There's an old telecoms joke that when electro-mechanical exchanges were first being replaced by digital, someone asked how many staff the new systems would need. The answer was that you'd have a man and a dog - the man would feed the dog, and the dog would stop the man from touching the equipment.
So far, this has been pretty good way to think about Twitter. It's something that was not so much created as discovered, and part of the genius of the early days was how few actual decisions were taken. If you don't quite know what it is or why it works, but it seems to be working, then touch as little as possible. Don't let the man touch the equipment - you might break it without realizing.
The problem is, of course, that it's stopped working, or at least working as well. Vastly more people have tried it and given up than use it today and growth has slowed to a crawl. So it is time to gain some self-confidence, slip the dog some artisanal locally-sourced salami and start fiddling with the equipment? And if, so, what's the underlying problem that you're trying to solve? Do you know?
Starting on Twitter has often reminded me of opening a blank spreadsheet in Excel - it can be anything you want, but you’ve got to put the work in first, and if you don't put the work in it's useless. Hence, Microsoft put a lot of work into product discovery, as you can see in this ‘file new’ window. "It can be this!" It can be that!" Some suggestions for new screens, views or filters for Twitter look a lot like this - a ‘baseball’ screen, say, instead of a ‘todo list’ screen.
On the other hand, one could also compare Excel to Photoshop. Both are powerful, sophisticated, complex tools that are unmatched in filling the needs of a core professional market. But Excel is also a very broad and flexible tool that people use for lots of quite different things. People make timetables and task lists in Excel. Photoshop use is nowhere near that broad.
Right now, Twitter only has those core users. It only has the people who use Excel for DCFs - it only has Photoshop users (and yes, not all Photoshop or Twitter users are the same, but it's a narrow set). Can it be as broad as Excel became - can it spread outside those core users? If so, will it get there by teaching everyone to do discounted cash flow models, or by showing them they can use it as a loan calculator and a shopping list and never touch formulae or make a chart?
That is, does Twitter need better ways to teach people and help them to think about it the way the current user base thinks about it, or should it create and enable entirely new ways to use Twitter? I think this is the fundamental question - does Twitter need better execution or a new approach? Is Twitter a better on-boarding flow, a ‘who to follow’ engine and a new tab or two away from growth? Or does it need to embrace fundamentally new ideas of what the product should be?
So, do you change the linear timeline, wherein the good stuff is buried after 20 minutes? How would that reconcile with the power of Twitter for realtime? Would having the 'right' tweets instead of a firehose of thousands per minute on a hashtag make realtime better or worse? Do you change the 'who to follow' model, that forces people to spend weeks working out where the content is, but gives you rich and meaningful interest data? How do you lure back in the vast numbers of people using twitter without even logging in? Amongst other things, those people are not participating in conversation (i.e. @ replies) at all - does that matter? If someone reads Taylor Swift's feed every day but never logs in, are they a more or less valuable user than someone who follows 50 people and logs in once a week?
Again - do you create products to make the 'hard' aspects of Twitter easier, or do you make them cease to matter? You can get value from Excel without knowing how to make formulae - how far should you be able to use Twitter without following people, or posting? How would you make that great too?
A core part of the development of Twitter (and Excel) has been the embrace of desire paths: look at what the users do and follow. But the desire paths by definition can only come from the people who are already using it. If you want more people to follow, you need to take a view yourself, for the first time, on what else Twitter could be.