Twitter's Levers: Driving Engagement and mDAU Growth

I started a position on Twitter (TWTR) a few weeks ago and took a hit from last week’s blood bath. I added a little more. In spite of the “bug” of misusing user data leading to over-earning, I believe the overall theme of higher engagement still holds.

As I wrote in my last post, Twitter has an “irreplaceable position in our culture and that should command a higher floor multiple than usual.”

Twitter is mostly used as 1) online discussion board, 2) source of near real time news. As an investor I’m intimately familiar with the former use case.

Twitter as an online discussion board provides unparalleled access. Where else can you debate with, and learn from the Nobel Prize winners like Paul Krugman? (yes, I know it’s easy to hate on him nowadays, but still…) Many of these thought leaders will patiently reply to your questions. Even without participating in discussions, just following the debates on Twitter can be more educational than years of reading textbooks.

Twitter is also the irreplaceable source of real time news. The best example I can think of is when Turkey was going through its failed coup a few years ago. Instead of waiting for mainstream media, people went to Twitter. From our computers and cell phones, we saw Turkish citizens flooding the street; we saw civilians unarming soldiers in tanks; we saw the momentum turning in favor of Erdogan and the ultimate coup failure. It was all live on Twitter.

So we have an incredibly differentiated asset here. What is something like that worth? At $30/share, the stock is still trading at ~17x LTM management adjusted EBITDA (big stock based comp add back here). It’s not cheap for sure. But there is a floor and at any given time, operating momentum can turn and the stock could go on a run. All I can do is hold and wait.

Where are We on Value Drivers?

Twitter makes money from advertising. The main financial drivers are:

• Number of advertising engagements

• Price per ad engagement




As can be seen above, ad engagements growth has decelerated in recent quarters while ad prices have been less negative over time. This is price elasticity in action. The market took its time to find the right price for Twitter’s ads.

It’s worth noting that much of the gains in ad engagements come from gains in mDAU (monetizable Daily Active Users). To see this, we further break down ad engagements into 1) number of users (mDAU), and 2) ad engagement per user



As seen above, mDAU growth reaccelerated in recent quarters. On the other hand, growth in ad engagement per user came down to mid-single digits yoy. It’s not clear why this is the case. I think part of it is tough comps. Another reason could be new users have not turned into “power users” yet. I believe product upgrades can re-accelerate this metric.

Which brings me to Twitter’s levers.


Upside Levers - Engagement Growth

Twitter is accelerating its pace of product development.

The company really cleaned up its user interface the past few months. Visible improvements include: lists are much easier to create, organize, and use; searches are easier to access; bookmarks are now in the main menu. The whole app is just better organized in a common sense sort of way.

These are low hanging fruits – fixes that’s should have been done long time ago. There was no reason why accessing lists should require 3 clicks instead of one.

The point though, is Twitter is finally waking up and getting things done. Combined with its "health" initiatives (reducing toxic tweets), these changes reduce friction and make it easier for users to find the content they want. This can lead to better engagements per user.

The upcoming interest based list is an example of reducing friction. We might follow specific users for their stock analysis, but it gets annoying when that person spend 20% of his time talking about his favorite politician. A list based on specific interest/topics (just stock analysis in this case) would side step that issue. This is easier said than done. How will Twitter curate only quality tweets? We’ll have to see how Twitter executes.

Upside Levers - User Growth

Twitter is also trying to increase mDAU and drive TAM expansion with new products. The most important ones are 1) sports events/highlights, and 2) live videos.

First, Twitter is now regularly showing sports events at the top of the app. Clicking into these you can see sports highlights and the latest commentaries. This is a proprietary format if you think about it. Sports highlights are not unique. Comments are not unique. But near real time sports highlights, with humorous takes from like-minded people – that is an unique format.

Second, and somewhat related, is live videos. After failing to integrate Periscope for years, Twitter is now actively showing live videos at the top of the application. Since live videos are at the top of the app, it is essentially guaranteed views, and will incentivize more creators to make more live videos. More supply will in turn draw more demand. Presidential candidate Andrew Yang took advantage of this a few weeks ago by broadcasting a live Q&A session.

Sports highlights and videos represent different use cases from old Twitter. As Eugene Wei noted in his classic essay Invisible Asymptotes, the traditional power user of Twitter are infovores who thrive on the texts. However, this group is a small subset of the overall population and represent limited TAM. I think video products can break that limitation by appealing to a different type of audience.

The levers discussed above are opportunities just within the Twitter application itself. Twitter can also let third party developers build on top of its one-way follow infrastructure and take a cut of revenues.

There are so many possibility for Twitter that I believe a paid subscription model should be the last resort.

I can see Twitter increasing its revenue mid to high teens CAGR for the next 5 years. A reasonable base case would be something like 5%-10% growth in mDAU, 5%-10% growth in ad engagement per user, and flat ad pricing. If TWTR just shows a little operating leverage (this is a software company after all!), it’s not hard for the stock to double.


Notes: 
1. Vergecast did a nice podcast with TWTR’s product head recently and talked about product opportunities. You can find it here. 

Journal For Week Ending 10/12/2019: Bought SMIT, Passed on UBER (and Others)

In the past 2 weeks, I continued to accumulate the Trade Desk (TTD) (thesis as stated in my last write up here). I also jumped on Schmitt Industries (SMIT). There's not much to say about this company. It’s a simple cigar-butt play as summarized in my tweet here:

wow. $SMIT just sold its main biz.

Waiting to see use of proceeds.

market cap <$12mm. What's left of the company?

Measurement segment, 3 buildings in Portland (40.5k sqft), big NOLs, ~$12mm of cash (including sales proceeds), and a CEO who really knows how to invest https://t.co/0PKB2tTBVA

— mspacey (@mspacey4415) October 10, 2019

So basically I’m buying a $12mm market company with $12mm in cash. I essentially get a bunch of stuff for free: buildings, NOLs, and measurement business. It's not that hard to imagine the whole thing could be worth more than $20mm, or ~$5 per share (Stock is trading around $3 at the time of this writing).

I also took smallish positions in TWTR and ROKU.

Twitter is something I personally find incredibly useful for learning and research. The stock is not cheap but the company has been making user interface improvements that I believe can increase user base as well as engagement levels. Twitter has a irreplaceable position in our culture and that should command a higher floor multiple than usual.

Roku is more of an opportunistic trade. I have been researching the whole connected TV space (along with Trade Desk) and thought ROKU was a good company, but held off from buying because it’s expensive. So when the stock price came down and filled some technical gap I took a gamble on it.


Some Stocks that I Passed On

I passed on a bunch of stuff the past 2 weeks. The most prominent ones are Uber, Chewy, and Spotify. They are all tempting buys as they are entrenched businesses with strong brands. All these stocks looked to be bottoming too.

Here I will comment on why I passed on Uber. Perhaps in a later write up I’ll share the same about Chewy and Spotify as well.

Uber

I am mostly concerned about the sensitivity of margins to prices (that consumers pay).

In 2Q19, management mentioned loss of $100mm for Ride Share. We can back out the fixed cost and see that this implies 20% contribution margin for core Ride Share business.

The thing is, this 20% is off their 20% take rate (as in Uber gets 20% of what riders pay). As a percentage of consumer spend it’s really 4% contribution margin.

Yes you can make up for low margins with volumes and still have great profit dollars. That’s not the problem. The problem is sensitivities - any change in pricing (or take rate) could wipe out the economics.

So far I’m talking about current data, but the real question is this: What does fully ramped economics look like?

We know this 20% contribution margin in Ride is a mix between a) fully ramped, very profitable cities like San Francisco, and b) not yet ramped loss making cities. What can we infer about fully ramped margins? Let’s plug in some numbers for illustration purposes.

• Let’s say 60% of tickets come from fully ramped big cities that are profitable; assume loss making cities have negative 20% contribution margin.
• So you do 60% x + 40% (-0.2) = 0.2
• Doing the math, this would work out to about 35% contribution for fully ramped cities.

Again, this 35% is off Uber’s 20% take rate, so expressed as a percentage of what customers pay, you’re talking about maybe 7% contribution margin, even when fully ramped.

After deducting fixed cost, you probably have EBITDA margin in the low single digits. Again, a little drop in customer pricing or take rate would wipe out the economics.

When I started my research I imagined Uber to be this high margin money making machine with software economics. That’s just not the case.