This is a story about insights, specifically about augmenting qualitative insights by adding a layer of social media on top. View this presentation to see how you can validate qualitative research insights on a mass scale using social media analysis.
Regression analysis: Simple Linear Regression Multiple Linear Regression
Augmenting Brands
1. Augmenting Brands with Real Time Data
Philip McNaughton, Face
A case-study about
using social media
data to scale up
qualitative insights
2. You can’t turn data into a story without joining it with other data
- Flip Kromer, founder and CTO Infochimps
3. This is a story about joining qualitative and quantitative data together
4. Where Social Media But we can
Research is now…. do more….
Monitoring
Campaign Tracking Validating
insights
Topic Buzz Scaling up
qualitative
Mining Insights learning
On a mass scale
More specifically it’s a story about the power of scaling up
qualitative insights with Social Media
5. A client came to us with insights developed
through qualitative research, and asked us to
validate, sharpen and prioritize.
6. We first took a traditional approach: a
survey > participants respond to
insights and tell us what they think
7. Great, but some limiting factors…
Artificial
Evaluative responses
Still limited sample size
No depth
8. This time, we wanted to augment this data with something a little different
9. We wanted to see whether the insights played out organically in the real
world, not in the artificial world of the survey and the focus group
10. We set out to see how
Social Media could
augment and validate
the learning on a
mass scale
11. Using our social media research tool Pulsar to pull in data from blogs,
forums, videos, social networks.
12. The Big Challenge; if we can’t ask a
specific question of social media users,
how can we ‘find’ a specific answer?
13. Starts by creating search terms that look very broadly at the
insights, their categories, and behavior around them.
14. Create a clean data set with non-original consumer content removed –
5000 pieces of content
15. Home as self- Home as Home as Home as
expression welcoming place flexible & versatile showing-off
Apply a code frame to each one of the 5000 relevant pieces of content,
matching each against each of the insights…
16. Obtain the relative size
of each insight’s
foundation in real time,
organic data
18. 1. Where it’s discussed
Qualitative Insight on a 2. How it’s discussed
mass scale – and we 3. Sentiment
never even had to ask 4. What categories are discussed
5. Relative presence of a category
19. Match our social insights
back with the quantitative
data, to create a 360 degree
data perspective
20. A word to the wise…
Qualitative Insight on a
mass scale – and we
never even had to ask Depends on penetration of SM
Data must be cleaned
Human process – labor intensive
SM is organic, but not the whole story
Still needs other data for control
21. 1. Scale up qualitative insights
Qualitative Insight on a
2. Mass organic qualitative insight field
mass scale – and we
never even had to ask 3. Cost effective validation
4. Dig inside every data point for depth
5. Dynamically track insights over time
Thank you. This is a story about using Social Media in Research, and of course we have heard a lot about that over the last couple of days. Ricardo from Mastercard yesterday talked about the journey from Questioning to ListeningToday I am going to talk about how questioning and listening can work together.
Aspractioners of social media research, its sometimes easy to be seduced into thinking that it has all the answers. Of course, that is very far from the truth. My thinking is that social media research works best of all when it is used in synthesis with other data sources. Not just as an add-on, but in a more directly integrated way with other research data – qual and quant.
Lots of things that social media is good for. But we can do more….There’s a lot of talk in the industry about social media research being a meeting point for qualitative and quantitative research.And at face we have been exploring one way of working at that intersection.
RB were feeling pretty comfortable with these insights, but they had been developed in a small scale way, and they wanted to find a way of validating, sharpening and prioritising
So we took a 2 step approach to this. Firstly, a relatively traditional survey approach.
Of course, this is always the challenge with social media. It inverts the traditional research process of question and answer. The answers are there, but how to filter and question the data in the way which finds the answers you need. This was particularly magnified in this case because we were looking for something very specific.
Of course, social media is primarily about filtering and questioning a potentially endless data field in the right way. The first thing was to narrow down a potential data set that we could search in.Focus on change and disatisfaction with the home environment
But this still leaves us with a huge amount of content that we know is roughly in the right area. How do we go about questioning this data set to validate our 4 insights…
And it actually a very traditional MR method > coding open-ended data.
Ultimately, what we were shooting for…. !
It’s great to be able to do that. But of course we can do even more….
Final point is the most interesting here. And it applies to social media not just in insight validation, but across segmentaion, comms tracking. So if I was to leave you with one bigger though about the benefit and relevance of social media research it’s that, the fact that it offers us a real-time, truly dynamic insight field. We just need to be as good as we can be at learning how to question it!