Thread by @TurnerNovak

$SNAP Q4 earnings today. Bad DAU numbers will tank the stock, but Snapchat has a core, sticky userbase and is being under-monetized.

Here's a thread on the Facebook Power Law and a look at how Snap could do up to $10 billion in revenue in the US alone in its current form.

Lots of going on at Snap (see link), but as someone who appreciates product and design, I've learned a lot following their product strategy (ie: creative new revenue generation in consumer social). Lots of lessons from what Snap has done right.

marketwatch.com/story/snap-ear…

To summarize my prior Snap earnings threads (below):

Snaps per Day is Snap's key metric, the US and EU are what matters for monetization, its ad prices are insanely low, and the market is not appreciating Bitmoji, Snap's AR platform, or the Snap Map.

https://twitter.com/TurnerNovak/status/1055543608894504960

For some context on why the US is so important for Snap, it helps to dig into Facebook's numbers.

FB is a living, breathing Power Law. Since its IPO, FB's userbase in North America and Europe has barely grown

… FB’s North American and European userbase is currently 30% of the total…

And FB’s North American and European userbase has always generated ~75% of its revenue.

Facebook had 2018 Average Revenue per User (ARPU) in US/Canada and Europe of $146 and $49, respectively, whereas Asia Pacific and Rest of World were $17 and $12 - even lower if you back out high GDP per capita markets like Australia, Japan, South Korea, etc.

US and Europe share of Facebook’s ARPU was nearly 90% in 2018

Lastly, Facebook’s ARPU growth in the US and Europe (green lines) are not only larger on an absolute basis, but have also always grown faster than everything else on a relative basis.

This is muddled because Facebook only reports "number of core Facebook users" and "aggregate company-wide revenue" in each market. Doesn't consider Instagram, Messenger, or WhatsApp; but revenue from the messengers is immaterial and Instagram monetization likely matches core-FB.

In 2018, FB had total cash outlay (P&L expenses + CapEx) of $30 per FB DAU, and $23 per "unique family of apps" DAU. P&L costs alone were $20 and $16 per DAU.

Considering a cash outlay of $16-30 per DAU, and ARPU’s of:

North America: $146
Europe: $49
Asia Pacific: $17
Rest of World: $12

All FB's free cash flow comes from developed markets, and it likely loses money everywhere else.

(Worth noting Facebook has disclosed a total of 116M fake and 255M duplicate accounts. No hints at where these are located, but heavy concentration either in/out of developed markets could further prove or disprove my point.)

nytimes.com/2019/01/30/tec…

Ben Thompson explains Facebook's business model very well. There's near zero cost to place another ad in a feed (or story stream). The more ad impressions, and the more it charges per ad, the more money it makes. Especially true in developed markets.

stratechery.com/2017/the-super…

As long as FB can keep a) creating new ad impressions and b) increasing prices (CPM's) in high income per capita markets, it will keep printing money at insane margins (potential issues with content licensing costs and regulations aside).

One of my biggest takeaways from FB's Q4 earnings call last week is that Zuck flat-out stated that Stories as an ad unit is dramatically under priced (which I've been preaching forever). As in "early days mobile ads" under priced.

https://twitter.com/bakergavin/status/1091008899086348288

The key with Stories, and short-form vertical content in particular, is it's a mobile video product structurally designed for ads. Whether its back-to-back pieces of unrelated 10 second videos, or 3-5 minutes of premium content, it lends very well to frequent, natural ad breaks.

High ROAS on Stories and high LTV on US/EU users is not only great for Facebook, Snap, TikTok, etc; but also any startup looking for lower CAC. Stories ads are significantly under priced, and offer a new acquisition channel for savvy marketers looking to put in the early work.

Based off reported DAU and MAU numbers, Snap's reach is about ~42% of core Facebook's in North America and 20% in Europe. Snap's growth has no doubt slowed, but Facebook's reach has barely grown in nearly a decade as it simply monetized exiting users in developed markets.

Easy to lose sight with FB posting its “2.5 billion users” numbers, but this reach comparison of <34 year olds in the US shows Snap is relatively even with IG and FB.

Here's a look at the reach in Snap's ad manager as of Jan 2019 - not DAU's, but anyone eligible to view an ad. Could argue it's roughly MAU's. Shows a big drop in Q4, but the footnote for both Oct and Jan says the data is extrapolated. Concerning, and by far Snap's biggest issue.

Snap's Android app is historically much worse than iOS, and is likely a reason Facebook was able to do so much damage. A redesign of Snap's Android app has been in the works for over a year, and will likely be a focus of the Q4 call as it's been expected to reverse user losses.

Just like FB is transitioning to "family of users" numbers, I won't be surprised if Snap does too. Snapchat's been ~flat over the past year, but Zenly appears to be growing in different markets, and last February's redesign accelerated Bitmoji adoption (now at ~134 million users)

As the reach of all these platforms plateau, investors and advertisers will pay more attention to cart conversion and ROI than just eyeballs. Reach will always be important, but AR-enabled "try before you buy", in-ad shopping, and location/visit attribution will be important.

Seems evident Snap's CPM's are too low. Users skew younger so you'd expect cheaper, but Snap Ads have very high engagement due to the nature of the app. Snap just needs to continue monetizing existing user behaviors and building out attribution.

https://twitter.com/TurnerNovak/status/1048790271520071680

While Facebook cloned Snapchat, Snap cloned FB's ad platform. Snap's is a few years behind FB, but its whole business model is built from the ground-up for mobile video, and Snap will benefit as vertical video ads are adopted across the industry.

https://twitter.com/herrmanndigital/status/1014169916138590208

Snap recently opened up programmatic, self-serve ads to its premium Shows, which will allow it to scale that product with relatively low marginal costs.

mobilemarketer.com/news/ad-age-sn…

Linear TV, which has no targeting and poor attribution, gives us hints at what CPM's in a mature digital ad industry may settle into. Network TV averaged CPM's of $43 in 2017, with CPM's ranging from $10 to $167 depending on the show and network.

Another takeaway from legacy TV is ad load. Network TV averaged 11 minutes of ads per hour, or 18% of total air time.

Based on an avg 4 minute show with three, unskippable six second ads, Snap has 4.5 minutes of ads per hour, or 8% of total air time.

Legacy TV metrics hint Snap could see Snap Show CPM's north of $20, with 2x the ad load. TV ads are about key pop culture moments, and Snap Shows are obviously not directly comparable, but they’re a new medium that matches new consumer mobile consumption behaviors.

Snap has stated the amount of time users spend watching Snap Shows has tripled since the beginning of 2018. The number of shows that reach a monthly audience of over 10 million has increased by the approximately same amount:

Q1: 7
Q2: 11
Q3: 21

Snap Shows are cheap ($35k-50k per episode) and quick to make. Heres a few scenarios w/ publisher margins at the high-end, paying full production costs under different CPM and view assumptions. Numbers get more attractive w/ more views and as overhead spreads over multiple series

Just like Netflix and Amazon push users to specific content upon launch, Snap can, and is, doing the same with Discover. All three can push views of whatever drives the optimal business results of each platform: Views, time spent, retention over time, etc.

Of course, Snap's userbase is the big unknown. Key to note 2018 wasn't its first user contraction - MAU's dropped 20% in the summer of 2014.

TV has also relied on increasing CPM's and ad load as reach declined the past decade, Could Snap do the same?

gigaom.com/2015/01/09/has…

Premium content, among other things like the UGC AR filters, social currency in Streaks, and a secondary camera roll in Memories, keeps users coming back to the app. Searching Twitter for “Snapchat only reason deleted” usually supports this:

twitter.com/search?q=snapc…

I always focus on "Snaps Created per Day" because a) it defines camera ad inventory, and b) a snap is a message (pulls in users with a notification) and a story (creates ad inventory).

It's the key measure of Snap's network effect / competitive advantage

https://twitter.com/TurnerNovak/status/1048790251601313793

AR content and camera ads are, in my opinion, a new medium akin to the newsfeed circa 2012, an immature format that is quickly gaining adoption. The usage is similar: users scroll vertically to interact and play with lenses (ads) as they did the newsfeed a decade ago.

I tweet this too often, but the camera represents a new medium for things like AR storytelling and commerce. It really only works though with a critical mass of users checking it consistently (aka Snaps per Day). Here's an example of storytelling in AR:

https://twitter.com/that_so_lyriee/status/1057130925358755840

Another way to measure Snapchat usage is to look at Snap's hosting costs. They were expected to fall over time with better pricing from AWS and GCP (bloomberg.com/news/articles/…), and they have, but they're flat since the redesign - when usage supposedly plummeted.

To touch on the Snap Map, it's a unique take on messaging that piggybacks on Snapchat’s social graph. Snap could naturally serve impression-based ads, introduce "map apps", and slowly shift currently un-monetizeable messaging activity to the map.

https://twitter.com/TurnerNovak/status/1028259401826820104

Here's how Snap's US revenue could look in 2019, plus a few future monetization scenarios.

I assume reach shrinks up to 20%, usage based on reports/leaks, CPMs and ad load hit low-end of TV, assume Bitmoji and Snap Map monetize via ads, and I exclude geofilters and Spectacles.

DAUs must stabilize, but Snap has lots of low marginal cost revenue growth ahead as it increases CPM's and impressions. Considering its self-serve AR and mobile video ad platform, at the very least its an interesting acquisition target compared to something like CBS or Viacom.

A few days late, but looking through the Snap Q4 earnings now. A couple highlights that stood out to me...

Snap stopped reporting Snaps per Day, but earnings call Q&A reveals they're still over 3 billion, or 16 Snaps per User per Day.

Snappables, Snap's AR games, were released in 2018, wonder what tractiion they're getting? 80M unique users played Tic-Tac-Toe in Q3, or ~43% of DAU's.

Snap reported 70%+ of DAU's played with or viewed AR lenses daily.

Over 300k community lenses have been created to date in Lens Studio, have received an averaged of 400k views each, for a total of 35 billion community AR lens views in 2018 and 20 billion in Q4 alone.

The Q4 explosion in AR lens views was likely spurred by the Lens Explorer launch mid-2018. Big question is how it evolves; does it simply stay as a way to preview lenses, or does it become something that more resembles an AR shopping mall or AR app store?

theverge.com/2018/7/10/1755…

iOS DAU's were up YoY and QoQ, and iOS time spent grew faster in Q4 '18 than Q4 '17.

Considering that, Android DAU's were likely down. The Android app rebuild is being slowly rolled out, which you could assume returns Snap to aggregate DAU growth when its widely released.

Revenue from premium content was up 2x in Q4 YoY, and users are watching 30% more of it daily YoY.

Snap stated it reaches 70% of 13-34 yr olds in the US w/ mobile video ads per month. NBC's Stay Tuned reaches 25-35 million per month, and 2/3 of those it can't reach elsewhere.

In Q4, over 60% of ESPN's Sports Center audience watched at least 3x per week.

Snap's new Original Show "Dead Girls Detective Agency" reached more than 14 million unique viewers, and 40% of those who watched the first episode watched the entire season (seems low to me??)

This Nov 16th article mentions 10 million viewers, so did Snap drive 4 million additional viewers through a) placement in Discover after it first aired, and/or b) word of mouth spurring search discovery? Indicates Snap’s premium content has a long tail.

pophorror.com/s2-of-the-dead…

Most interesting to me isn't even the in-app performance of any Snap Shows, but what it could mean outside Snapchat. Could eventually license IP, and @basche42 has made good points in the past on how Snap could use them to launch an OTT subscription service (I'm not sold on this)

Snap has grown the # of advertisers every Q since launching its self-serve ad tool in Q2 '17. Important, because more ad diversity increases engagement rates on all ads (don't see same ad over and over again), and improves advertiser ROI, which allows Snap to charge higher CPM's.

This was also the first quarter since Snap IPO'd that BOTH pricing and ad impressions were up. Impressions have driving revenue growth over the past , could we start seeing pricing trend up as well?

Lastly, Snap's contribution margin was 89% in Q4, which means for each $1 in additional revenue compared to Q3, it cost Snap ~11 cents on a product-level (hosting, revenue share, etc). Not surprising, as it was 80%+ all of 2018. Hints how scaleable Snap's business model is.