Data of the heart
Last Autumn, I spoke at Predict Conference in Dublin.
Rather than talk about predictive analytics, a subject I know little about, I wanted to make a case for data of the heart, rather than data of the head.
Why “data of the heart”?
Arguably, it’s because we are dealing with the most complex of data sets, the human being.
Researchers as "Meaning Seekers"
Like any research, the goal of qualitative research is to seek meaning.
An accusation that I've had levelled at me from some client engagements - is that the type of qualitative research we do is less scientific because it doesn't speak the language of statistical significance and probability.
While that's true, I'm not sure if quantitative techniques or tools can uncover what’s going on in people’s heads or their hearts. For that, you need qualitative research.
The approach we take looks like this, which at a level of abstraction, is probably quite similar to any data approach.
- Define the research question.
- Select the approach.
- Gather data.
- Analyse & synthesise (model it).
Ultimately, for us at Each&Other, it’s about moving from research to informed action
The why of a Cholera outbreak in the 19th century
It’s the mid 19th century and a doctor, John Snow, couldn’t convince other doctors, scientists and local officials that cholera, a nasty bastard of a disease, was spread when people drank contaminated water.
Then 1854 happened.
616 people died in and around Broad Street in London's Soho and Jon Snow used the event to help him prove how cholera spread.
He started with some quantitative data - how many people died and then, using a geographical grid, mapped out the deaths from the outbreak.
While often cited as one of the best examples of data visualisation the story generally stops there. Except it doesn’t.
Enter qualitative data
While John Snow suspected the Broad Street pump as being the source of the outbreak, there were certain anomalies in the data. To cover these, he continued to sleuth study using qualitative techniques.
While the dead tell no stories, he went to the living. In particular, a local brewery. No men working there had got sick or died and he wanted to know why. It turns out that the brewers had their own well. He discovered this by asking questions and by direct observation (what we call "contextual enquiry" in the business).
This was important because it helped Snow rule out other possible sources of the epidemic besides pump water.
Why, where and how
Yet, town officials couldn't find how sewage had actually got into the well.
Snow had successfully answered why and where, but not how. And this is where it gets interesting.
Enter a reverend, Henry Whitehead, who took on a quest to discredit John Snow ("you know nothing, John Snow"). His hypothesis was that it was sin and divine intervention. I'll leave that one there.
Following a similar line of questioning, Whitehead found out that a woman had washed her baby’s nappies near the well.
Same approach, then and now
This method worked then as it does now. A good example is the classic Clayton Christensen milkshake story.
Quantitative told one story about milkshake sales, leading to the wrong or incomplete conclusion where jobs-to-be-done interviews filled in the gap discovering richer and deeper insights.
A recent-ish example
When working with a bank a few years ago, analytics told us how often people logged into their accounts and it also told us that they mostly did nothing.
It doesn't take a genius to infer from this that they weren't in fact doing nothing. They were checking their balance. But this didn’t answer why, nor where. The data didn't give us context.
We interviewed a number of banking customers and the types questions we asked were
- Tell me about the last time you logged into your account...
- What were you doing at the time?
- Where were you?
These interviews uncovered answers to the classc
- Why?
- What?
- When?
- Where?
- How?
What was important about this was that it gave us context around the behaviour.
It turned out that by far the strongest emotive reason for people using their banking app was that while they were in a store, they wanted to make sure they had enough for a transaction, and if they didn't to transfer between savings - to avoid the embarrassing "transaction declined" scenario.
And it was with this context or meaning that we made a really strong case for what seems a really simple feature.
From meaning to meaningful.
It’s not a quick balance feature, it’s actually an “embarrassment saving feature”. And it's through simple, curious and open research questions can we uncover real meaning, the data of the heart, so that we can go on to create something truly meaningful.