#reThink13 : Lead Into Gold (And Back Again)
Tom Ewing with his run down of the themes so far in re:Think 2013: the rise of Data Synthesis
I messed up. In my write up of day 1 I identified CivicScience and Google Consumer Surveys as two great companies who had cracked the problem of distributed surveys – serving extensive surveys one question at a time.
That’s all still true. So what’s the problem? Well, I framed it as CivicScience emerging in the wake of Google’s survey product. They did not. CivicScience’s launch predates Google’s survey launch – and they serve a different business need anyway. Brand stories matter and authenticity matters, so I’m glad to set the record straight on this.
By Tom Ewing
A conference isn’t complete these days without a new word for researchers. No sooner had I fully adjusted to being a ‘data whisperer’, than I go to the ARF and discover I’m now an ‘insight alchemist’, turning base data into gold. Some on the Tweet feed objected to the magical overtones of this coinage. Some objected just because it sounded very silly. But if your Twitter bio needs something to go between “rockstar” and “ninja”, don’t let me stop you.
There’s a germ of truth in it, of course. Themes at a conference this sprawling can be tricks of the schedule, but a bunch of presentations concerned themselves with how we synthesize datasets, understand large-scale data, and turn distributed, cloud-based data into something that looks and acts like research.
Google Consumer Surveys are the masters of this last. Their pitch – given on Monday in the noisy “Insight Theater” mini-stage – is familiar enough now that you can almost forget how radical it is. It’s not the level of accuracy, the aggressive development and uptake, or even the 10-cent cost per impression that’s most interesting: it’s that Google have cracked how to make surveys “bite-size”. One of their clients is running an 150-question tracker study through Google Consumer Surveys – that’s 150 questions served one at a time to different people. The wisdom of doing an 150-question tracker in the first place is doubtful, but this distributed research is something new, and Google are cunning enough to put the blandest possible skin on it. Their outputs are still resolutely and reassuringly un-flashy, but don’t let that fool you.
However, even Google can’t expect the field to themselves forever, and one of the most interesting ARF presentations came from Pittsburgh’s CivicScience, a start-up which deals in the same kind of distributed polling as Google, but aimed more at Enterprise to Google’s DIY. There’s a key difference in the user experience, too. While Google creates Surveywalls viewers need to get through to reach content, CivicScience’s polls are integrated with publisher content, blending into a news page or blog seamlessly. Millions answer the questions, are served cookies, and CivicScience uses the information for predictive analysis. They offered entertaining examples: people who like Taco Bell, we learned, are 34% more likely to support legalized pot.
The gold, though, is in the less intuitive results. Working for T-Mobile, CivicScience explored the revealed preferences of “persuadable” customers thinking of switching away from AT&T. They isolated the traits which made such people different – from favorite shows and restaurants to attitudes to change and novelty – and then used the same polling tool to test potential offers which might nudge them over the edge.
Just as Google offers radical tech which can be used for Jurassic-era trackers, so CivicScience rested this highly impressive data mining exercise on a very old-school claimed intention question. It’s a theme of the ARF this year, a hidden streak of conservatism amongst the new and shiny – disrupt the methods of data collection, without overturning the assumptions made to draw meaning from that data.
Bite-size, cloud-based surveys are at the leading edge of data synthesis, but for many researchers integrating data sets is still a more artisanal process. Research Now and Warner Music showcased a project they’d done on this year’s Grammy Awards, which mixed survey data, social media, web panel data and sales data. This is the 360 degree consumer view we hear so much about, and it let Warner Music prove the ROI of Grammy nominations and performances very clearly.
Terrific, solid work which any research shop would be rightly proud of – but this was also one of those case studies where there was an unfortunate disconnect between the power of the methodology and the actual insights gained. The key finding is – and I hope you’re sitting down for this – that an appearance on the Grammys leads to more web searches, higher awareness and a sales boost for artists. Alas, useful research is not always exciting.
Can it be made so? That was the theme of the day’s final session – in which researchers got to put on yet another hat, this time taking the role of “storytellers”. The session had its moments, but ended up confirming to me that “storyteller” is often a word that’s used more than actually thought about. SapientNitro, first up, talked about data visualization – yet another type of synthesis, which can be a fantastic tool for storytelling but is not a solution in itself. It didn’t help that the examples were somewhat whiskery – Taylor Swift and Kanye at the Grammys, anyone? – but the main issue was a failure to answer the question that comes after you’ve made your sexy data viz: “Now what?” Gatorade have a war room made up of wall-to-wall dashboards and hot visuals, but as the Q&A confirmed it’s mostly just to titillate the executives – the real work goes on elsewhere. Visualization is only the first stage in the story process.
Charles Kennedy of ABC followed, using visual metaphors to show the power of visual metaphors in storytelling. This was a light, elegant, enjoyable presentation – too skimpy, I thought at first, but someone pointed out to me that at least Kennedy had left his listeners with an actual tip, so it lands in the “win” column. With the power of metaphor dancing in my head, I popped downstairs to hear GfK talk about the promiscuous consumer and the many, many metaphorical kinds of brand relationships people now have – from “flings” to “ex-friends” to “enemies”.
Oddly, in this panoply of human-brand relationships, the only metaphor missing was the type of human intimacy where money changes hands. Which reminded me of the lesson of the insight alchemist – metaphor is great, but be careful it doesn’t glamorize or obscure reality instead of illuminating it.