Monday, October 24, 2016

What should Twitter be?

Twitter is in the news for the wrong reasons. On one hand, the subscribers growth is slowing, and profits are not up to the mark. There is confusions as to the business model and if it can ever make money. Google, Salesforce, Disney and others were mulling acquiring Twitter. I think Twitter is more valuable than even Facebook (if you ask me) and it will be a core-architecture for the social web. Here are my thoughts on Twitter's business model and how. 

[Shameless plug: The background for the discussion is my interest in business models and my ideas of firms and how they create value. The frameworks are detailed in my book "Understanding Firms: A Manager's model of the Firm" - please buy it. ;-)]

About my twitter use first
I have been a twitter user for some years now. I don't over-tweet but I guess I should be in the approximate middle as to tweeting. But when it comes to consuming the tweets, I think I am a super-user. I have neatly segregated lists and I scan them using Flipboard and Tweetdeck. I also like to share what I read on / through Twitter. You can find/follow me @rahuldeodhar by clicking my username. So let us begin.

----

Tweets fall into following categories -
  1. Link to content from the web - example linking to videos etc. [Sharing domain]
  2. Comment on content on the web (on twitter / outside twitter) [ Commenting domain]
  3. Opinion/Views about things - including witness view, etc. [Content domain]
  4. Personal updates - status updates [ Personal domain]

The Twitter Stream is like a river with all these mixed up from various sources. It is impossible to make head or tail of it if you want to read it. It is like watching the river from a bridge. It is all ok for some time but is not critical - it is a leisure activity. If you really want to do something interesting with it - you can't. If you want to track trouts - well from your perch you cannot. If you are a master user you are like a diver facing the current trying to analyse the water. Whatever analysis you do is useless because the stream has changed by then. 

So that is why advertising on Twitter is so damn difficult. As difficult as it is to interrupt a diver studying water with a TV commercial. The diver doesn't dive that often, and when he does, he doesn't want to look at your TVC. It is very difficult to read the twitter stream - ads make it worse.

Twitter is like a monologue for most of the people, most of the time till others start commenting, sharing and you start getting reactions (not the button based reactions but real comments). Then it becomes interesting. If you are too popular, it turns nasty (sometimes) and resembles a bar fight or a cake fight from Charlie Chaplin movies.

---

But Twitter can become more relevant
Twitter can become the default commenting engine of the web - become disqus++. Twitter sits at the junction of comments and sharing. This is win-win for websites that generate comments as well commentators and for twitter. If someone is posting a long form comment, the first 140 characters will only form part of the tweet. This will force people to summarise their comments and it will be easier for authors and really interested parties to parse the detailed long-form reply at the web-page itself. [swelling the stream]

It can also become default Reviewer of choice. Reviews are essentially comments to something - either product, services etc. Ease of being able to pinpoint what we are reviewing remains a challenge. So say when you are reviewing a Phillips table lamp should it be tagged to product page in your country or to product page of Phillips international - well those things is what Twitter R&D spend of $800m should be used for. It can substitute product reviews in amazon, sites like good reads etc. [swelling the stream]

But it may be better to avoid hosting content. A few good things Twitter has done is to integrate photographs and videos into the stream. Twitter can choose to partner with YouTube and Google Photos or say Flickr for it or it could go on to become a content development platform - like medium (say). To me, there is value in letting content reside with YouTube or Flickr and using the Tweets itself to gather data. There are many arguments as to whether sequestering content behind login walls is good or bad. (Facebook likes it, google is fairly open). I prefer open architecture. It is like building cities v/s building walled communities - cities are much nicer. [swelling the stream - though not too much]

If you note carefully, Twitter can address two strong models - create once publish everywhere (COPE) and Diverse Information Sources in one stream (DISOS). I concocted the last one so apologies if it sounds clumsy. While the first allows easy of creating content, second allows ease of consuming content. Remember for advertising models consuming content is important. 

Twitter also need to universal Tweeting, specialised Stream reading. It already has universalised tweeting. You can tweet from any app/webpage etc from phone or computer. Or you can go to the Twitter website or its app and tweet from there. I use the Twitter website / App for reading the stream rather than tweeting itself. But reading is cumbersome. Tweetdeck is one way of segregating content based on usernames and hashtags. But even Tweetdeck is difficult to parse. Flipboard is easier to parse but a  bit weird in formatting. Twitter needs to develop apps that help user read the twitter stream better. [Reading the stream]

Twitter stream-reader, must complete the information picture the way Microsoft's Photosynth compiles the photographs from various sources. Twitter is it's information equivalent. If I pivot the twitter stream on a person, it should give me the subject-clustering of tweets of that person. If I pivot the twitter stream on a topic, it should people-cluster the tweets into groups - my followers and within that based on my lists. In both these views we should have ability to go back at least one day (for consumers). [Reading the stream]

---

How to Generate Revenues
Twitter has been a platform of choice for news-dissimination. Twitter must take it to its logical conclusion. Twitter can replace the newswires - all of them in one go. For this, Twitter needs to use pre-identified usernames. It already does that with verified accounts. It will need a customised App for distributing and reading newswires. I cannot see why it cannot be done. By itself newswire business is about $3-5 billion with possibly about 12-14% profitability. For twitter, it may be more profitable. [separate newswire business]

Twitter based News channel is also a possibility since the news is there in the stream, videos are there and you can combine those using pre-selected usernames (handles) and tags. It can set up programming automatically just by relevant people tweeting about it. Facebook and YouTube have started live video dissemination, but in a stream they make no sense. But curated videos allow you to create topical channels. Twitter should also be able to hash YouTube existing video library into a proper playlist of sort. Advertising through this will be easier. [reading news stream]

Twitter Stream-Reader Pro can be a fully loaded stream-reader that can help clients get easier view on the data stream in terms of their relevance. Imagine Ford Mustang Twitter Stream-REader Pro (FMTSR Pro), it will read the streams about Ford Mustangs and then give you detailed analytics. If I was Twitter, I would hard code "Ford Mustang" into the this FMTSR-Pro and charge Ford for yearly use. The same App with modified hard codings can be deployed for others say Lego. For Glaxo or Pfizer it can drill down doctors and non-doctors into the categories. Twitter currently does sell the stream analytics but the revenues from that is quite low - about 10% of the revenues. I would presume it should account for 80% of the revenues. So there is a lot of potential in this. [ corporate stream reading]

Customer Service Pro can mine the stream for companies listen for customer complaints and engage the customers using DM. This can be even now but I can see Twitter being able to add analytics to aid the customer services. Personally I had super experience from Hyatt who solved my problem through twitter. From then on I personally air my grievances on Twitter so that if any company is listening then I can get help quickly. Indian ministers use twitter to solve emergencies - Railways and External Affairs ministries are quite active. I presume companies would love to use Twitter to solve their customer's problems. This function should ping companies when a customer tweets about bad experience he is having with the company. It should allow the companies to set criteria as to when alerts are triggered from the stream reader.

Twitter Topic Tracker can be a reader that tracks specific topic - say wind surfing, pottery or something. At present users have to create the lists as per their own specification and these lists can be public. The problem is in a list of economics we get general tweets (say happy birthday to my daughter) by economists but miss the economics tweets by other people who are not in the list. There should be some way of fixing this. Advertising in the viewer of this Topic Tracker will be more relevant and therefore more lucrative and sensible. [ to curate better advertising]

Twitter can create publicly sourced Subscription Magazines just like Flipboard or Paper.li with relevant tweets compiling into readable magazine. Twitter can take share of these subscription through a twitter-owned store and distribute subscription to contributors (original and those sharing them), using some acceptable equation. (So content creators take 75% of the revenue based on reads, clicks, shares etc. the individual share of creators can vary While those sharing get 1% of share of the content creators and Twitter takes 20%). Using the stream, Twitter can compile edited-book like special editions giving complete spectrum of opinions on selected topics. There can be advertising within these magazines and revenue sharing with content creators. So for example, Twitter can compile a special edition on "Manufacturing Policy" or "Dodd-Frank Bill" by experts and add value to the journalistic discourse. [content curation and advertising]

Twitter can also give a Event Live-view using tweets by general public (those by reporters go through the wire I am presuming) and create a Stream reader view that gives overall picture based on live tweets as to what exactly is happening. This might require integration with AI algorithms to parse value of information shared by a particular Tweet. Currently, search results that give "top tweets" tends to tell us this but it needs improvement. [Stream Reading] 

---

Now with all these developments, I think Twitter should have been more valuable to news websites. It is indeed sad that Twitter cannot monetise itself better. In the present post I did not want to consider the operational parts of Twitter strategy because John Hampton has already considered them in his two fantastic posts Some comment on the Twitter buyout rumours and here Measuring how bad Twitter is. I hope Twitter heeds them. John Hampton is seldom wrong. Fred Wilson and Union Square Ventures were early investors in Twitter. I hope they understand the possibilities and take corrective actions.
 

Disclosures: I have no investment in twitter.