The magic of starting up

23:57, Jun 10 2014
Lightning Lab programme director Dan Khan

Ever wondered why there's a handful of mobile apps that just work so much better than others?

You've got no idea how: it's as if by magic.

Take for example Zite, which has changed the way I consume news. The app provides me a personalised stream of recommended articles that are not only relevant, I want to read them.

Over time it has learned my preferences so well it now shows me entirely new things I will be interested in reading with almost perfect accuracy.

Another miracle product is Dropbox. How do I share my photos and files with family, friends, and colleagues? I simply drop them into a folder on my computer. That's it.

For me this is the wizardry of the app. There's a certain hidden pleasure in not understanding how it works but instead submitting to the fact that it will happen every time you open the app, like waving a wand.


The user experience is taken to the most abstract level and the simplest solution to a problem is found.

Sending cash has become alchemy with Square. The service makes transferring money to someone as easy as sending them an email by cc'ing with the required amount in the subject line.

Square then sends you an email asking you to fill out your debit card details on their secure site, and asks the recipient to do the same with their bank details. A day or two later the money is transferred.

Such simplicity of solution and clarity of purpose seems magical.When designing their products many app builders optimise one or two steps in a user's workflow, with each step providing a little additional value.

In building magic apps you need to think differently, by optimising all of the steps into one and providing significantly increased value across all of those steps.

Rather than collating and presenting complex information to allow the user to make a more informed choice, magic products learn the user's behaviour and make the choice for them.

If I peel back the conjuror's curtain on Zite, they are looking at what I read - how long I stay on a page, whether I like the article, and if share it with friends.

They then likely generate some profile of the articles and categories I read, finding others with similar profiles to me and use what they read as a good basis to match what I'd like to read next.

Sure, they could ask me what I want to read, but would this solve my problem of figuring out what I should read next?

It's these inner workings and algorithms that become the intellectual property and competitive advantage of your company, which is another good reason to keep this voodoo inside your startup black box.

Henry Ford famously said that if he asked customers what they wanted they'd say faster horses, but it took an innovative leap of faith to deliver the Model T Ford.

The same question applies when building startups - how can you create a magic product that seemingly automatically solves a user's problems in an educated and simplified way without just building a faster horse?

Dan Khan is programme director at startup accelerator Lightning Lab