Thinky.done – early learnings

We’ve closed the garage door on our first experiment of 2014 over at thinky.do and there’s a post about what we learnt about bitcoin from our Open Wallet Experiment there. A few weeks ago we went public about our rebooted approach to experimentation, so what have we learnt about learning one month on?

First off, constraints are both good and bad, or, more accurately, helpful and limiting. We set ourselves the goal of thinking up and launching an experiment, in public, in a 4 week period. And, Yay us, we got it out of the door. Just. We might have had a better conceived, better executed experiment if we’d given ourselves more time, but we might also still be in idea generation phase, filling up whiteboards with hypotheticals and possibilities instead of results and learnings. We did it, it’s done, onto the next doing.

Second, the subject of the experiment. The extended Labs team were absolutely certain that Bitcoin was the right subject for our first foray. Everyone was talking about it, none of us understood it properly, this was our chance to learn. And learn we did. We now know how to buy it, look after it and spend it. We’ve also learnt that bitcoin is a hard thing to think about and a difficult tool to use for experimental purposes. Getting to grips with bitcoin took time and the technical restraints meant several ‘pivots’ before the Open Wallet Experiment got out there. And while we’re not bitcoin billionaires, we’re in a better position to talk to clients about the benefits and drawbacks of cryptocurrencies than we were in January.

And lastly, how we work. We couldn’t have done anything without help from a number of people. Colleagues in BBH, partners outside (particular thanks to the guys at MediaMonks for talking us through bitcoin practicalities), people who emailed and commented on the blog and our G+ page, all helped tremendously.

And so, on to Experiment No2. Trying to remember what we’ve learnt already, and not forget that each month we’re starting over, all over again.


Learning Bit by Bot

Tyrell: Would you … like to be upgraded?

Batty: I had in mind something a little more radical.

Tyrell: What … what seems to be the problem?

Batty: Death.

Blade Runner, 1982

Robotify.me – what we did, what we learned and what we’re doing now

In December 2012 we launched robotify.me, an experiment to test our hypothesis that seeing social media behaviour visualized could actually influence and change those behaviours. Perhaps, we asked ourselves, data visualisation might reveal surprising nuances of social media behaviour which might otherwise be overlooked?


How would it feel to compare activity – likes, links, retweets, checkins, photos – with the rest of the group’s data? Would the transparency of the visualisation cause any changes in social behaviour? Would inveterate retweeters be shamed into posting more original content? Could we encourage more checking in, more posting of photographs, more liking by visualising the effect that it had on the robot?

Robotify.me was also another opportunity to learn and experiment with process. Could we create a service rather than a campaign? Could we work fast and lean and create a mvp? Could we create a product without a brief, without a client?


A little over a year on, the answers to some of these questions are in. The first thing to say is thanks. Thanks to the team who worked so hard (and gave their time so generously) on robotify.me and thanks to everyone who took part in this project. Thousands of you created robots and we loved seeing the project come to life, reading the tweets, hearing your thoughts and feedback on this thing we’d made.

Much of what we learned is displayed in the infographics accompanying this post and some of our early learnings were incorporated into changes we made live on the robotify site in the early go-live days and weeks. Perhaps our major learning was to do with storytelling – if we wanted people to learn a little about themselves we should, perhaps, have shown more, and told more explicitly. Knowing when to intrigue and when to explain is something we will take with us in the future.


We also learned that when you have a team with demanding day jobs it’s impossible to schedule daily scrums and the focus and scheduling required for an iterative workflow are not easily applied to side projects. When we plan future Labs experiments (and more on that very, very soon) we’ll definitely be thinking about the sorts of projects that lend themselves to a leaner approach. Stretch is good, but restraints will help define scope from the very beginning.

So, we’re going to be pulling down the shutters on this particular garage and disassembling the robotifier, cleaning down the work surfaces and wiping down the whiteboard in preparation for a new swathe of Labs experiments, robotify learnings fresh in our minds. We’ll be keeping the service up in it’s current form for another month, so you can still create a new robot, revisit your robot mirror-self or download and print out your robots for your digital files.

Finally, thanks again for supporting our Robotify.me experiment.

Bleep. And out.

Crash Test Dummy

Just as it is very easy to have an opinion about art without knowing how to draw, it’s very, very simple to talk knowledgeably about ‘digital’ without knowing anything about coding – the logic underlying every website and digital product we’ve ever used, tweeted about and, often, criticised.

So as part of Internet Week Europe and with Google Creative Labs we held a Coding for Dummies workshop which was an opportunity for 40 or so people to learn at the feet of some true coding ninjas and take their first, shaky steps along the path of geek enlightenment.

We started from the basics, quickly learning that 40 people cannot transfer a file to the same server simultaneously. We covered basic html, the fundamentals of server-side and client-side interactions moving smoothly onto CSS and javascript embeds before Googler @monocubed wowed us with some experimental HTML5 projects that might, right now, be a little beyond our abilities.

The afternoon didn’t finish with the class able to recreate We Feel Fine or launch an alternative blogging platform. But what it did was give everyone the confidence to go and have a look at a webpage’s source code and the beginnings of understanding why things on the web look and behave the way they do. A bunch of people can now go and play with code, launch a page onto the internet, tweak it, break it and maybe even fix it again.

Below are @tomux’s slides and at the end of them we’ve added a few of the pages that some of the dummies-no-longer created. We enjoyed ourselves so much (and still have so much to learn!) that we hope to do this again some time – keep an eye on our twitter for details.

Huge thanks to everyone who came along and special thanks to Googler’s @tomux, @monocubed and @potatolondon and BBHers @mrjonandrews and @jimhunt_ for patient teaching and technical prowess.

We know code-fu.