HackBits by Dane Lyons

You have 2 options

1) Give up now and return to your regularly scheduled Facebook browsing.

The problem with blogging is it takes persistence. People start blogs all the time, write a handful of posts, and give up because nothing seems to be happening. It takes 50+ posts just to get your feet wet.

How can you get over the hump?

Just get out there and do it.

Didn't work?

It's ok, brute force isn't for everyone. If you need that little extra motivation to keep you going then I recommend finding a blog buddy. Find a friend who wants to start a blog and encourage them. Suggest ideas for posts. Comment on their work. By helping them, you'll get inspired for your own blog and they'll help push you.

After you've got a buddy, sign up for PowerUp and start tracking your blogging activity with your friend. By seeing a visual representation of your work, you'll naturally want to do more.

People have unique tastes. They expect you to figure out their love for chocolate, cookie bits, mint and raspberry syrup pretty damn quick. They have little desire to explicitly express interests and they vote with their indifference.

What does that mean for you as a developer?

Machine learning is not a silver bullet for user engagement. People don't have the patience to sit around as you guess wrong over and over again. You might be able to buy yourself some time with transparency, good messaging and incentives but getting the UX and the backend right at the same time is non-trivial.

Shortcut the learning process. We're not living in a world of isolation. There is probably a metric ton of data available for every one of your users the moment they sign up. Just looking at public social data alone can teach you a lot about the person. Sure a lot of the time we're spending our time talking about trivial, completely irrelevant topics but we're also leaving behind important clues in the process.

Pay attention to indirect associations. You aren't always going to have a treasure map of public data for a user but don't let that stop you from testing a hypothesis. Amazon made a killing on suggesting items to users based on purchase behavior from other users who viewed a particular item.

Do things the old fashioned way. Have humans analyze the data to make recommendations. Humans are great at identifying signals an algorithm wasn't trained to consider.

So what is friendsourcing?

Above are some of the traditional ways you might take advantage of user data to make better suggestions and improve the overall experience. This is probably not a new thing but I'd like to throw one more option into the hat.

In some cases you can incentivize friends to make suggestions. Nobody has more context than a friend. They've got all the right signals and already know exactly what you like and what you don't like. We should be finding better ways to take advantage of this insider knowledge. This doesn't mean we should develop a system for friends to sell each other out for cold hard cash...but we can create a system where friends can help each other in a more meaningful way.

I see 2 important use cases here:

Friends can take on tasks that they enjoy doing for their friends that dislike those particular tasks. An obvious scenario here might be a friend shopping app.

If you're using a great app you might be willing to help pre-train the app to your friends likes/dislikes at the time of sending an invitation.