@jtimed asked us on Twitter the following question:
It’s getting easier and easier to collect photos, video and GPS (heck – even barcode capture will be mainstream by 2012) as consumers become more savvy with smartphone technology and handset sales continue to increase. This will only continue to rise, as will the adoption of tablet devices built on Google’s Android and Apple’s iPad platform.
We don’t have all the answers – but we want to find them with you.
Industry trend watchers continue to talk about what to do with the large volume of qualitative data that is being captured (not to mention the space it requires to store them all) and it’s an issue that industry leaders and innovative market researchers are continuing to advance. The key point is that the information provided in photos, video, GPS and barcodes (and other forthcoming advances) in real time is valuable in offering richer insights into the consumer’s experience which is valuable to clients and researchers alike.
We can offer market researchers the opportunity to provide a holistic view of the consumer experience validated in real time as they take their smartphone from home to work, to the supermarket, cinema, restaurant and back home again. If your clients knew this was available for diary studies, mobile mystery shopping, CAMI and mobile panel, wouldn’t they ask for it?
With apps-based research that is appropriately designed to answer the research question, for the audience, for the device and for how the respondents use it, it is possible to conduct qualitative research on a quantitative scale. We want to work with you to make this happen.
@jtimed and others in our community – we’d love to hear your thoughts on this industry challenge.
It’s a thought that has been echoed by market research professionals we met at NetGain 5.0 and the CASRO Online Conference in Vegas. It’s definitely a challenge, and we are only scratching the surface, but we’ve started to find some success by looking outside of MR for solutions (necessity being the mother of invention, and all...).
Firstly, photo and video data is qualitative in nature so you can’t get around coding it in some fashion to get the most out of what you’ve collected.
Firstly, photo and video data is qualitative in nature so you can’t get around coding it in some fashion to get the most out of what you’ve collected.
Even in a very open unstructured research design, we often include a few quant type questions which helps respondents to code the content of the photo or video for us according to pre-defined options. Or you can quantify a qualitative preference like below.
Of course, this approach doesn’t capture everything, so you then might go to a coding specialist software and service like ‘Ascribe’ by Language Logic. They innovated in text coding, but have applied efficient time saving code book techniques to multimedia coding, too. Still, this is a time-intensive process...
So back to the quantifying of your qualitative data... Microsoft provides MS Pivot, which we’ve used to pull in metadata (i.e. survey answers) associated with photographic data – then you can start to do some dynamic sorting and get a feel for the quant story behind the mountains of qual data you’ve collected. We first explored this with Ipsos last May on our Great British Weekend (MCASI) project.
So back to the quantifying of your qualitative data... Microsoft provides MS Pivot, which we’ve used to pull in metadata (i.e. survey answers) associated with photographic data – then you can start to do some dynamic sorting and get a feel for the quant story behind the mountains of qual data you’ve collected. We first explored this with Ipsos last May on our Great British Weekend (MCASI) project.
We asked respondents what they were doing, who they were with, and to take a photo and GPS capture of where they were over the course of a day. Pivot allowed us to filter the images by response data such as:
• Location
• Family
• Friends
• Food
• Media
And other activities that people were engaging in.
It’s a dynamic sort and looks like this (sorry for the lack of sound - we didn't have time to make it look very polished - we just wanted to respond to you!):
When you look around to other tools like Google Maps, you can start to create other insightful views like what follows:
From here you can see stories evolve from the visual trends such as people watching media, feeling relaxed, eating with family, being excited about being out with friends etc..
It’s getting easier and easier to collect photos, video and GPS (heck – even barcode capture will be mainstream by 2012) as consumers become more savvy with smartphone technology and handset sales continue to increase. This will only continue to rise, as will the adoption of tablet devices built on Google’s Android and Apple’s iPad platform.
We don’t have all the answers – but we want to find them with you.
Industry trend watchers continue to talk about what to do with the large volume of qualitative data that is being captured (not to mention the space it requires to store them all) and it’s an issue that industry leaders and innovative market researchers are continuing to advance. The key point is that the information provided in photos, video, GPS and barcodes (and other forthcoming advances) in real time is valuable in offering richer insights into the consumer’s experience which is valuable to clients and researchers alike.
We can offer market researchers the opportunity to provide a holistic view of the consumer experience validated in real time as they take their smartphone from home to work, to the supermarket, cinema, restaurant and back home again. If your clients knew this was available for diary studies, mobile mystery shopping, CAMI and mobile panel, wouldn’t they ask for it?
With apps-based research that is appropriately designed to answer the research question, for the audience, for the device and for how the respondents use it, it is possible to conduct qualitative research on a quantitative scale. We want to work with you to make this happen.
@jtimed and others in our community – we’d love to hear your thoughts on this industry challenge.
Powered by Compendium
You see, we can now do this with unstructured qualitative text. We can categorise millions of social media posts using the words that people use. These developments in text mining and analytics lead me to conclude that we should be able to process photos and videos in a similar way. It is great that we can use tools such as PowerPivot to sort photos and videos using quant type questions, textual comments, GPS location etc. However, just like Google Images, this is grouping together photos and videos using structured tags. As you mention, we have to code the data in some way, and in the future we will be able to do this using the pictorial content itself.
Wouldn’t it be great if we use object recognition to group similar photos and videos together, independently of any existing metadata? Developments in speech recognition and text analytics will allow us to categorise videos based upon what people say within them. Experts in machine learning and computer vision will find ways to cluster images, based upon the visual similarities within them. And developments in quantitative and visual semiotics will allow us to understand the semantic relationships contained within images, applying structure to automate the analysis. These developments may take time, but the future has a habit of arriving sooner than we expect.
We now have the opportunity to collect photos and videos on a vast scale, and we can currently organise them using external tags. Developments in text analytics mean that we can now automatically analyse social media posts using the words they contain. In the future, we will also be able to automatically analyse photos and videos using the pictorial content within the images they contain. This will dramatically reduce the time and costs involved in using photos and videos in research, allowing anyone to conduct research this way. Finally, it may even become a mainstream alternative to survey-based research, but that is a thought for another day!
I suspect the technology will be developed for other more commercial purposes first - such as voice, picture or video driven search. Are you listening, Bing? ;)
How do you automatically analyse social media posts using words that are ambiguous, especially in particular segments where casual interactions can subvert traditional interpretations of dialogue?
I don't think using photos and videos in research is expensive at all. In fact, asking someone to press 'record' on their own device and tell us what they think is much cheaper than recording a focus group, or a personal interview and you can ask them to pre-sort the category their content falls into before you even know that they've completed the survey. It all depends on the research question. There's a large scale mobile digital ethnography project I will tell you about some time soon that meant we were conveniently with the respondents at all times, pulled out of their pockets & purses and communicated with in their "dead time" waiting, watching tv etc and the content they shared with us (which we digitally filtered to notice themes) was more valuable than through any other method (unless they each had a tv crew filming their every move). You're right in that the next step is to, as an industry, develop an "ordered reference standard" where the qual data or the validated qual content can be reliably structured. We know the magnitude and the enthusiasm is there to make this quant -worthy!