About 40 years ago, Warren Buffett said it best:
Disintermediation: The End of Ad Agencies?
This past week: a major advertising agency holding company, WPP, announced worse than expected results and lowered its long-term earnings per share growth guidance. The stock dropped a lot. I think it had its worst day in about 20 years. However, unlike some of the other trends we’re talking about here – there wasn’t a big spill-over into other ad agency holding company stocks. For example, Omnicom shares are quite a bit higher than the low they hit of around $65 sometime last year. Following WPP’s poor results, I got a lot of emails about the trend toward “disintermediation” in advertising.
To explain the concept, here is a line from Facebook’s 10-K:
“Marketers pay for ad products either directly or through their relationships with advertising agencies, based on the number of impressions delivered or the number of actions, such as clicks, taken by users.”
In the past year, some of the world’s biggest ad agency holding companies have paid billions of dollars to Facebook and Google on behalf of clients. As was the case in the age of classified ads in newspapers, small ad buyers always buy directly while some of the world’s biggest brands buy through ad agencies. While agreements between agencies and clients can vary in terms of how the agency makes money, buying ad space on behalf of clients can do two things for ad agencies: 1) It creates commissions and 2) It creates “float”.
Disintermediation is a fancy term for cutting out the middleman. In the case of brands buying on Facebook and Google properties – this includes things like YouTube and Instagram – it can mean three things. One, the company that owns the brand (for example, McCormick owns French’s mustard) could deal directly with Facebook or Google in a special, privileged way. In other words, people at McCormick could develop a relationship with people at Facebook in some way and this could result in a deal that helps raise awareness of the French’s mustard brand on Facebook. Two, the company that owns the brand could buy ad space directly in a non-privileged way via auctions in which all sorts of bidders participate. This saves the company on commissions and may improve working capital use. And three, the company could in-house certain activities related to running a digital advertising campaign. Digital ad campaigns had historically generated the most dollars of revenue for ad agencies per dollar going to the media outlet showing the ad. In years past, we know the same amount of client spending could generate 2-3 times more revenue for an agency if the money was spent on digital instead of TV.
Based on WPP’s most recent results, this last form of disintermediation is the one we see hitting ad agency results now. Giant corporations that have big brands are doing more digital work in-house. This doesn’t rule out the possibility that bigger and worse forms of disintermediation are coming next. But, it’s the in-housing of labor intensive work related to digital advertising that is showing up in 2017 results.
The Duopolists: Facebook and Google
When you take all of Facebook and Google’s properties together they account for a large share of the U.S. online audience each day. Last year, the two companies may have had 60% of the digital advertising market in the U.S. That’s obviously not as high as what CBS, NBC, and ABC had in the U.S. TV ad market 50 years ago (often 90%+). However, Facebook and Google do have similar positions in many other countries. And you have companies like Tencent with big positions in countries where Google and Facebook are not major factors. Of course, when you move down the list past Google and Facebook – ad share drops off completely with a very fragmented market for the other 35% or so of digital ad spending in the U.S. In fact, the other 35% to 40% of the digital media market that doesn’t go to Google and Facebook is much more fragmented than other forms of media had been. There were never 10 TV stations, radio stations, billboard companies, etc. that had 2% market share in a local area. On the internet there are such small players.
The big difference between advertising on TV 50 years ago and advertising on Google and Facebook is the targeting tools that Google and Facebook provide. These companies are moving toward disintermediation simply by making some of the targeting that was part of an ad agency’s job something anyone can do using a lot of technology to help them. In fact, that seems to be the strategy at Facebook where the company will depend on increasing ad prices (which requires increasing ad effectiveness) rather than increasing users, increasing time spent per user, or increasing ads shown per user per hour.
So, the trend is toward disintermediation. Does that mean an investor should avoid ad agency stocks and buy the shares of Facebook and Google?
Handicapping: Why Betting on the Favorites Often Fails to Pay
For investors: The problem with doing that is price. You can be right that Google and Facebook will increase their share of the digital ad market and digital advertising will increase its share of overall advertising, and yet your bet on the winners out there in the world of business may actually underperform some other investor’s bet on the losing business.
Because, you’re placing your bet in the stock market. And the stock market offers tremendously different odds on ad-related favorites like Google and Facebook versus an ad-related longshot like Omnicom. How much optimism is baked into the stock prices of Google and Facebook? How much pessimism is baked into the stock price of Omnicom?
Here, we’re talking something like the difference between getting 2 to 1 odds on Facebook and Google versus 8 to 1 odds on Omnicom. There’s no real way to translate a stock market bet into horse racing terms, because over very long time horizons stock market bets pay off differently than horse races. Short-term value bets are easier to think of like a horse race. Long-term growth type bets are hard to think of as a horse race.
Here, as I’ll show you in a moment – we’re lucky enough to know that Facebook and Google can’t actually grow for very long. For growth stocks: These businesses are actually very, very close to total maturity. We’re analyzing an unusually short race here. This isn’t the Belmont we’re handicapping. It’s a sprint.
Outrunning Multiple Contraction
High growth businesses trade at high multiples of free cash flow. Low growth businesses trade at normal (or lower) multiples of free cash flow. Over time, high growth businesses become slow growth businesses. So, over time growth stocks see their price-to-free-cash-flow multiples contract.
As you can see from the table, if you bet “with the market” on the bright future for Facebook and Google and decide not to bet “against the market” on the bleak future for Omnicom – you have to pay for it. What do you have to pay?
Well, taking the high growth side of the argument forces you to also take a series of positions that tend to do badly in the market. Here, betting on Facebook and Google means betting on stocks that already have market caps of $500 billion to $750 billion. In addition to that, their price-to-free-cash-flow levels are too high for a slow growth stock. So, the dollar amount of revenue growth you’d need at these companies to give you a market beating return in the stock market is especially high. At Facebook and Google, you’re going to need revenue growth that’s even faster than the annual return you want in the stock. You want a 10% stock return? You’ll need more than 10% a year revenue growth to get there. You want a 15% stock return? You’ll need more than 15% a year growth to get that. And so on.
Let’s say a no-growth stock has a price-to-free-cash-flow of 15 (that’s a 6% to 7% free cash flow yield). In reality, the price-to-free-cash-flow you see here is not adjusted for stock option grants to employees. So, that 15 times price isn’t as cheap as it appears at any of these companies (including Omnicom). If these companies give away 1% to 2% of the business to employees each year – that comes right out of your annual return as a shareholder.
So, what size would Facebook and Google have to grow to in 5 years, 10 years, etc. for your investment to work out here? Remember, if we know the price-to-free-cash-flow multiple is going to contract at some point, then we know free cash flow has to grow faster than market cap – and you are only going to make money (unless the company buys back stock or pays a dividend) from market cap growth.
If you want a 15% return in Google and Facebook over the next 5 years, free cash flow at these companies needs to grow at a little over 30% a year. That’s if the companies end up trading at a price-to-free-cash-flow of 15. Many people reading this will believe Google and Facebook will never trade as low as a price-to-free-cash-flow multiple of 15. However, the continued success of these businesses pretty much guarantees they will one day trade at such mature company multiples.
Now, I don’t think investors are actually betting on free cash flow growth of 30% a year at these companies – this would mean, Google and Facebook would have to quadruple their free cash flow in just 5 years. What I think people might be betting on is:
1. A return lower than 15% a year in these stocks
2. A holding period longer than 5 years in these stocks
3. That Google and Facebook will always trade above a price-to-free-cash-flow of 15
Point #3 is wrong. It doesn’t feel wrong now, but it’s wrong. Google and Facebook are very fast growers – so, it feels like they should be fast growers forever. However, unlike something like Amazon – Google and Facebook will run out of fuel pretty fast if they continue to grow. All of the profit at Google and Facebook comes from advertising. The two companies combined have over 60% of the digital ad market and the digital ad market is already over 40% of the worldwide advertising. Historically, ad spending grows at the same rate as nominal GDP. Recently, ad spending has grown slower than GDP.
What Happens After Google and Facebook Eat the Ad World?
What’s undeniable is this: Omnicom, Facebook, and Google get the majority of their money from advertising. In fact, of the three companies, Omnicom gets the most money from non-advertising (but still corporate communications) activities. In the long-run, ad spending won’t grow faster than nominal GDP. Therefore, Omnicom and Google and Facebook are all eating from the same pie and that pie is not growing faster than the overall economy.
It’s easy not to notice this. As investors, we look at the past record rather than future projections. And we compartmentalize our thinking by company, industry, etc. People aren’t putting Google and Facebook in the same compartment as Omnicom. But, in the end: it’s all just ads. You might think that digital advertising is somehow different from overall advertising or that Google and Facebook are somehow different from digital media outlets in general. But, in the long-run they’re not. And the faster the companies grow now – the shorter the long-run gets.
What I mean is this: Google and Facebook may be better than the overall digital ad market now. But, pretty soon they will be the digital ad market. Likewise, digital advertising might be a better space than advertising overall – but, pretty soon, digital advertising will be overall advertising. Facebook and Google are becoming the worldwide ad market. And the worldwide ad market is mature. So, Facebook and Google are becoming mature.
For example, let’s say that over the next 5 years Google and Facebook increase their share of the digital ad market from about 60% to about 80% and digital advertising increases its share of total advertising from about 40% to about 60%. If that happened: Google and Facebook would now be capturing 48% of worldwide ad spending instead of 24% now. Ad spending grows over time. Let’s be optimistic – relative to its recent growth – and say global ad spending grows 5% a year. That would mean Google and Facebook would be eating double their portion of the overall ad pie in 2023 relative to 2018 and that pie would be 28% bigger (in nominal dollars). So, the increase in revenue for Google and Facebook would be a little over 20% a year for the next 5 years. That’s a good result if you’re buying a stock at a price-to-free-cash-flow of 15. It’s even a good result if you buy at a price-to-free-cash flow of 20. You could make 15% a year that way (because, the price multiple collapse as the stock ended its growth phase would only take about 5% a year off your returns). But, if you pay 30 to 35 times free cash flow – a 20% annual growth rate over the next 5 years gives you a return somewhat worse than 10% a year but better than 5% a year. Generally, you don’t want to make the kind of bet where you need 20% annual growth in the underlying business to drive 8% annual returns in the shares you own.
Again, this assumes the share prices of Google and Facebook eventually collapses to a price-to-free-cash-flow of 15.
They have to.
Let’s go back to the concept of handicapping to explain this. How will the market handicap Facebook and Google’s prospects after another 5 years of 20% annual growth?
Well, we know Omnicom – which eats from the same pie as Google and Facebook – is a close to no-growth company right now (its share of overall ad spending isn’t increasing) and it’s valued at about 9 times free cash flow even when the market knows the company will pay less taxes in the future than it has in the past. Google and Facebook are priced at about 30 times free cash flow. Those prices are more than justified if past growth continue. But, we know it won’t.
Once Google and Facebook account for nearly all of the digital ad spending pie and digital ad spending accounts for nearly all of worldwide ad spending – they would have a future that looks exactly like Omnicom’s future today. The companies would be unable to grow faster than the market. And, of course, any shift to any other kind of advertising would eat away at their existing free cash flow. That’s the situation Omnicom is in now. It can’t outgrow the ad market. And any disruption to the ad market hurts it. Well, if Google and Facebook keep growing much longer they too will reach the point where they can’t grow any faster than the worldwide ad market and any disruption to the ad industry will take earnings from them.
In 2028: No Ads Run Anywhere but on Google and Facebook – What are the Companies Worth?
Let’s look at a hypothetical 10-year future where that’s exactly what happens. Put Omnicom aside. After all, if disintermediation happens on Facebook and Google and the world starts advertising only on Facebook and Google properties – then, there will be very little need for ad agencies. Maybe Omnicom will be a disaster as a stock. But, does that necessarily mean Google and Facebook will be great stocks?
At some point, Google and Facebook will be no growth companies. What if we reach that point in 2028?
Again, we’ll assume a 5% growth rate in worldwide ad spending. This means the world will be – in nominal dollars – spending 63% more on ads in 2028 than it does now. This time, we’ll be assuming that Facebook and Google go from 60% of all digital ads to 100% of all digital ads. And we’ll be assuming that digital ads go from 40% of all ads to 100% of all ads. This assumption literally means that every ad in the world as of 2028 will appear on either a Google or Facebook property. No other advertising will exist.
Over 10 years, Facebook and Google would grow their revenue by a little over 20% a year at which point they’d have eaten all the world’s advertising pie. Absent dividends and buybacks – a big and unfair assumption – Google and Facebook shares would probably return about 12% a year over those 10 years in which they successfully consumed the entire ad world.
Betting on Global Domination
That looks like a terrible bet to me. Honestly, I’d like to make more than 12% a year in stocks. I’ll take a sure 10% over an iffy 15%. But, how sure is the assumption that two companies with 60% market share go to 100% market share and the market they serve goes from 40% of the industry to 100% of the industry all within 10 years.
Obviously, it can’t literally happen. But, there are powerful network effects here. I’ve seen how addictive these media properties are compared to the media properties of old. An outcome close to total domination could be close to inevitable. Maybe Google and Facebook will never get to 100% of digital and digital will never get to 100% of all ads. But, it might actually be likely that Google and Facebook get to 80% of digital ads and digital ads get to 80% of the total ad market.
I want to stress one thing here. I’m not saying Google and Facebook are bad stocks. But, I am saying that a truly long-term investor can’t make anything approaching a fortune just from growth at these companies. You need good capital allocation. To outperform cheaper stocks (like Omnicom), these stocks will eventually have to buy back share and pay dividends. Even if they do that, their stock earnings multiples will fall. A lot of the growth in the underlying business has to first go to offsetting this multiple contraction before you can start profiting from it.
Of course, margins could expand. At Facebook especially this seems a likely outcome. So, Facebook could grow earnings faster than revenue. But, we still end up at the same place. Either: these companies won’t be ultra-fast growers in the years ahead or they will become slow-growth stocks very quickly.
It’s literally impossible for Facebook and Google to grow at 20% a year – or anything like that – beyond 5 or 10 years. Maybe they’ll find other things to do. But, all they’ve done historically is make money off ad spending. And they won’t be able to extract growth from that business after 5 to 10 years from now. Obviously, they could grow much longer if they grew much slower.
On a recent podcast I said, “It doesn’t matter what a company’s worth when you buy it. It only matters what a company’s worth when you sell it.”
That’s a strange thing for a value investor to say. But, it’s true. You can buy Facebook and Google as growth stocks. But, if you’re a long-term investor, you can’t sell them as growth stocks. By the time you sell these stocks, they’ll be done growing.
The other issue, of course, is that once dominant these two companies will be on the wrong side of any future disruption in the ad industry. If you have anywhere from 25% to 100% of the world’s advertising pie – you’re going to find it very hard to be the one who isn’t losing 25% to 100% of revenue to whatever new thing comes along.
Would I buy Google and Facebook today?
No. The stocks are too big and too expensive to offer good odds. If they were this expensive, but much smaller – they might be good bets. And if they were this size but much cheaper – they might make good bets. But, big and expensive is bad in the stock market.
What about Omnicom?
There’s a real risk of disruption. The stock is definitely cheap versus other public companies. It’s even now a tiny bit cheap (especially when you factor in lower future taxes) compared to where it has traded historically. I bought Omnicom in early 2009. It’s 9 years later and the future looks less certain to me than when I bought it back then. And the stock is actually more expensive now than it was then. I’ve said before that I’d definitely consider Omnicom below $65 a share. As I write this, the stock’s at $75 a share. I have to warn people reading this though – even at $65 a share, Omnicom isn’t as good a bet as when I bought it at $27 in 2009. It looks relatively attractive because most other stocks are so unattractive today.
How about handicapping though?
Let’s look at the odds you’re being given on Omnicom. The stock has a 3% dividend yield. It can lower share count by 2% a year. And then an expansion in the price-to-free-cash-flow ratio from 9 when you buy it to say 15 when you sell it could – if it happens over 5 years – would make you another 10% a year. In other words: it’s possible – I won’t say likely, but I’ll definitely say it’s a real possibility – you could make 15% a year in this stock even if organic revenue growth is awfully close to 0% a year.
The hurdle the underlying business of Omnicom has to clear to get me a 10% to 15% a year return while I own it is much, much lower than the hurdles Facebook and Google have to clear to get me the same 10% to 15% a year return. Facebook and Google are ultra-wide moat businesses. But, at today’s prices, I wouldn’t choose those stocks over Omnicom.
You can reach Geoff by emailing him: firstname.lastname@example.org, following him on Twitter: @GeoffGannon, or listening to his podcast. His stock specific write-ups appear on a subscriber supported website: Focused Compounding.