September had a lot of ups and downs; end ended with a lot of downs. My high-water mark, my portfolio was up about 47% YTD. However, the last week of the month really hammered me. Ended the month down half a percent.
I know my portfolio is a bit different than many here. For example, I don’t own some board favs like Lightspeed & Cloudflare. At the same time, I hold decent sized positions in some that aren’t found here as much (Sea Limited, Snowflake, Roku, Fubo).
I REALLY like Lightspeed, Monday and Cloudflare. If I had 10 positions, those would probably be my last three. However, I have a very difficult time selling any of my top 7 conviction stocks because my thesis is still place.
Also, I don’t generally have smallish positions (under 10%). The reason I don’t like under 10% positions as much is that I want my positions to have a nice impact when/if I’m right.
That said, a few of my positions have been basically “dead” money for many months. I wrestle between what I see as the inevitable long-term gains and the fact that these positions weigh me down while others like NET, LSPD & MNDY are doing well.
YTD
37.6%
September
-.5%
Monthly performance Allocation Aug Allocation
Upstart 38.1% 20% 12%
Sea Limited -5.8% 17% 18%
Data Dog 2.6% 16% 16%
Roku -11.1% 12% 14%
Crowdstrike -12.5% 12% 18%
Snowflake -0.6% 12% 12%
FuboTV -17.8% 10% 11%
Transactions:
I didn’t make many transactions in September. I did sell some Crowdstrike shortly after their earnings. I put 2/3 of that money into Upstart and the rest into Fubo. This brought Upstart to my 3rd/4th largest position. Then the rest of the month, UPST ran up which made it my largest position. I’m glad to have it at the top.
Position comments: Focused on not as heavily held ones
Sea Limited
This stock is a MONSTER and they continue to have great news. Their recent announcements about entering the India and EU markets are very exciting. They have proven to execute magnificently! They keep chugging along at triple digit growth rates and that appears to be the case for the forseeable future. They took a bit of a hit due to raising capital with a small dilution of shares. The purpose of this is for cash needed for their expansion. They have a HUUUUUGE opportunity in front of them and are executing on it.
Roku:
Roku has shrunk from my largest position to my 5th largest. This is not due to me selling, rather due to the fear of their hardware growth pushing the price down and other positions of mine doing well (especially Upstart, DataDog). They reported a great quarter with 81% revenue growth. The platform revenue is over 82% of total revenue. This will keep climbing.
Advertising dollars continue to flow to CTV. Roku is best positioned to capitalize on this. The other day I was watching The Roku Channel and for the first time, saw an interactive ad. I wasn’t able to move fast enough to do the interaction as I was in the middle of feeding the family. The fear of declining hardware sales misses the big picture of the advertising. This is similar to the fear Apple had for years when ipad & iphone sales declined. Wile they ARE winning the CTV platform wars, they don’t HAVE to in order to be a huge winning company. If you’re on twitter, @FromValue does an excellent job explaining this.
Fubo:
Another CTV position. Sometimes I think I should consolidate and only have 1 CTV position and move those funds to SaaS. I’m not sure which one I would choose. I used to have 3 but sold out of Magnite. However, the future of both these companies is VERY bright. Fubo is EVERYWHERE. I see their commercials EVERYWHERE. From announcers during games doing Fubo promos to bigger commercial pushes. Their recent deal with the New York Jets (who play in New Jersey, which is one of the markets they’re approved in) is another in a line of shrewd moves. Q4 is when they’ve been talking about their sports betting coming online. In hindsight, I probably should have had a smaller position in them until about now, knowing the sports betting was coming.
One thing to realize is that they are placed to grow and do very well with just their video streaming and advertising businesses. HOWEVER, the sports betting is what will push them over the top.
Snowflake
Recently was a darling on the board but fell out of favor after posting 104% YoY and 100% annualized QoQ (almost 19%) growth. Guidance is along the same but slightly lower that previous growth clocking at about 9.5% QoQ as opposed to the previous guidance of 10.5%. Their revenue retention rate ticked up to an UNHEARD of 169%. These are all FANTASTIC numbers.
As many of you know, the reason for the falling out of favor here is mostly due to the falling RPO Remaining Performance Obligation number. This number annualizes to the low 30% range. I guess the fear is that revenue growth will suddenly shoot down instead of the normal gradual (see CRWD, DDOG before its recent re-acceleration). Perhaps some fear of the usage based model and linking that with Fastly. I dunno. I think Saul said it best that a lot of these fears are overblown but he’s more worried about the complexity of seeing the future here, and that’s fair. There are a lot of us on this board with a background in data, big data, analytics etc… This is my background/career as well. I see the vast majority of people with this background and a good understanding of data and big data staying in Snowflake. Perhaps that removes some of the complexity for me.
To me its pretty simple and its really 2 pieces. Data explosion and reporting complexities.
- Data Explosion I’ve said this example before. A decade or so ago, I started working as a Database Administrator for a small ECommerce company. When I started, the size of the main database was about 10 GB. All databases combined was maybe 12-13 GB. I left that company a bit over 5 years later and the data sizes went from that number to over 2 TB. 200x data growth! In 5 years. One small company. I’ve since moved on to consulting. This experience is typical.
- Query complexities… My background is as a database administrator. I specialize in improving performance. People in my field have a joke that goes something like this (its a joke because its not possible to do this). “I CAN FIX THAT PERFORMANCE PROBLEM, GET RID OF THE DATA.” When there is more data, it takes more power to execute the same queries. It is relatively rare to find a software engineer who really understands this. One story to illustrate this. When I started at this company a decade ago, there were times throughout the day where the e-commerce website would be unusable; essentially causing an outage. Nobody knew why because the code had not changed. When I dug in, I found very poorly written queries that was the main cause of the issue. A simple modification made the query go from eating up ALL the resources (CPU, Memory) to it being barely detectable. The code worked fine when there was not much data, but when you add more data to it, it brought down the website.
I work in consulting now and have worked with hundreds of companies, big and small. I have seen software companies and the quality of their products. To put it nicely, it is VERY rare to find a well written application. There are some. I’m usually surprised to find some of the more simple performance best practices being utilized. As we know, as software is very sticky, its difficult for a company to change their platform. So what happens when the application isn’t performing well? The most common and fastest way to ‘fix’ the problem is to throw more hardware at the problem. The traditional method would be to get more CPU, more Memory, faster disk. This is a VERY expensive option, however, it is worth it for a hospital or whatever to make sure that their Radiology software remains usable (this actually happened to me not long ago).
So how does this relate to Snowflake? Well, Snowflake is BRILLIANT on how they take advantage of this. Its more like death by a thousand papercuts. I call it their EASY button. To make a particular code run faster in Snowflake, click a button. Now, every time that code runs, it’ll be DOUBLE the price. Now, 6 months down the road, when there is even MORE data to run that query on, the query slows down some more. Oh, click that easy button again, now we’re paying 4x the original cost. Its kinda like morning classes in college. The first time you skip a class to sleep in, the easier it makes it to do the next time…
THIS IS WHY RPO IS CHEWED UP FASTER THAN CUSTOMERS EXPECT.
So… to wrap Snowflake up… What are the chances revenue growth is gonna quickly drop off?
I always start these writeups thinking it’ll be a shorter one this time.
Good luck to all in October.