Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.
Despite the media chatter, I am still bullish on Twitter – as should be any investor who understands the social network’s fundamentals and true potential. Twitter has the highest revenue growth rate of any tech firm with over $2 billion in sales over the last year. And at today’s market cap, Twitter is an incredible bargain.
The company has enormous untapped potential to impact the world and create value for investors and partners — far more than short-term investors probably realize. But to unlock that hidden potential, some significant product and business model evolution may also be necessary.
I truly want the Twitter ecosystem to succeed. And it is in that spirit of support and optimism that I’m offering a number of ideas below that could help Twitter not only regain its former growth curve but surpass it. I’m breaking down my detailed playbook for the company into three sections:
Let’s jump straight to part 1.
1: Improving the signal-to-noise ratio on tweets
One of the primary challenges to using Twitter today is finding content you want and getting attention for your own content amid all the noise. I believe the changing ratio of reward to effort in Twitter is one of the underlying reasons engagement rates are not what they once were. Another way of thinking about this is that the signal-to-noise ratio in Twitter needs to be improved. There are several ways to fix this:
Smart personalization and filtering
The first thing Twitter needs to do is give users better ways to filter their timelines so they can improve the relevancy of what they see.
To do this, add machine-learning based personalization to the Twitter user-experience so that users can teach Twitter what they want to see (and what they don’t want to see).
How to do it:
- Twitter should add Netflix-style thumbs up (“more like this”) and thumbs down (“less like this”) feedback actions on each Tweet.
- Based on this feedback, Twitter can learn machine learning to learn and adapt to each user’s changing interests and prioritize content for them.
- In addition, by analyzing what each user has tweeted, retweeted, liked, or taken action on in the past, as well as who they follow and who follows them and takes action on their tweets, the system can learn even more deeply about changing user interests and priorities.
By personalizing the user experience, Twitter can vastly improve the signal-to-noise ratio for each user. This will help to restore the reward of participation by giving users more relevant and timely content when they use the app. This in turn will yield greater response rates to tweets, which not only rewards other users for tweeting but ultimately also rewards advertisers for their tweets as well.
A new metric for influence
Another way to filter and improve Twitter’s signal-to-noise ratio is to provide a better way to measure the value of a user.
Follower counts made sense as a filter in the early days of Twitter, but with the rise of fake Twitter accounts, Twitter followers for sale, and Twitter bots that generate followers, nobody trusts follower counts anymore.
Verified users and featured users is another approach, but that tends only to apply to celebrities or major brands. What about the 99% of the rest of the users?
Twitter needs to add its own influence score, like what Klout (a company I helped start) pioneered, to provide a better way to filter users and their tweets by influence.
A Twitter influence score would add another level of social reward to participating in Twitter, especially if this score doesn’t simply favor massive celebrities with enormous followings but rather is a true measure of a user’s expertise and potential to drive downstream engagement on topics.
How to do it:
- Influence should be a measure of potential downstream engagement that a user generates, and it might also take into account a formal measure of expertise and a measure of downstream influencers who follow a user.
- Users should have a cumulative influence score, as well as a sub-score for each topic or hashtag they are influential for.
- Ranking users on a topic by their real influence would create a competitive cycle where users would once again try to be in the top 1000 or top 100 for topics on Twitter.
- Every hashtag (and even every proper noun) on Twitter should have an auto-generated portal page with a leaderboard that ranks the influencers and content for that topic.
- Twitter could provide a badge for people who rank in the top 100 or top 1000 on a topic. This badge would appear on their Twitter profile and next to their favicon, and they could even put it on their resumes or on LinkedIn and other sites.
Implement congestion pricing
The signal-to-noise problem in Twitter has a negative impact on engagement and reward to consumers, publishers, and advertisers. What this comes down to is the fact that there is no cost to Tweet. A spammer can send thousands of useless messages at no cost drown out legitimate tweeters who don’t spam.
When you have a limited supply of something (in this case, user attention) and a potentially unlimited demand for it (in this case, content providers who want their attention), the solution is congestion pricing.
Uber, for example, has shown us all how congestion pricing works. We can also see congestion pricing in the rise of pay lanes on busy highways and in auctions for limited ad slots on search engines and TV networks. It’s time for Twitter to do this with tweets.
How to do it:
- Every tweet is essentially an ad to get attention from someone.
- Anyone should be able to tweet for free – which is like posting an ad for free. But free ads should get lower exposure than paid ones.
- Anyone should be able to optionally pay to boost a tweet to buy it higher visibility.
- Boosted tweets would be displayed at the top of feeds in a special section according to a dynamically priced auction run by Twitter.
- Boosting a tweet should be easy and built into the Twitter publishing UI/UX and API.
- Promoted tweets are simply the most premium boosted tweets – and they appear in an even more highlighted section above the regular boosted tweets.
- Users should be able to pay into their Twitter account to buy points that they can spend to boost posts.
- Points can also be earned by engaging with Twitter (see next section).
- Not only would congestion pricing make a ton of money for Twitter, but it would solve the signal-to-noise problem almost overnight. Some users would be upset and say this is unfair. Too bad. It’s better for Twitter as a whole, and ultimately for them too — because it would make Twitter far more usable and engaging for everyone.
Add a points economy
There are several ways users can engage in Twitter. In the early days of Twitter, as I’ve mentioned above, the potential reward per amount of effort in engaging was higher, and this served to spur a lot of engagement. But today the potential reward for engaging has decreased – it’s simply harder to get attention than it was before. Twitter can solve this by further gamifying engagement to provide more potential reward.
How to do it:
- Twitter should implement a virtual currency system within Twitter, where users can earn and spend points (call them “Seeds” perhaps?).
- Twitter should reward users by enabling them to earn points when they do things that benefit Twitter and Twitter advertisers, such as posting tweets that earn a certain number of responses from other users, clicking on ads, sharing ads, or hitting “achievements” such as getting 20 retweets on a post, or getting 10,000 real followers, or ranking in the top 100 or top 1000 influencers on a topic, clicking on a certain number of tweets per day, etc.
- Users should also be able to earn points when other users optionally give them “tips” on Tweets they like.
- Users should be able to buy points for cash as well, by paying into their Twitter account.
- Users can spend their points to boost posts (essentially this is a micropayment to increase the visibility of a post), buying actual ads (promoted Tweets), or on giving “tips” to other users on posts they like.
- Points could be redeemable at an exchange rate for cash, or goods and services, just like loyalty programs for credit cards and frequent flyer programs.
A simple points economy would help to increase the potential reward of engaging in Twitter and could be very profitable to Twitter as well. Every user would become a potential paying advertiser. Tens of millions of users, each spending a few dollars a month, translates into meaningful revenue.
Improving the noise-to-signal is just one area where Twitter could greatly improve. The other two areas the company should focus on are:
2: Enabling better search and collection of tweets
Twitter’s content paradigm is long overdue for some evolution. Here are some ideas for breathing new life into it.
Enrich tweets with richer metadata
CEO Jack Dorsey’s recent proposal to expand the length of tweets beyond 140 characters is a great first step, but adding some extensible metadata fields to tweets can have an even greater impact in the long run.
How to do it:
- I’ve written elsewhere about how adding some very rudimentary semantic metadata to tweets could have a huge impact on how Twitter content can be used and on the value of the content.
- The basic idea is to add support for machine tags in tweets and extensible metadata fields to tweets.
- Links, usernames, and hashtags are metadata and should no longer count towards the character count of a tweet.
Lead an open standard for cards
I’ve written elsewhere about how important cards are as the new medium for mobile content. Twitter is well positioned to create and shepherd a true open standard for cards. Whoever does this first wins a lot of karma points. But more importantly, they get to define the standard to their advantage. Twitter should seriously consider doing this.
How to do it:
- Twitter should release an open standard for cards, the next medium after Web pages.
- The standard should support a separation of design, content, and metadata.
- The standard should support machine readable semantics.
- The standard should be given to the W3C at some point.
Make tweets more reusable
Currently the Twitter user experience is a river of news where as soon as something falls below the fold it starts to lose value at an accelerating rate. Twitter content is just too ephemeral; it has a shelf life of perhaps a few seconds — a few hours at best.
But what if there was a way to extend the shelf-life of a tweet, making it valuable for longer? Not only would that provide more potential reward to the author of the tweet (or the advertiser), but it would also provide more potential eyeballs (and ad impressions) to Twitter in the long run.
How to do it:
- To accomplish this, there needs to be a way for users to privately collect and share tweets they want to keep and reuse as knowledge in the future.
- Simply add a “Collect” action to every tweet. Users could collect tweets into their own collections, or to collections for groups.
- When you collect a tweet, you could also add your own tags (private to you and/or your group) and comments. Collections would also be searchable in many different ways.
- Collect tweets like you collect bookmarks — for hobbies, projects, teams, organizations, and events. Search them instantly to find your knowledge when you need it. Suddenly, collections of tweets become searchable knowledge.
- Better yet, open up these collections via the Twitter API as well so that third-party apps can create new and innovative ways to organize, search, and visualize this knowledge to make it even more useful. All of this drives more engagement to Twitter and Twitter content.
Tweets would have longer shelf lives and would generate more future value to their authors and to Twitter. And by the way, Tweet collections would still run Twitter ads. And since every Tweet would be a live dynamic card, it could update and provide new content and even provide analytics to the author or publisher when it is accessed again.
3: Focusing on being a network not a destination
Twitter’s business model has evolved a lot, but I still believe there is a larger opportunity for the company hiding behind a paradigm shift that hasn’t happened yet, but could.
The ultimate 100 billion dollar question is “what is Twitter, really?” Is it a media company? Is it a portal destination? Is it an app? Is it an infrastructure? Or is it a network? How you answer this question makes all the difference in the world.
I think the best answer is that Twitter is a network. It’s a network for distributing and accessing the world’s real-time information. This is a much bigger idea than being just a destination or an app.
Twitter destinations and apps are simply endpoints that provide views into this network. The real value of Twitter is the network and the network effects it enables.
Ironically, during the first years of its ascent, Twitter behaved more like a network than a destination. And that mindset correlated directly to Twitter’s massive growth curve at the time.
It appears to me that when the company began to shift towards thinking about itself as a destination, not a network, that ecosystem growth, user growth, and engagement began to level off. Maybe there is something valuable in this observation?
Let the ecosystem innovate
Bill Joy famously said, “Innovation happens elsewhere,” and he was right. There is always more innovation potential outside an organization than within it. The key is to figure out how to catalyze it and profit from it.
Twitter is ideally positioned to do this — as long as it thinks of the business model as a network not a destination.
How to do it:
- Relaunch a new version of the public APIs.
- Let small and large developers compete on an even playing field to invent new uses, interfaces, and features that drive adoption and engagement for Twitter.
- Make Twitter extensible with apps that developers can create and run inside Twitter, like Facebook apps in Facebook.
However, while the reasons were legitimate, the side-effect was that a thriving ecosystem of legitimate third-party applications was turned off. That ecosystem happened to drive a lot of long-tail use-cases and traffic to Twitter, but more importantly, it was constantly innovating new user interfaces and features for Twitter. It was an incredible engine of innovation.
By finding a way to restart that ecosystem, Twitter could energize thousands of developers and millions of users to help it innovate. Let a thousand flowers bloom. These were the core founding principles of Twitter, in fact. As I have sometimes said, “Twitter’s future is really its past.”
Unlock Twitter’s monetization potential as a network
I have written elsewhere that I believe that thinking of Twitter as a network yields a bigger return for investors than thinking of Twitter as a destination. Destinations are old-paradigm concepts. Networks are the future. In fact, the network IS the destination.
Instead of trying to get all the eyeballs to come to a single portal-style destination site or app, Twitter can be a much larger and more entrenched global utility by being the network powering millions of different destinations and apps.
The key to this working is for Twitter to be open while still controlling the content standards, monetization, and advertising spots, across the network.
I have for many years been refining what I believe could be a better model for doing this. Twitter can be bigger, and make a lot more money, as a network, than as a destination.
How to do it:
1. Reduce barriers to spreading Twitter content
- Twitter content is like a virus that spreads Twitter monetization opportunities. So let the content spread more easily, as long as it carries Twitter monetization with it.
- Anyone should be able to reuse and redistribute Twitter content for free via the Twitter APIs (which would need to be modified and reopened to support this concept). The APIs would be similar to the previous public APIs; they would not provide the full firehose.
- Third-party apps (yes, even third-party Twitter client apps) would be permissible as long as they don’t change the original content (which should come pre-formatted as cards from Twitter, where the authors get to control their formatting).
- Augmenting the content provided by Twitter with additional metadata or context (such as related links, comments, etc.) would be permissible as long as the original card content is not obscured or altered.
- In addition, third-party apps should be required to display Tweets in the order provided by Twitter, unless they pay to buy out the ad space (see point 3 below).
2. Anyone who displays Twitter content in a consumer facing app MUST carry Twitter’s ads
- Any app that displays tweets must agree to carry Twitter’s ads in the content.
- They must only display Twitter ads (also provided as cards) from Twitter in or around that content. No other ads would be permissible in Twitter content. However, it would be permissible to run ads in their own pages or apps outside the Twitter content (not inline).
- Ads would be cards served by Twitter and would be displayed on a variable frequency within any list of tweets. The frequency would be determined by Twitter’s algorithms.
- Twitter ad cards would be served in a targeted fashion by Twitter. Apps that use Twitter content must allow Twitter to place and read targeting cookies on their users as well.
3. Apps that don’t want to run Twitter ads would have that option by buying out all the ad space from Twitter
- App providers that wish to carry Twitter data without Twitter’s ads, would have the ability to do so. But they would have to buy out 100% of Twitter’s ad slots in the stream of tweets they display. This would be done by calling a different content API from Twitter that would bill them for each ad that Twitter does not display.
- Companies that pay to buy out Twitter’s ad inventory on tweets could insert their own ads within or around the tweets they display.
- The beauty of this is that Twitter would be paid for every ad slot in this scenario. In fact, this could be potentially more profitable for Twitter because 100% of its ad inventory would be paid for in any app that chose to do this.
4. Twitter would still charge a CPM for full-firehose data
- Full firehose data has significant costs to Twitter to host and serve. It’s quite a different animal than rate limited or time limited API pulls, which should be free like they used to be).
- However the CPM to buy firehose data from Twitter could be dynamically priced and therefore lower than it is today, because Twitter’s total daily volume continues to grow, making the relative value of a tweet lower than it was before.
- You can think of this just like inflation in economic terms — the more tweets there are, the less valuable a single tweet becomes, so the cost to buy a tweet should be adjusted down accordingly. Effectively it would be a built-in volume discount for firehose customers.
- This will make it possible for brands and analytics companies to continue to afford to analyze all the data they need to analyze without the price constantly climbing as data volumes increase.
- Even with a dynamically priced CPM for firehose data, Twitter would still grow data licensing revenue because more firehose use would result from a lower cost per Tweet.
- Coupled with the other suggestions in this article, this could enable data licensing to become a much larger portion of Twitter’s revenue mix in the Twitter-as-network paradigm.
Become the world’s real-time marketplace
Perhaps the biggest untapped potential for Twitter to grow its revenues isn’t advertising but commerce. If Twitter adds richer semantics, as I’ve suggested above, it could become the world’s biggest real-time marketplace. Think of it as Craigslist + eBay + Alibaba combined, and all in real-time. This could ultimately be a bigger play than advertising for Twitter.
Think about it this way: Twitter is already a marketplace for attention — in particular, attention to content. But what if it could also be a marketplace for products and services?
How to do it:
- There would need to be a standardized taxonomy of cards with the appropriate metadata fields for describing various commercial offers and requests for different types of goods and services.
- There would also need to be a way for users to create requests to get cards that match their criteria. Users would need ways to create weighted parameterized searches such as “I want to find offers of used Tesla cars for sale, and price is more important than mileage.”
- There would need to be a marketplace portal for navigating categories of these types of commercial offer cards. For example, there would be categories for real estate, personals, services, cars for sale, consumer electronics, jobs, and every other commercial category of interest.
- Twitter would provide and standardize this categorization scheme, but an ecosystem of third-party apps and services could also provide interfaces and tools for participating in this marketplace.
- In addition, by enabling users to pay to make their Tweets more visible, marketplace participants could compete and spam could be weeded out economically. This is a win for everyone.
- Twitter could provide optional payment settlement, purchase protection, escrow, premium verified seller and buyer programs, and other transaction support, with appropriate fees charged to the seller and/or buyer — an additional massive revenue stream for Twitter.
- Most importantly, users would need to be able to choose what channel they were looking at — so if they only want to see content, not commercial offers and requests, they can do that. On the other hand, if they are looking for real-time deals or classified ads, they could do that efficiently without having to pore through irrelevant content.
The ideas above, if adopted, could go a long way to restoring Twitter’s former growth rates and could also help Twitter realize the massive untapped potential of the network it has built. I am hopeful that Twitter’s management team will adopt some of these suggestions in coming years. If this happens, then today’s Twitter stock price is an incredible bargain!
Nova Spivack is a technology futurist, serial entrepreneur, and angel investor, focused on next-generation of search, AI, big data, and the Web. He was recently ranked in the top 20 futurists based on an analysis of social and Web influence. He is cofounder and CEO at Bottlenose.
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.