View profile

Twitter Write-Up

Expand Your Circle Of Competence!
Expand Your Circle Of Competence!
Jack resigns from Twitter. Parag Agrawal will be the new CEO effective immediately.

“we’re sitting on top of the world’s largest depository of human thought… all delivering in real time. And the applications of that are really wide ranging.” - Matt Derella, Twitter Customer Lead
Twitter’s market cap of $37 billion can buy “the world’s largest depository of human thought”.
Since the IPO in 2013, Twitter has been relatively flat trailing the S&P 500. Twitter is worth a look at this price to see if a turnaround can make it a good investment.
Twitter History
To understand Twitter’s history, you have to go back to even before it was founded… to 2004. When Noah Glass came up with the idea to build the next podcast technology, Odeo; and Evan Williams who funded it with the millions he made from selling Blogger to Google.
Odeo’s early employees were anarchist hackers, great at coding but saw work as a temporarily pit stop before being involved in the next protest event.
In the summer of 2005, Steve Jobs announced podcasts were coming to iTunes. The news meant that Odeo would not see the light of day. Majority of the Odeo employees would be let go, while the remaining would pivot to a new idea— Twitter, the future of status update.
The pivot to Twitter took two weeks to build. The hacker mentality of getting spaghetti code to work at any cost compounded technical debt and would plague Twitter for years to come. 
Jack’s First Rodeo
Twitter has had four CEO changes during its first nine years. It’s difficult to build a stable culture in a young company when a game of musical chairs is taking place at the CEO position.
Ev was the main financier of Twitter and it was assumed he would become CEO because he had previous leadership experience at Blogger. Instead, Ev decided to make Jack CEO so that he can focus on his pet project, Obvious Corporation, an incubator for Web start-ups. Ev would be available to Twitter for guidance and vision.
Internal resistance sprang up when the idea Jack would become CEO emerged. Many people believed he wasn’t experienced to run a company, but Ev commented that Jack could learn on the job. Jack would be the first CEO of Twitter and Ev would be chairman. It would be the first time employees look up to Jack as a leader.
Employees describe Jack as quiet and averse to conflict, preferring to work in the background. His strength was envisioning the big ideas.
In his book “Hatching Twitter”, Nick Bilton interviewed previous Twitter employees on Jack’s leadership style.
Being a natural born leader was a skill Jack lacked early on. When sending a draft email setting the company’s goals, each milestone started with “I want to be able to…” or “I want…” or “I…” The draft email “I” was replaced with “we” to address the company.*
Ev recounts a story of Jack lacking attention-to-detail. “Jack, who had been managing expenses on his laptop, had been doing the math incorrectly. When Ev learned about this, he asked a friend and seasoned entrepreneur, Bryan Mason, to meet with Jack and show him how to manage the company’s books, but Bryan spent the entire meeting at a whiteboard with a marker explaining the basics of accounting.”
Employees did not know who to go to to make decisions, Jack or Ev. Jack’s commitment to Twitter was questioned when server outages displaying the “fail whale” were increasingly reoccurring, leaving work early for personal activities, and still not a dollar in revenue.
Ev was becoming more hands-on seeing that Twitter’s problems weren’t going away. He saw that Jack didn’t have what it takes to take Twitter from a start-up to a real business. In 2008, Jack was fired as CEO and moved to a passive chairman role and “silent” board seat to divert outside speculation that inside Twitter was a mess. 
It was Ev’s turn to sit in the CEO chair.
*The first management book Jack read years later was “The Score Takes Care of Itself” by Bill Walsh; who coached the 49ers from the worst franchise in sports to a dynasty that won three Super Bowls.
Ev’s Turn (2008-2010)
Users grew massively from 5 million to 71 million under Eve’s watch. What didn’t grow enough was the 110 employees and management team. Partially due to Ev’s bias of hiring friends and paranoia of hiring talented management that might betray and replace him. 
Ev was slow at making important decisions. In a past interview, he shares, “the mistake I made was definitely under hiring, both in quantity and in experience in several areas, for a long time.”
By 2010, internal gossips of Twitter having no direction casted fears over everyone. Ev felt blindsided when the board fired him. Dick Costelo would become CEO and bring with him his business experience of consulting for 8 years and selling his startup, FeedBurner, to Google.
Dick Costelo (2010-2015)
Costelo brought the first dollar of revenue into Twitter shortly after becoming CEO by signing deals with Google and Microsoft to display tweets. In his five years as CEO, he took Twitter from $0 to $2 billion in revenue.
Many of the core advertising foundation that is still used today were built under Costelo. Twitter was the first company to pioneer ads displayed in the feed. At that time Facebook ads were displayed on the side of the webpage. Ashish Goel, a professor at Stanford, came up with the brilliant idea of charging advertisers using tweet engagements. Standard measurements like cost-per-impression and cost-per-engagement would be converted to Twitter actions through likes, retweets, and replies.
In an interview podcast, Dick Costelo reflects back on Twitter after five years that’s worth a listen. (LINK)
Twitter IPO
2013 Twitter IPO
2013 Twitter IPO
Twitter IPOed on November 7, 2013 with an initial public offering at $26 per share. By the end of the year, it reached $75.
“A multiple on $75 to what we were doing in revenue, even the EBITDA at the time was bananas.” - Dick Costelo, former CEO at Twitter
Investors were keen on Twitter’s slowing revenue and user growth. The hype of becoming the next Facebook crashed into reality as did the stock price. Down over 60% from a year ago.
In 2015, Chris Sacca, an early Twitter investor which at one point owned 18%, made public a 8,500 word essay on What Twitter Can Be. The core message has aged well and worth a read. Other parts did not, like Periscope will be the most important acquisition in Twitter’s history. (Periscope was shut down in March 2021 due to usage decline.*)
8 days after Sacca’s essay circulated the internet, Mr. Costelo resigned.
*It’s interesting that Kayvon Beykpour, co-founder of Periscope, decided to remain at Twitter as Product Lead after Periscope was shut down. Usually emotions run high and entrepreneurs would move on to the next big thing. Continuity in the product lead position is important as there has been 6 product lead changes since 2014.
Jack is Back (2015-2021)
After seven years, Jack returned to Twitter as interim CEO . During that time he co-founded Square, a payment company helping SMBs do business. The board was impressed with Jack’s execution and needed a founder’s vision to get Twitter back on track. 
After three months, the board decided to make Jack the permanent CEO. He would split his time between Square and Twitter which has received criticism over the years.  
The board looked for 4 key criteria in the selection process:
  1. A world class leader with a track record of success in the consumer internet technology and media sectors.
  2. A person able to articulate a clear direction for the company’s future and a plan for how to get there, one that can both guide the company and inspire employees, users, investors and the public at large.
  3. An executive with a demonstrated record of execution and thoughtful risk taking, who can get the most out of product and engineering teams while prudently managing the company’s financial resources.
  4. We sought a leader who understands the importance of company culture and knows how to foster it.
In what would be a rare occurrence for a founder to do, Jack gave up one-third of his Twitter holdings to the benefit of current and future employees. He did the same also at Square previously.
The period after Jack’s return can be split into four phases: Direction, Health, AdServer Rebuild, and Product Innovation.
Direction (2015-2017)
Twitter’s top of funnel has remained robust over the years. Top of funnel describes millions of people that interact with Twitter who are new or who have not logged in for 30 days or more.
2016: “when we look at the top of our funnel every day, we’ve said on the calls in the past, it was millions of people that are new or resurrected, either not currently user in the last 30 days that comes to the top of our funnel every day and that’s a massive number. That translates into over 400 million on a quarterly basis and 700 million on an annual basis of that translate into actual users. If we saw a decline on the top of our funnel that was significant, your concern would start to concern us as well, as well as other factors, but we have not seen that. And we think it’s a competitive advantage and it actually reinforces that top of the funnel dynamic that we benefit from and that still creates such a unique opportunity for us.” - Anthony Noto, Q4 2016 earnings
2021: “Our top of funnel remains extremely strong with more than 2M people coming to Twitter every single day that are people who are either new to the service or who have not logged in for 30 days or more.” - Kayvon Beykpour, Analyst Day 2021
People are attracted to Twitter because it’s the fastest way to see what’s happening and what everyone is talking about around the world.
Jack first task as CEO of Twitter was to give direction and prioritize the product road map. The leadership team flattened the organization to elevate engineering, product, and design to report to Jack.
Twitter trimmed their fat with two significant lay offs to focus on its core strength. That meant gradually taking away resources from products like Fabric, Vines, and Periscope until they were either shutdown or sold.
“So we’ve been reviewing our road maps to make sure that we are making Twitter easier for everyone around the world. Currently, the product makes people do a lot of work to realize the value, and there are a lot of initiatives aimed at making sure that people can immediately get value out of Twitter and get a lot of power from it as well.” - Jack, 2015 CEO transition
Health (2018-2019)
The health of the public conversation was deteriorating from harassment, abuse, and misinformation. In 2018, Twitter publicly announced they were removing fake bots and prioritizing health over product improvements. The health initiative would impact user growth near term but long-term would create a more sustainable social platform.
Technology like machine learning and AI can help lift the burden from users self-reporting, but improvement on health initiatives will always be on-going effort.
Twitter has done more than other social media companies on health by allowing users to revert their feed to chronological timeline. This allows you to see tweets only for the accounts you follow. They were the first to stop accepting money from political ads and the first to label misinformation tweets.
In an interview with Kavyon, Twitter Products Lead, he explains that Twitter is rethinking the fundamental incentives on the service.
“Things going viral on Twitter can be a really awesome thing. But oftentimes, unhealthy content can get viral more easily precisely because of those mechanics. So one of the things we’ve been thinking about is whether we have the right balance of incentives within the core product experience. Putting our rules aside for a moment, just as an example, there isn’t really a disincentive today to being a total jerk on Twitter. And that’s a product problem. Like that’s something that we should think about in terms of what are the right metrics to use to tip the scales a little bit and create a disincentive for people to behave in a way that we might consider abusive.” (LINK)
AdServer Rebuild (2019-2020)
More on the AdServer rebuild in section “Greatest Risk is Behind Twitter”.
Product Innovation (2021 and on)
The openness and cadence of product developments has accelerated in the last year. Some of the products I am excited for are:
Communities - Twitter has been taking cues from subreddits at Reddit as they are testing communities.
Communities will be great for all Twitter users who want to engage deeper into a niche community. It might be intimidating to tweet for the whole world to see and less so for a smaller community. For example, asking about how to filter honey inside the “Bee Keeping Community” and getting a reply will be magical for a less followed account.
Humans natural tendency to belong in groups to survive, share, and learn from has remained in our DNA since the beginning of time. Communities is just another way to easily find your “group”.
Spaces - A different form of casual podcasting with audience participation that includes all the uhms and uhhs. I’m excited to see what new use cases the world can create with Spaces.
Twitter Blue - Pay $2.99 USD to get premium features like undo tweet, categorized bookmarks, easier experience reading tweet threads, most-shared top articles, custom navigation, etc…
All these features are great for the power user but will probably not have enough value for the casual user. $3 a month is almost half a Disney+ subscription.
If 1% of Twitter’s 211m mDAUs are power-users and subscribe to TwitterBlue annually, that will bring about $75m in revenue. 10% about $750m.
I’m less optimistic on how much Twitter Blue can bring to the bottom line. From Tony Haile’s tweet, it seems all of the $3 will go to news publishers to fund an ad-free experience. With the goal “to earn publishers 50% more revenue per person than they would make from serving ads.” (total adds up to $3.51. Will Twitter be paying the extra .51 cents?)
Tony Haile
Twitter Blue launches today in the US with Ad-Free Articles. You can now get an ad-free experience on the WaPo, LA Times, Rolling Stone, The Atlantic and hundreds more. Your subscription directly funds the journalism you read.
Thoughts about future of Twitter
“…the best way to measure our success is mDAU” - Kayvon Bykpour, Product Lead
Twitter’s ambitious goals by end of 2023:
  1. Doubling the number of features per employee that directly drive either mDAU or revenue
  2. At least 315 million monetizable daily active users (mDAU)
  3. At least $7.5 billion annual revenue
Twitter has grown mDAUs at low teens percentage consistently with a one-time boost from COVID. There needs to be an acceleration of growth in the high teens if Twitter has a chance of reaching their goal of 315m mDAUs by end of 2023.
Revenue grew from $2.2B in 2015 to $3.7B in 2020 at a 11% CAGR. In Q3 2021, Twitter reported 211 mDAUs and $1.26 billion in revenue (13% and 37% y-o-y increase respectively). Video formats have been more than 60% of advertising revenue for several years. 
After a full year of the completion of the AdServer rebuild and minimal impact from IDFA, revenue continues to grow.
Management believes that long-term, Twitter’s business can achieve 40-45% EBITDA (excludes traffic acquisition costs) and 15% GAAP operating margins. It has achieved those metrics on a few quarterly occasions but never for the full year results.
In 2020, ad revenue was 85% brands and 15% direct response with SMB making up less than 15% of ad revenue. MAP revenue grew 50% y-o-y to $95 million in Q4 2020 and made up 8% of advertising revenue. The goal would be to achieve a healthy mix of 50/50 brand and DR mix like SnapChat.
Twitter’s strength is attracting advertisers to launch products and increase brand awareness. In the past, when a user with high purchase intent sees an ad on Twitter, the transaction would occur off the platform. Shop Modules solves this problem. In the near future, users can be purchasing products on Twitter.
Twitter Shop Module
Twitter Shop Module
A big lever that remains to be pulled… would be utilizing users’ direct signals expressed following Topics/Interests. This would be a leap in DR ad monetization once turned on.
Rich Greenfield at LightShed Partners asked if Twitter is using any of the signals from users following Topics and where they are in the process. Jack responds:
“I think one of our biggest priorities in the company right now is personalization. So we’re going to put a premium on finding all the right signals to make sure that you’re not just seeing more relevant ads but you’re seeing more relevant Tweets as well. These have a very similar systems, so I don’t think it’s all that far off and then we can start using more and more of these signals to increase the relevance of what we show. But this is the greatest opportunity for us in terms of relevance and that drives everything from growth and usage but also to our advertising business.”
Advertising TAM
“The global advertising market at the end of 2020 was $490 billion dollars. Digital advertising, where we participate, accounts for roughly half of this market, or $280 billion dollars. And this is the ad category expected to grow the most in 2021. Within the digital ad market, the brand and direct response, excluding search, represents roughly $150 billion dollars of opportunity in 2020. And we believe DR, which we also call Performance Advertising, is larger and growing faster than brand, especially considering the acceleration toward e-commerce we’ve seen in the past year.” - Matt Derella, Twitter Customer Lead
Twitter makes up a very small share of the global digital advertising total addressable market. The main thesis of increasing advertising TAM is to increase mDAUs to attract more advertisers. To increase mDAUs requires product improvements. Twitter Communities is a product that should increase mDAUs meaningfully.
Greatest risk is behind Twitter
The greatest risk Twitter faced was when management decided to rebuild the ad server infrastructure in 2019. The stakes were high— $3 billion in revenue. If the rebuild was done poorly, you could have a worse performing system than before.
“A great advantage of software systems is that they are extremely adaptable. There comes a point in the evolution of complex software systems, however, where that malleability hinders growth instead of facilitating it. At some point, software will approach a stage where it’s no longer serving its purpose — helping people. 
This was the case for the Twitter AdServer in early 2019. After 10 years of iterative development, the system was too inefficient to further evolve with the organization. When it started, we were an extremely small team of engineers, serving a single type of ad format (Promoted Tweets), generating around $ 28M of revenue. Today, Twitter’s Revenue organization consists of 10X more engineers and ~$3B of revenue, supporting multiple ad formats - Brand, Video, Cards. 
Low velocity for launching new products and tightly coupled cross team dependencies with high overhead costs added to the growing complexity of the organization. This demanded a fundamental change in order for us to scale any further.” - Twitter Blog (LINK)
Promoted Trends was a past example showing the difficulty of introducing new ad products. As a parameter, it relies on country location rather than the normal Candidate Ranking step. In the old monolithic ad server, Candidate Ranking stage could not be be skipped. Hacky workarounds had to bypass this stage in order to get Promoted Trends to work.
An accumulation of technical debt increased, snowballing the problem for future employees.  
Old AdServer infrastructure
Old AdServer infrastructure
Releasing new ad products was stressful. New and even experienced engineers required an understanding of the whole ad pipeline causing hiring and retention of employees to be difficult.
In order to scale, Twitter desperately embarked in an ad infrastructure rebuild in 2019. The solution would be to break up the monolithic legacy infrastructure and separate the ad products into microservices. The huge risk was if redesigned poorly, would lead to reduced productivity than before.
New AdServer infrastructure
New AdServer infrastructure
The new AdServer benefit was increased ad development velocity that could scale. New engineers are not required to understand the whole ad pipeline allowing faster contribution when hired.
“As we continue to grow our ad business, adding a new product will be as easy as plugging and playing these horizontal platform services as needed.”
Tradeoffs were needed for the new ad infrastructure:
  1. increased hardware cost: double the cost to compute but within 5% of the ad revenue generated
  2. required to grow product development teams
  3. engineers were taken off product development to support building the new ad infrastructure
  4. increased complexity in ranking and auction dynamics 
By mid 2020, the ad server rebuild was complete. During the year and a half rebuild, the monolothic AdServer continuously ran to support Twitter’s $3B ad business. Traffic was slowly funneled into the new ad server and compared to the old AdServer.
The results?
  • No single team is the bottleneck for most planned projects, and more than 90% of the projects planned in ads serving can be executed.
  • Experimentation is much faster - online ranking experiments in the ads serving space are now being run 50% quicker.
Risking all of Twitters ad revenue to rebuild into a horizontal ad platform that scales was risky but required for growth.
“the milestone that we passed of rebuilding our ad server was pretty critical for us. We’re really happy to put that work behind us. As we’ve been talking about, rebuilding that ad server allows us specifically to experiment a lot more and, thus, increase our development, obviously, because we can launch new products, get feedback from advertisers, evolve them very quickly, refine them and then get them to market. So everything has really been dependent upon that, including our experiments with direct response, starting with MAP.” - Jack Dorsey, Q2 2020 earnings
Twitter's weakness is its moat
There is no starter-kit for the interest graph on Twitter. New users must slowly build up their interest graph by sending signals through likes, retweets, following accounts, and Topics. Conversely, Facebook’s interest graph is your friends in real life that are stored on your contacts list. It’s easy to import and gain immediate value.
The interest graph dynamic causes Twitter to scale slowly but be more resilient. This creates a network effect that is difficult to replicate and serves as Twitter’s moat. Many companies have tried to create a product like Twitter but didn’t gain enough traction and failed.
On-boarding process is still an issue management has been trying to solve since Twitter was founded. In 2014, Instant Timeline was introduced to new users where they are prompted to choose high-level topics to tell Twitter what they are interested in and populates their feed. (Topics is a similar product) Because Twitter is about interests, it takes longer to understand the user. That’s a problem. If Twitter started out with a well defined interest graph of the user from the beginning then it can bring value to new users minutes after downloading the app.  
The solution?
I’m not sure it can be solved. If Youtube, Instagram, and TikTok were open to sharing their first party interest graphs, maybe. But they will never give up valuable user data to help a competitor.
The interest graph dilemma Twitter faces is the same for all social media platforms that rely on interests rather than contacts.
A positive is that Twitter has yet to use the signals from Topics. A negative is that Twitter has yet to use the signals from Topics.
Jack resigns from Twitter
Jack’s farewell letter to Twitter employees left a lot to be interpreted.
“Founder-led” companies are often worshipped to a fault. Costco led by Jim Senegal, Amazon by Jeff Bezos, Walmart by Sam Walton— all led by founders who built a great culture that will survive well after they are gone.
Twitter did not reach the pinnacle of “founder-led” culture. It fell short. The past founders and first employees with a hacker mindset, CEO betrayals, and stability of the company made it impossible to build a great culture. I think that’s why Jack wrote, “I’ve worked hard to ensure this company can break away from its founding and founders.”
This will be the first “smooth” CEO transition without anyone being backstabbed.
By May, Jack will be fully separated from Twitter, riding off into the sunset, never having to deal with Twitter’s problems ever again.
Will Twitter still be here in a decade?
Humans curiosity to learn about whats happening around them will always remain. News from the newspapers are a day late, radio and t.v. are hours behind, Twitter is instantaneous— it can’t get faster than that.
In ten years, we might be tweeting from inside the Metaverse. But whatever the new medium might be, I’m confident the future management will make the transition like the past; from SMS message to desktop to smart phone.
Circle of Competence
I would like to thank you for making it to the end in 140 characters or less… and I just did.

Thank you.
Don’t miss out on the other issues by Expand Your Circle Of Competence!
Did you enjoy this issue? Yes No
Expand Your Circle Of Competence!
Expand Your Circle Of Competence! @OofCompetence

Circle of Competence newsletter expands investors' knowledge on great companies.

I spend over a hundred hours reading books, annual reports, and earnings transcripts to create one company profile per month.
Top 3 most viewed posts:
1) The Teachings of Nick Sleep
2) Baidu, Buy One Get Two Free
3) Will Alibaba be a $2 trillion dollar company by 2030?

You can manage your subscription here.
In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Created with Revue by Twitter.