Profitablity per distribution channel

Atlas is now selling its pearls through 3 different channels :
- regular auction (Kobe), with an average price of 73 $ during H1 25,
- low grade auction (Bali), with an average price of 21 $ during H1 25,
- other channels (probably mainly retail), with an average price of 155 $ during H1 25.

As the average cash cost was 42 $ during H1 25, we can see that :
- low grade auctions are largely loss making (loss* of 21 $ per pear),
- while other channels have a high profitability (gain of 113 $ per pearl, vs 30 $ only for regular auctions).

During H1 25, 32 % of pearls were sold via low grade auctions which generated a loss of around 1.9 m$ (vs a loss of 0.4 m$ estimated during H1 24).
This was significant vs a total EBITDA of 7.1 m$.
This loss has partly hidden the great performance the company had with non auction sales.

Non auction generated 5.7 m$ of revenues (30 % of total revenues) and 4.18 m$ of EBITDA (59 % of total EBITDA).
So, despite the large price decrease, ATP was able to increase its non auction sales by 43 % (+ 107 % for volumes) and its non auction EBITDA by 26 %.

One of the surprise has been that they have been able to increase their non auction sales without affecting much their cash cost / pearl (+ 6 % yoy).

Going forward, we can imagine a scenario :
- where the loss from low grade auctions will decrease, as the low grade auctions represented 32 % of the volume sold in H1 25 vs 15 %e in H1 24,
- while the company will probably try to increase the share of pearls sold out of auctions.

Developing the retail channel has a double interest for the company : increasing its earning potential, but also largely decreasing its cyclicity (as retail prices are much higher than wholesale prices and should continue to cover cash costs even at the low end of the cycle).

* I define profit/loss per pearl as average price - cash cost.