Sometimes, when you have your head down working on customer suggestions or platform improvements or bug fixing or other stuff, you can miss a really great idea. Or those great ideas go into the ‘sometime’ pile (aka ‘the pile you never have time to deal with’).
Which is why, at the end of December 2013, the dotmailer product team took some time out of their usual work and experimented with totally new ideas instead.
Internally, we called this the ‘dotmailer hack week’ (although back then it was actually only three and a half days). And it all went so well, we did it again at the end of last year – only bigger and better. Nearly every engineer and designer in the company took a week off their normal work, and focused instead on building lightweight prototypes of things we’d perhaps otherwise not build.
We got to ask ourselves: ‘What would really help our customers? What would change their lives? What could we do that no one else is doing?” – and the whole dotmailer workforce took on the challenge, generating almost 100 ideas.
Nearly all would have helped marketers in some way (although one person did suggest we write and perform a dotmailer song), but together we whittled them down to just 19 hacks.
And a week later – on the day before Christmas – we all huddled together and watched three hours of new, alternative, bleeding-edge marketing technology solutions. We also had a company-wide vote to find out what our collective expertise thought was best – and I’m proud to share the winner, and some of my other favourites, below.
And the winner is…
The amusingly titled IAmBadWithNames scooped the top prize. It was dreamt up by Dikaios Papadogkonas, and was hacked by Dikaios and Steve Lillis, who both work on the back-end technology that powers dotmailer. I had a few words with them recently on their hack.
Me: What was the big idea?
Dikaios: We store user navigation data after a contact follows a campaign link. We wanted to put this data into a probabilistic model, and then build a query engine that could answer complex questions and identify navigational patterns. Specifically, we wanted to build clustering and classification algorithms that allowed us to group users into patterns, and associate new users to one of these patterns. In essence, we wanted to model how likely it would be that customers who clicked a link would then go on to convert.
Me: How did you think it would help our users?
DP: We don’t currently provide any analysis on the data we record. By doing so, we wanted to show that it could help our users understand how their website is navigated, how contacts actually used it after receiving a campaign, and whether the contacts reacted as expected.
Me: What was the hardest part of the hack?
DP: In a small amount of time we had to build a large number of algorithms from scratch and visualise the result in a meaningful way. Even a small bug could create a completely meaningless result.
Me: And what about that name?
DP: I was hoping that by using the name ‘IAmBadWithNames’ someone would help me pick a better one. But they didn’t.
And some more…
Elsewhere, FTPr (pronounced eff-tee-pee-er) also caught our imagination. It was hacked by Garry Marsland, Boris Maslennikov and Ilya Polyakov, whose day jobs are spent working in our custom integrations team. Garry said, “Although we have two methods for importing contact data – the web application and the API – we still deal a lot with automated SFTP”.
The reasons are clear: “The web application is easy to use, but requires someone to log into dotmailer and perform the process manually. The API allows people to automate the process, but only with a developer’s help. Both methods have a 9MB upload limit, and neither method will automatically create address books or contact data fields.”
He went on to say, “The idea behind FTPr was to provide a third method for importing. One which was flexible, had no real limits, and which could be used manually by someone sitting in front of their computer or plugged into an automated export from another system.”
Attempting to make learning dotmailer easier (and, dare I say it, more enjoyable) was the dotmailer achievements hack, worked on by Danny Rendle, James Marais (both back-end developers) and Sani Chan (our visual designer).
According to Danny the big idea was to, “give our users a sense of progress through lighthearted feedback, and provide an opportunity for them to discover other features”.
And finally (because I’m at risk of writing dotmailer’s longest ever blog post), I’ll leave you with dotmailer for Slack. Ok, this was my idea – but the execution by Simon Letch, a back-end developer, means it’s worth a mention. Slack, for those of you in the dark, is a team communication tool. It’s used by the products team here extensively – and we wondered, ‘What if marketing teams used it? How could it help them get more out of dotmailer?’
I asked Simon for his thoughts. “We wanted our users to receive dotmailer notifications through a Slack channel that they created. Things like campaign creations and sends, contact imports, subscribes and unsubscribes along with campaign reports a number of hours after each campaign was sent.”
But Simon didn’t stop there: “Grander plans formed the more I thought about it. I built an ‘alive’ dotmailer that users could talk to via a bot on Slack – so they could ask dotmailer for information about a particular contact or a particular campaign.”
Here’s to 2015
Of course these are just one-week hacks, and are a long way from being production ready. But the product management team (that I’m part of!) is already thinking about the real-world applications that some of the hacks may work for. So don’t be too surprised if you see things appearing in the app that are inspired by what you see here.
And I’m sure towards the end of this year we’ll do it all again. If you have something you’d like to suggest, pop it in the comments sections below.