potential value creation of between $3.5 trillion to $5.8 trillion from artificial intelligence, as illustrated overleaf.

Source: Appen, McKinsey Global Institute
Added to this, so-called “deep learning” algorithms are at an early stage of adoption and benefit from much larger data sets than traditional machine learning algorithms. This presents a further likely driver of increased demand for Appen’s services.
Against that backdrop, Appen is the global leader in the provision of human annotated data sets. It services 8 out of the top 10 global technology companies, including Google and Facebook, albeit on a non-exclusive basis. This is a clear indication that the business provides high-quality services within this sector.
This is further highlighted by the financial track record. For the three years to December 2017, Appen delivered a compound annual growth rate in sales of 48% and net profit of 76%.

Source: Factset

Source: Factset
What is the company worth?
We know that applications for artificial intelligence represent a huge and increasing market and that Appen has been a beneficiary of this growth to date. However, forecasting with any accuracy on even a two to three-year time frame is very difficult.
Moreover, there are identifiable risks. Approximately 90% of Appen’s revenues are generated from only 5 very large, well-resourced customers. These customers could demand lower pricing as volumes increase, which did occur with one customer in the second half of 2016. Further, Appen has a number of significant competitors and could potentially lose a large customer to one of these competitors. A third risk is relationships with the contracted crowd, as highlighted by the multiple lawsuits which Uber has faced concerning its relationships with its contracted drivers.
To illustrate the valuation uncertainty, we ran a discounted cash flow scenario analysis. The key variable in this analysis was the number of years before Appen’s recent three-year growth rate slows to a market average type growth rate, in line with nominal GDP growth. If this time period is three years, we estimate that the company is fairly valued on its current CY2019 PE of 27.0x. However, if the time period is five years i.e. a more extended period of very high growth, then the company is 65% undervalued at the current price.
We met with management following the recent result and were encouraged by the growth opportunities of the business and the near-term outlook. Our conclusion at this stage is that we have more questions to have answered and work to do to reduce the above-mentioned valuation uncertainty.