No. of Recommendations: 5
Selling ads is not a moat. The business seems about as sans-moat as they come.
Ad revenue is 100%? It was like 97% a few years ago - so they're less diversified?
The ads are all of the revenue but, as you point out, the ads are not the moat. This is the case with Google, or any ad business. There is a business in place that takes nearly all the expense, which is not not the "product". The actual "product" is the public's attention. That's what sold to the "customer" which is the ad purchaser. So the expense goes to creating the moat, from which the product (you and me, we're the products) is sold to the customer (those placing ads). That moat that takes nearly all the expense, in teh case of Meta, is providing a free service to communicate/publish. That is where the moat is attempted to be developed. (so they merely sell that "product" (you and me, etc) to the "customer" (the purchaser of adverts) sitting on the moat, which is seen as invisible to some financial analysts wanting to find a moat in the revenue device itself.
In the case of newspapers, it's the same. But now the moat was the distribution and daily habbit of users buying newspapers for below-cost, or free. In the case of Google, it also the same - having the best search, along with daily habbit of people used to using that search. Then selling ads on top of that. In the case of Meta, it is wanting (or needing) to stay in touch with friends over instagram/whatsapp/threads/(insert future product altered/coded, and added to their social networks).
For newspapers, the moat of distribution dominance didn't erode from competition, but rather the model of paper being replaced with online print. So that killed their moat. Those purchasing newspapers were looking at News Limited's moat as the dominance of the network, and got it right that competition didn't kill the moat from other newspapers.. but got it wrong that the moat itself would evaporate from competitors entering from a different domain (online print) which newspaper gurus were not looking carefully enough at.(*)
For Meta, I believe their moat is not only far more dominant than any newspaper ever reached in the past (News Limited under Murdoch being one of the larger), but profoudnly less fragile and more organic. I wouldn't have observed that 10 years ago, but their ability to merge/create products in their network the last 10 years has proven to me that they are extremely resilient to moat-evaporation.
Google gets 101b views/month, YT gets 53b (April 2026). FB and insta get 10% of that.
FB ads are garbage, and all meaningful content (that is, not cats and people doing dances)
is on YT - all the finance channels there get much more traffic than FB things, and all
the real how-to videos are on YT.
The quality of ad targetting today with all the leaders (Google, Meta, etc) is not great but the actual important thing for investors is that it is both massively able to be improved, and improving. With meta earnings $230 billion today, at high margin, and the ad targetting not great - then imagine what it would earn when the targetting is actually far more effective?
When you set up an ad campaign with Meta, you usually set it to run across all the services. It defaults to all, and you can untick based on your targetting. That includes:
1. Threads
2. Instagram (Feed, Stories, Reels, Explore pages,
3. Messenger (main chat tab (inbox) or as sponsored messages).
4. Facebook (Feed, Stories, Marketplace, Video Feeds, Right-Hand Column placements)
5. Meta Audience Network (third-party mobile apps and websites that have partnered with Meta to display their advertising)
6. Whatsapp, future services etc.
Their ad business has grown to catch up to Google's recently and Meta has the faster rate of trend-growth for ad revenue so they'll be likely to surpass by quite a bit. Google is a fantastic business also.
PE of 21 is *relatively* cheap, but so is a NY strip at $20/lb when filets are $40/lb.
It depends on the earnings 10 years away. For the case of Meta I would be very surprised if earnings per share were not 4x (nominally) what they are now, in 2036. The earnings multiple doesn't need to rise to have a great return.
CEO is weird. Weird bunker building in Hawaii, weird annoyances of neighbors
in CA real estate. Made awful decisions with the metaverse, and the board,
if it was functional (which it is not, it has no power) would have tossed him.
His style is offputting to say the least, Metaverse flopped - though he pulled out without anything catastrophic happening relative to their overall scale. What two things I do like about the CEO is that (1) he is willing to admit he is wrong (not so much in words, but in actions) and chance course when he makes a bad move, rather than being stubborn and not changing a bad decision - this agility is extremely important. (**) Making staff numbers tighter, despite the temporarily lower morale, etc. (2) He is constantly searching energetically for new expansion ideas/opportunities, rather than being conservative - this is also important, as Phil Fisher would like to underline also.
His voting share rights are obscenely skewed against you and I retail investors.
I would expect that if there was a more diverse board, the company would not necessarily be stronger for it. Shareholders are sometimes, contrary to the usual believe, shorter-term in their thinking than owner-management, and can vote against investments that have zero return for the next 3 months, but a huge return on investment over several years. That was the case with Google in the past, when it had massive cloud investments that the shareholders were skeptical about 2 years ago, and revising their memory (most accurately as "In the past thought it was poor spending, better if they bought more stock") to a newer version, inconsistent with past beliefs, but consistent with the present facts (that "I always thought is was obvious thing to spend so much capital on building the their AI models/infrastructure").
I have a a couple of friends who own businesses purchasing a huge amount of AI compute since Feb this year. These two firms are sepnding a lot of money on AI (presently towards Google, but they'd be happy to flip to Meta if they offered their computer to customers also, which I'm almost certain they will) not because of hype, but because they are getting immediate customer benefits thar far exceed the LLM expense. They'll be some misspendings, but overall I expect the benefit to such mid-layer-businesses (online shops, brokerage software, etc) wanting LLM access will remain quite larger than the expenses itself. What happens, though, as with past manufacturers forced to upgrade to a new technology (where you go out of business if you don't upgrade) is that the mid-level businesses don't beneift so much (as they all upgrade so competitive position is the same as before) and the main beneficiaries are the tech sellers, and the end-custom actually using the products created by the mid-level businsses.
- Manlobbi
(*) and (**): Note the overlap of these two points. You want a CEO that is paranoid about any competition, other sources of keeping the public's attention (this is the environment for revenue, not the revenue itself), and a CEO that biases towards too much activity (erroring on getting it wrong sometimes, rather than erroring on just being conservative).
