Building bridges with the machines: the future of qual

The chasm between humans and machines is rapidly growing closer at a speed that most in the insight industry don’t appreciate.

Indeed, with more and more of our lives lived online, and data capture, analysis and reporting becoming more automated and ‘intelligent’, machines will get increasingly better at analysing our digital selves, what this means for businesses, and predicting what we will want in the future.

It’s easy to see the impact of machine learning on quantitative insight as businesses invest more in automation and artificial intelligence to understand the data they have on their customers. Just take a peek at the latest GRIT report to see all the buzz around machine learning, automation, and other variants of AI.

But what’s the place of qualitative insight within this tectonic shift?

The challenge

A standard narrative runs that qual is in a good place to add value, as all this quantitative data needs contextualising, people need actually observing and stories need telling. This is evident in the current popularity of UX research and ethnography.

In reality, ‘traditional’ qualitative insight is under threat if it doesn’t adapt in the next 5 years. We need to add value. This means we need to build bridges to the machines, cross that divide and work together. Those that don’t will face redundancy.

The evidence

Don’t believe me? With perfect voice-to-text transcription imminent (see my fav app, here), it’s only a matter of time until machines analyse mountains of transcripts, dig around for patterns and then spit out a report in a pre-written template. Indeed. this is already happening to a certain extent with various AI-driven platforms in the market.

The classic focus group seemingly hangs around based on the premise that robots can’t interact with people in an intuitive or meaningful way. However, robots will soon be able to do the classic qual moderator role in a good enough way.

Consider the following:

  • We know most people are used to using apps and interacting with others on their mobile phone – often to communicate in an abbreviated form (WhatsApp, TikTok, Snapchat)
  • Again, going back to the GRIT report, it shows a great deal of interest in chat-bot based surveys from clients and suppliers
  • We know online qual communities are popular, cheap, and effective
  • And we know that many questions – even in qual – can be standardised

 

So it’s actually not far-fetched to see chat-bots and then more sophisticated AI variants taking on more and more qual. It would be in highly structured, online and one-to-one situations at first, and then it would branch out into increasingly more complex scenarios. Given that focus groups themselves are entirely unnatural — despite our best efforts — it’s reasonable to assume our participants would adapt over the long run.

Building bridges and adding value with qual

Qualitative insight is going to have to be agile and play to its strengths to stay relevant in the medium-to-long term.

At Relish, we are prioritising the following areas to future-proof our business and add value for clients:

  • Good insight design. As machines do more of the work, humans are going to be needed to 1) programme the insight design and 2) verify and build off outputs produced by machines.
  • Observational approaches (mobile and ethno). Machines will struggle to get up close to the moments of interest and capture nuanced or unpredictable behaviour.
  • Bigger picture cultural analysis and sharp semiotic thinking. Understanding what is happening in culture and design and how this works to create meaning in a given time will be sorely needed to complement reported consumer data.
  • Creative thinking approaches. Methods that do not generate any new primary insight, but use existing primary and secondary insight as a starting point, and use qualitative thinking technique to push past what we currently know and into new ideas.
  • Curation of insight. As the amount of data clients have access to increases, there is value in showing them their own data in new ways to encourage engagement, prompt thinking and help draw new connections.
  • Being iterative, messy and playful. No matter how good machines get at approximating human interaction, they will struggle to get a handle on elements we know are important for getting the most from participants. Things like: harnessing playful activities, humour, and accepting that we don’t get to ‘the answer’ in a logical straight line all the time – especially via conversation.
  • Using our senses and building things. Engaging our participants with all their 9 senses is important to elicit meaningful responses, as is using projective activities and being ‘hands on’ in allowing participants to build and create. Machines are likely to struggle across these areas for a good while yet.
  • Creating first-class customer experiences or user journeys. The make-up of a good customer experience or a seamless user journey is complex. Being able to evaluate these, draw on cross-category examples, and pull apart the nuance is likely to remain in the classic qual domain.
  • Telling stories and landing insight with impact. A lot of what we do as qualitative insight practitioners is to tell stories. This will remain important as a way to achieve business impact. Also, using a bit of flair and theatre to captivate, entertain and make insights sticky will remain big value-adds.

 

The future is an interesting one. Smart agencies will remain nimble, become more consultative, and harness technology to provide more valuable insight for their clients.

Relish are up for the challenge and we’re looking forward to building bridges and exploring them with our clients.

Neil