Quant Equals Qual
3 likes674 views
The document discusses the importance of integrating quantitative and qualitative data in UX research, emphasizing the need for faster, more relevant analysis to inform decision-making. It critiques the current state of UX research as slow and costly, leading to missed insights on customer behavior. Key points include the necessity of combining different data sources and the challenges posed by data access and privacy regulations.
Quant Equals Qual
- 1. August 2015 UCD CONFERENCE - HUMANITY IN DIGITAL LANDSCAPES @LolaOye Quant = Qual Why we should all love data
- 2. ‣ My background is primarily Qual. ‣ Depending on your leaning (Qual or Quant), you may feel my points are unfair. ‣ I’m more interested in what’s next, than what was. As you listen, bear in mind: 2
- 3. 3 “I assumed that the time would come when there would be a science in which things could be predicted on a probabilistic or statistical basis. Isaac Asimov, Author of the Foundation Series, inventor of Psychohistory
- 4. Quantitative 4 /ˈkwɒntɪˌtətɪv,-ˌteɪtɪv/ Relating to, measuring, or measured by the quantity of something rather than its quality. In UX Research measured by: • Numbers • Amounts • Trends • Increments • Statistics We treat quantitative data as inherently summative. It lack’s the nuance that experiences are built on. Qualitative /ˈkwɒntɪˌtətɪv,-ˌteɪtɪv/ Relating to, measuring, or measured by the quality of something rather than its quantity. In UX Research measured by: • Narratives • Stories • Superlatives • Inferences We treat qualitative data as both formative and summative. But it is open to bias and slow to analyse.
- 5. 5 UX research is losing relevance. It takes too long and is too expensive. Too many “UX people” don’t have research skills. We no longer have the right skills for the emergent business & technology context
- 6. 6 “Given two or three data points, our minds can construct an alternate reality in which all of those data points make flawless sense. Five UX Research Pitfalls, UXMagazine. Elaine Wherry, 2010
- 7. Indication not Inference Causation not Correlation
- 8. “Gary King, Harvard University, Director of Institute for Quantitative Social Science, 2013 Big Data is not about the data. 8
- 9. 9 We have a lot of quantitative data that we need to analyse better. We have a lot of qualitative data and research but we can’t harness these insights in real time. Our quantitative data and qualitative data are not joined up. We can’t keep waiting 3-6 weeks for data analysis. UX & Product people need to stop burning budget on research! We need big data. We got a Hadoop, please bring instructions.
- 10. HEY, WHAT CAN YOU TELL ME ABOUT OUR CUSTOMERS’ BEHAVIOUR? WELL WHAT DO YOU WANT TO KNOW?
- 11. I WANT TO BE OPEN, WHAT CAN YOU TELL ME? LOTS. YOU NEED TO BE MORE SPECIFIC.
- 12. OK, WHAT IS OUR CUSTOMERS MOBILE BEHAVIOUR? DO YOU WANT EVERYTHING?
- 13. ERM, YEAH…? OK, THAT’S GOING TO TAKE 6 WEEKS. EXCEL FILE OK?
- 14. ARE YOU SERIOUS OR JOKING? I CAN’T TELL. I’M NOT JOKING.
- 15. ‣ They didn’t know how to get the data they wanted, how it was stored or how we would be able to use it ‣ The ‘user’ was us…UX & product folks who need to make informed decisions to prioritise services and features ‣ We’d never built a big data system before, so we had to learn quickly! This brief had spiky bits: 15
- 16. 16 User Experience Architect Statistician/Data Scientist Back-End Developer Front-End Developer UI Designer “Unicorn” Developer Client Bringer of whisky.
- 17. Nobody really knew what they could get. So they didn’t know what they could ask for.
- 18. You have to start with qualitative questions you’ve always wanted to answer. There will be huge gaps in insight because some databases don’t play nice. The Data Protection Act. It sucks. 1. 2. 3.
- 19. What the business has always wanted to know, but never knew it could get: 19 PERFORMANCE • App downloads / app usage / app feedback (1view) • Activity in digital products over time (spot trends, dips and peaks) • Customer activity (frequency) across channels • Cross product offering and cross channel purchase behaviour (360 view) • Touchpoint usage: feature use, feedback on features, drop off points and completed journeys • Segments x value earned by channel EXPERIENCE • Customer’s comments over a determined period of time • Behaviour across channels over a period of time Use data to segment users by behaviour, not spend • Impact of launches or push notifications on behaviour over time • Movement between transactional segments: crossing demographics with purchase value/frequency …..and lots more.
- 20. POS Website Clickstream On-Site Customer Reviews Loyalty Card External Product Reviews Daily Sentiment Analysis Social Media Search Data App Analytics Postcode Lookup Dozens of UX Insights DATA SOURCES
- 21. 21 Red: Data sources we don’t have access to or don’t know how to access, or for which the source itself is unintelligible to us or totally unknown. Amber: Data sources we have access to but which have problems that severely limit their use. Green: We have access to these sources today and the data is clean. 53% 10% 37% 30 Data Sources | RealTime | Periodic | Historical
- 22. 22 Their quant stuff we could plug in Qualitative (Unstructured) Data The bits we made…all re-usable and open! ACIXOM
- 23. 23 REAL TIME SALES SITE CATALYST HISTORICAL SALES AXIOM DEMOGRAPHICS TWITTER APP FIGURES OPINION LABS
- 24. 24 Can I see how many vouchers have been redeemed this week and how that compared to last week? REAL TIME SALES SITE CATALYST HISTORICAL SALES AXIOM DEMOGRAPHICS TWITTER APP FIGURES OPINION LABS
- 25. 25 Can I monitor the impact of the app launch on customer’s voucher redemption behaviour? REAL TIME SALES SITE CATALYST HISTORICAL SALES AXIOM DEMOGRAPHICS TWITTER APP FIGURES OPINION LABS Can I monitor the impact of the app launch on customer’s voucher redemption behaviour?
- 26. Can I monitor the impact of the app launch on customer’s voucher redemption behaviour? 26 REAL TIME SALES SITE CATALYST HISTORICAL SALES AXIOM DEMOGRAPHICS TWITTER APP FIGURES OPINION LABS Can I monitor engagement across different channels in relation to feature releases and I can I overlay that with channel specific sentiment? Can I monitor the impact of the app launch on customer’s voucher redemption behaviour?
- 27. 27 Can I monitor app downloads over time, find out who is using them and what the overall sentiment is about the app? REAL TIME SALES SITE CATALYST HISTORICAL SALES AXIOM DEMOGRAPHICS TWITTER APP FIGURES OPINION LABS Can I monitor the impact of the app launch on customer’s voucher redemption behaviour? Can I monitor engagement across different channels in relation to feature releases and I can I overlay that with channel specific sentiment? Can I monitor the impact of the app launch on customer’s voucher redemption behaviour?
- 28. 28
- 29. We’re not allowed to know everything we would like to know. Get over it.
- 30. It’s really difficult to separate interesting behavioural data from “Personally Identifiable Information”. 30 THE DATA ITSELF • Geography vs location • Individual & household vs personalisation • MAC addresses THE ANALYSIS • Sentiment + Sales by geography • Usage + Ratings + Sales by product ….the more you cross and the more ‘accurate’ it is, the closer you get to effectively breaking the Data Protection Act.
- 31. QUANT QUAL
- 32. 32 “The secret of genius is to carry the spirit of the child into old age, which means never losing your enthusiasm. Aldous Huxley, Author & Philosopher