having held a large parcel from $3 up to the top and watched it...

having held a large parcel from $3 up to the top and watched it fall back here to $8.50 (and added a bit at circa $20 then again around $13 from memory), ive again added more today at $8.50.

ive no idea where the SP will go in the next few months, and i agree the CEO and company could have done better in terms of articulating their strategy recently.

my purchase today was based on the following investment principles:
1) EV/EBITDA below 10x (MC now around USD800m / EBITDA ~ USD 80m), which is pretty cheap for any company, let alone a company that is the global leader in one of the megatrends that will define humanity over the the next 10-20 years, ie machine learning)
2) its oversold, i.e. sentiment is nearing or has reached the point of capitulation, where many impatient holders who said they would 'ride out the rough patch' are panicking after a tough 6-12 months and 'cutting their losses', including no doubt the fund managers who are worried about short term underperformance and losing investors to index funds (lol)
3) they are the global market leader in their field (data annotation for search, image, speech, autonomous driving etc), a market which is predicted to grow at CAGR 30%+ for the next decade
4) they havnt lost any of their big tech clients, the projects are just shifting, but even the old projects eg search will require constant updating as datasets (i.e. internet content for search) continuously evolve, so the target for ML algorithms
5) those new projects will inevitably require large datasets for training as they mature, for which Appen is one of the only companies with the massive scale (millions of on-demand crowd annotators) to deliver. I work in statistical research and one of the fundamental truths of ML and model building is that the strongest determinants of model (i.e. ML algorithm) accuracy is the size and accuracy of the training dataset and validation/ calibration datasets. i dont see another competitor providing a superior alternative proposition to APX to drive the tech companies to change their dataset provider, especially given the strong track record and trust APX has established in these customer relationships. in fact, i think part of the reason APX often comes across as 'vague' in their presentations and investor calls, is because they value discretion and know their customers demand confidentiality in terms of new projects and directions, thus they cant speak freely about their projects.
6) their business model doesnt have the rapidly expanding margins of scale (unlike say promedicus) because their margins are modest; they are essentially a specialised service provider providing labour-intensive human data annotation; by definition they cant completely automate the data annotation, because it must by its nature be human annotated datasets. They are, however, working to enhance and 'partially automate' the process to speed it up and reduce the manual workload required by their crowd of annotaters; they are also enhancing checking systems to reduce annotation errors in order to optimise the accuracy of the dataset. They also have the advantage of using an 'on demand' crowd on a casual basis so that as work temporarily fluctautes to the downside, they can simply 'bench' their crowd of annotators which means their fixed cost base is low in terms of labour and this reduces the cyclical downside risk (i.e. unlike say a miner or Kogan recently, they dont make a loss due to fixed staff or inventory costs, when sales cycles fluctuate to the downside briefly).
7) for those who arent aware, there is currently a war emerging between the tech giants in terms of fighting for 'ad spend'. google and facebook have enjoyed a comfortable duopoly in terms of dominating global advertising spending for the past 5 years. now, they are under attack from the emerging tech companies who also are gaining a larger share of 'eyeballs' and 'facetime' of users on the internet who can thus attract advertisng spending away from them. key players include tik tok (bytedance) who now have over 1billion users who spend a larger time on their app, as well as amazon who are attracting direct advertising spend within their website/ app/ platform, alibaba for the same reason, as well as other players such as spotify through music and podcasts, etc etc. all this means that facebook and google need to invest more heavily in their search and relevance algorithms to compete and to maintain superiority.

i could go on, but i dont want to overburden people with my thoughts.

anyway, the point is that the SP of APX suggests it is ex growth, which is true on a fy21 vs fy 20/fy19 basis, but im my humble opinion is rubbish on a 3-5 year view.

i think this is a great long term risk reward entry point. an even better price may be offered in coming weeks or months, but $8.5 is compelling enough for me to take a parcel today! the key now is patience. im flipping my sentiment back to buy from hold.

FYI i also bought a big chunk of alibaba HK9988 today at deliciously discounted prices below usd 140 per share on a p/e of around 15 for a stock growing at 30%+ in a massive 1bill population that is early in their transition to an advanced consumption society, in which alibaba is the dominant market leader as the number 1 online marketplace (aka amazon). enough said...ignore the short term noise, even if evergrande goes under, alibaba will be a lot more expensive in 5 years time and common prosperity will support not reduce their long term value proposition.