Charles Dowd will be on a Data Driven Product panel at the FutureScope Conference on May 10th. FutureScope is Irelands only conference that promotes collaboration between entrepreneurs and large enterprises.
Plynk is money in a message. With Plynk, money moves instantly just like text or images do. When you send money between people it moves instantly, they have the money right away and they can use it directly online You also get a MasterCard which allows you to spend online or in apps.
Plynks product design is driven by data. We never ignore it.
6 years ago I was working in Facebooks Menlo Park HQ with Mark Zuckerberg as he pivoted the company from desktop to mobile. Why? The data told him if he didnt Facebook would die.
All staff were shown a usage graph of the future which said ‘Desktop is dead, mobile is everything, we don’t have a real mobile product.
The reaction from Mark was decisive. Let’s change the entire direction of the organisation from the top down. Were mobile first from now on.
The future wasn’t set, but everyone saw that the analytics on user behaviour were only going one way. When extrapolated forward just a few years it was obvious what needed to change. The entire company then switched its focus towards solving that problem. The switch from desktop to mobile had been flicked.
In many companies, the inability to build products using data is a cultural problem. If the culture of your organisation is command and control, then the data that you capture is usually used to confirm your bias. In Facebook the data came from the bottom up was taken on board by leadership and the rest is history.
Ive now taken the lessons I learned at Facebook and try to build a culture of data-driven design thinking at Plynk. One recent example as an issue in our sign up process. We had a choice to make when the user first opened the app: when they sign up should the user choose a conversation to join or should you actually force people into the sign-up conversation?
One view was that you shouldn’t make decisions on behalf of the customer. The other one was that if you dont, youll lose people on sign up. We decided to run some tests and it turns out that theres a 25% conversion difference between putting people directly into the chat and leaving them with a choice. So the data said: ‘Put them into the chat directly! That’s a real decision point.
We have also used data to find problems. We discovered a really obscure bug that wasnt being reported by our crashlytics. By examining the same process on two different platforms, the one on iOS delivered a 12% lower conversion rate than it did on Android. We were initially thinking ‘Oh, there must be something terrible about the way iOS works’ but it turned out that there was a bug that we hadn’t noticed now we’re able to fix. Data pushes you towards asking new questions about what you’re actually doing.
On a larger scale, lets say there are twenty different variables. You can build twenty by twenty different tests and then run them all in parallel. You can then actually see where the individual binary decisions are made. Your machines test the various different combinations of your parameters and then you ultimately find the one that says ‘This one wins, this combination wins’, and then you go with that.
What you should be looking for is not to drive people towards what you want them to do, but to find the things that they want to do. Make sure that they align with what you’re trying to achieve and optimise for the features that drive engagement.