Wow, thank you so much. I just discovered your channel and am simply blown away by it. I've read (and watched) a lot of educational content on UX, but it seems that noone else manages to give such clear, concise and actionable advice. (Well, maybe Erika Hall is close :). Looking forward to reading your book soon!
Hi David! This is super helpful for me. I have just conducted seven interviews and was trying to identify the segments/patterns. While my supervisor tried to lead me towards targeting them as "students "and "knowledge workers," I always felt something was not right. However, I could not find a better way to describe the segments I saw. Your method of shuffling and comparing cards seems a good step ahead for me. Thank you for sharing.
Hi David! Very helpful and timely information. I have the same situation as the person who asked the question. I interviewed two types of users, individual consumers and businesses. Do I look at the similarities and differences between users in the two different groups, or split them out as if it were two different studies? Thank you!!
Segment by needs, definitely. Segmenting by pain points I'm less sure about. I'd want to understand what was behind the pain point, which is usually a behaviour or a need.
Very interesting. Thank you. If you were to try to assess the business value of each segment or simply try to identify the % that each segment represent, you would apply a quantitative research methodology after the initial segmentation phases ? Also, Would that be the way to identify you a main or primary segment?
Hi Sam, I have read your question in the comments and I would like to understand it properly since I am also interested in the topic. But I would say that what you are asking has been already done (correct me if I am wrong please). Because in this case, I understand personas come from quantitative research already done. Tyler already has lots of data, and now needs to refine by the properties worth for the product, and create persona definitions (to continue with journeys, stories, etc). In my opinion quantitative is done mainly before product development and after product development for metrics. During product development qualitative should be done for task definition and refining, and also for iterative testing. Again, please correct me if I am wrong. Thanks.
@@popejonas178 Pope Jonas hi, thank you for your message. I am sorry, I wasn't very clear. To me, personas And / or segments need to be, as recommended in the video, be behavior based. And I think you need qualitative to identify behavior. Once you have the variables or the dimensions identified and the segments created, my question was, would the next step be quantitative studies to evaluate the % each segment represents and measure the value (LTV for ex.) For each segment. I think qualitative and quantitative research are often required / useful to get a better picture and lead to a better defined problem, focus on the most business impacting problem and also in evaluating the solutions. Would love to hear David Travis thoughts on this.
When I've needed to get quantitative estimates of the the size of these segments in the past, I've used a survey. I describe the personas and then ask my respondents to self-identify with one of the personas (or choose "none"). If you do it this way, it's very important to make sure you have a random sample of participants and that you chase up non-responders.
Thanks for another interesting tip! Do these behavior-based segments ever clash with the segments created by the business/marketing department? And if so, how do you reconcile them?
It depends: it's not unusual for this to be the first time that anyone has thought about segmenting users. When they do conflict, I point out that UX segmentation has a different purpose to marketing segmentation: it's about understanding usage not influencing purchase.
David, thank you this is very insightful! How do you make sure all the participants are compared to each other? If I have 40 participants in my persona research, how do I “pick my participants a couple of times”? How do I do that? 3:33
Just pick a couple of participants randomly (say participants 11 and 32) anf then ask: what do they have in common that makes them different from (say) participant 6? Rinse and repeat until you have a set of characteristics on which you can grade your whole sample. (40 people is a very healthy sample for qualitative research. Kudos!)
@@DavidTravis Thanks David! I guess what I'm confused about is are we comparing every single participant to each other in this compare 2 to 1 format? If so, how do I do that with a sample size of 40? Appreciate you taking the time to respond.
Hi David, how would you go about setting up screeners from a testing service like UserTesting? I am setting up unmoderated tests like a click test and have screeners for gender, education, salary, etc. Can your method be used to identify patterns in behavior for a particular state (US) in order to get the best representation of that state? Thanks for your time!
@@DavidTravis What if we are setting up screeners for an in-depth interview? Can this method be used for defining behaviors in screeners for interviews? Also, how do we use this information to decide sample size for interview.
Wow, thank you so much. I just discovered your channel and am simply blown away by it. I've read (and watched) a lot of educational content on UX, but it seems that noone else manages to give such clear, concise and actionable advice. (Well, maybe Erika Hall is close :). Looking forward to reading your book soon!
I'm glad it's helpful, thanks for the comment. 😁
Hi! David, thank you for sharing your approach. It's very useful.
Glad it was helpful!
Hi David! This is super helpful for me. I have just conducted seven interviews and was trying to identify the segments/patterns. While my supervisor tried to lead me towards targeting them as "students "and "knowledge workers," I always felt something was not right. However, I could not find a better way to describe the segments I saw. Your method of shuffling and comparing cards seems a good step ahead for me. Thank you for sharing.
I'm glad it was useful.
Very well explained. Thanks Sir.
You're welcome. 😊
Helpful as always. Thanks, David !
Glad it was helpful.
Hi David! Very helpful and timely information. I have the same situation as the person who asked the question. I interviewed two types of users, individual consumers and businesses. Do I look at the similarities and differences between users in the two different groups, or split them out as if it were two different studies? Thank you!!
If you know these are different categories of user, then I'd treat them as separate studies.
Thanks for this David, quick question, couldn’t needs or pain-points be used in lieu of behaviors? What do you think?
Segment by needs, definitely. Segmenting by pain points I'm less sure about. I'd want to understand what was behind the pain point, which is usually a behaviour or a need.
Very interesting. Thank you. If you were to try to assess the business value of each segment or simply try to identify the % that each segment represent, you would apply a quantitative research methodology after the initial segmentation phases ? Also, Would that be the way to identify you a main or primary segment?
Hi Sam, I have read your question in the comments and I would like to understand it properly since I am also interested in the topic.
But I would say that what you are asking has been already done (correct me if I am wrong please). Because in this case, I understand personas come from quantitative research already done.
Tyler already has lots of data, and now needs to refine by the properties worth for the product, and create persona definitions (to continue with journeys, stories, etc).
In my opinion quantitative is done mainly before product development and after product development for metrics. During product development qualitative should be done for task definition and refining, and also for iterative testing.
Again, please correct me if I am wrong.
Thanks.
@@popejonas178 Pope Jonas hi, thank you for your message. I am sorry, I wasn't very clear. To me, personas And / or segments need to be, as recommended in the video, be behavior based. And I think you need qualitative to identify behavior. Once you have the variables or the dimensions identified and the segments created, my question was, would the next step be quantitative studies to evaluate the % each segment represents and measure the value (LTV for ex.) For each segment. I think qualitative and quantitative research are often required / useful to get a better picture and lead to a better defined problem, focus on the most business impacting problem and also in evaluating the solutions. Would love to hear David Travis thoughts on this.
When I've needed to get quantitative estimates of the the size of these segments in the past, I've used a survey. I describe the personas and then ask my respondents to self-identify with one of the personas (or choose "none"). If you do it this way, it's very important to make sure you have a random sample of participants and that you chase up non-responders.
Thanks for another interesting tip! Do these behavior-based segments ever clash with the segments created by the business/marketing department? And if so, how do you reconcile them?
It depends: it's not unusual for this to be the first time that anyone has thought about segmenting users. When they do conflict, I point out that UX segmentation has a different purpose to marketing segmentation: it's about understanding usage not influencing purchase.
David, thank you this is very insightful! How do you make sure all the participants are compared to each other? If I have 40 participants in my persona research, how do I “pick my participants a couple of times”? How do I do that? 3:33
Just pick a couple of participants randomly (say participants 11 and 32) anf then ask: what do they have in common that makes them different from (say) participant 6? Rinse and repeat until you have a set of characteristics on which you can grade your whole sample. (40 people is a very healthy sample for qualitative research. Kudos!)
@@DavidTravis Thanks David! I guess what I'm confused about is are we comparing every single participant to each other in this compare 2 to 1 format? If so, how do I do that with a sample size of 40? Appreciate you taking the time to respond.
Hi David, how would you go about setting up screeners from a testing service like UserTesting? I am setting up unmoderated tests like a click test and have screeners for gender, education, salary, etc. Can your method be used to identify patterns in behavior for a particular state (US) in order to get the best representation of that state? Thanks for your time!
I wouldn't aim for representativeness in a usability test. Here's why: userfocus.co.uk/articles/myth-of-the-representative-sample.html
David Travis thank you!
@@DavidTravis What if we are setting up screeners for an in-depth interview? Can this method be used for defining behaviors in screeners for interviews? Also, how do we use this information to decide sample size for interview.