Qualtrics Text IQ Training

00:37:50
https://www.youtube.com/watch?v=jA4E5MD0YEI

الملخص

TLDRThe video presents a training session on Qualtrics' Text iQ, led by Karen Sin. It covers the significance of text analysis in understanding stakeholder feedback, the functionalities of Text iQ, and its evolution from heuristic algorithms to machine learning models. Karen explains the process of using Text iQ, including sentiment analysis and topic detection, and emphasizes best practices for collecting qualitative data. The session includes a live demonstration of the Text iQ tool, showcasing how to navigate its features and analyze data effectively, ultimately helping organizations make data-driven decisions based on qualitative insights.

الوجبات الجاهزة

  • 📊 Text iQ helps analyze open-ended feedback effectively.
  • 🤖 It uses machine learning for real-time improvements.
  • 📝 Best practices suggest limiting open-ended questions to three.
  • 📈 Organizations can track trends over time with Text iQ.
  • 🔍 Sentiment analysis provides insights into stakeholder feelings.
  • 💡 Text iQ automates the categorization of qualitative data.
  • 📚 Resources for training are available on the XM Basecamp website.
  • ⚙️ Manual adjustments can enhance sentiment accuracy.
  • 📉 Text iQ eliminates the need for manual tagging of comments.
  • 🔗 Data can potentially be exported to other analysis tools.

الجدول الزمني

  • 00:00:00 - 00:05:00

    The training session on Text iQ by Qualtrics is introduced by Karen Sin, who emphasizes the importance of the platform in enhancing user experience and feedback analysis. She outlines the agenda, which includes an overview of Text iQ, a demo, and resources for further learning.

  • 00:05:00 - 00:10:00

    Karen explains text analysis as a method to gather feedback from stakeholders, highlighting sentiment analysis and topic detection as key techniques. Text iQ is introduced as a feature that automates the process of extracting insights from open-ended feedback, making it easier for organizations to understand customer and employee experiences.

  • 00:10:00 - 00:15:00

    The benefits of Text iQ are discussed, including its ability to uncover insights quickly, eliminate manual labor in analyzing comments, and classify sentiment automatically. Karen stresses the importance of collecting quality data and suggests limiting open-ended questions to optimize responses.

  • 00:15:00 - 00:20:00

    Karen outlines the process of performing text analysis with Text iQ, which includes getting the right comments, creating text topics, using sentiment analysis, and taking action on insights. She provides an example of how to formulate open-ended questions to gather meaningful feedback.

  • 00:20:00 - 00:25:00

    The demonstration begins with accessing and navigating Text iQ. Karen explains the different modes available in the Text iQ pane, including editing and analysis modes, and highlights key components such as fields, topics, and response lists.

  • 00:25:00 - 00:30:00

    Karen continues the demo by showing how to add questions with open-text responses and create text topics based on those responses. She explains the iterative process of tagging comments and using topic recommendations to streamline the analysis.

  • 00:30:00 - 00:37:50

    The importance of taking action based on insights from Text iQ is emphasized, as it helps organizations identify best practices and areas for improvement. Karen discusses how to integrate Text iQ data into larger dashboards for comprehensive analysis.

اعرض المزيد

الخريطة الذهنية

فيديو أسئلة وأجوبة

  • What is Text iQ?

    Text iQ is a feature developed by Qualtrics to analyze open-ended feedback, helping users understand both the 'what' and the 'why' behind qualitative data.

  • How does Text iQ improve text analysis?

    Text iQ uses machine learning models that self-improve in real time, making it more accurate and actionable compared to traditional heuristic algorithms.

  • What are the key components of performing text analysis with Text iQ?

    The key components include getting the right comments, creating text topics, using sentiment analysis, and taking action on text insights.

  • Can Text iQ analyze sentiment in open text responses?

    Yes, Text iQ automatically classifies sentiment, providing a macro-level understanding of how respondents feel about certain topics.

  • What are best practices for using open-ended questions in surveys?

    It's recommended to use a maximum of three open-ended questions to optimize responses and minimize drop-off rates.

  • How can organizations take action based on insights from Text iQ?

    Organizations can mobilize teams to address key findings, leverage best practices, and prioritize pressing issues based on the insights gathered.

  • Is it possible to export Text iQ analysis to other tools?

    Yes, Text iQ analysis can potentially be exported to other analysis tools like Tableau, but confirmation is needed.

  • How can I access training resources for Text iQ?

    Qualtrics users can access on-demand training, webinars, and support pages through the XM Basecamp website.

  • What should I do if I notice inaccuracies in sentiment analysis?

    You can manually adjust sentiment scores in the Text iQ tool to ensure more accurate representation of feedback.

  • Can Text iQ track trends over multiple years?

    Yes, Text iQ allows for trend analysis over time, enabling users to filter and visualize data across different periods.

عرض المزيد من ملخصات الفيديو

احصل على وصول فوري إلى ملخصات فيديو YouTube المجانية المدعومة بالذكاء الاصطناعي!
الترجمات
en
التمرير التلقائي:
  • 00:00:02
    >> All right, so, good afternoon everyone.
  • 00:00:06
    Welcome to the Text iQ training offered by OCE.
  • 00:00:10
    Thank you for joining us.
  • 00:00:13
    We have Karen Sin who is a representative of Qualtrics, who will be presenting to us today.
  • 00:00:19
    And Karen, if you don't mind just sharing a little sharing a little info about yourself,
  • 00:00:23
    and then go right ahead and get started.
  • 00:00:25
    >> Okay, perfect.
  • 00:00:26
    Thanks, everyone for having me.
  • 00:00:27
    So, good afternoon everyone.
  • 00:00:29
    I want to quickly introduce myself, my name is Karen Sin, and I will be your very engaging
  • 00:00:33
    and exciting trainer for the day.
  • 00:00:36
    I am the head of Customer Success in the Federal Practice at Qualtrics, and in that role,
  • 00:00:40
    I'm really just dedicated to ensuring
  • 00:00:43
    that your team gets the best use and value out of our platform.
  • 00:00:47
    So, I am very excited to be here today to provide an overview of one
  • 00:00:52
    of our powerful features, Text iQ, and hope that you can get meaningful use out of it.
  • 00:00:56
    Timing couldn't be more perfect for this demo.
  • 00:00:59
    Our Qualtrics platform just launched a new iteration in TextIQ
  • 00:01:03
    that incorporates some strategic updates here
  • 00:01:06
    and makes it more accurate, actionable, and agile.
  • 00:01:09
    So, just as an update, the feature continues to evolve from a heuristic-based algorithm
  • 00:01:14
    to true machine learning models that really self improve in real time and continues
  • 00:01:20
    to be increasingly more and more accurate.
  • 00:01:22
    So, we have about 90 minutes allocated to this training.
  • 00:01:26
    I don't anticipate we'll need to take all that time,
  • 00:01:29
    but it will be perfect to leave room for questions.
  • 00:01:31
    So, just to give you a quick summary of the agenda for today's training.
  • 00:01:35
    I'll start by going through a brief introduction just around Text iQ and text analysis
  • 00:01:40
    in general, do a full walk-through, or demo, around how to use the feature,
  • 00:01:44
    and then point you to some key resources that you can use when you start to go
  • 00:01:49
    through this process on your own.
  • 00:01:50
    So, just also note that within this deck, I've also included a list of best practices for you
  • 00:01:57
    to also reference when you're going throught the process.
  • 00:02:01
    Okay, so, I'll go ahead and get started.
  • 00:02:02
    I just want to make sure everyone can see my screen okay.
  • 00:02:08
    >> We can, Karen.
  • 00:02:09
    >> Okay, perfect, and we'll just start with the introduction here on text analysis and Text iQ.
  • 00:02:16
    Okay, to start, I just want to take a moment to explain what text analysis is.
  • 00:02:20
    So, text analysis is, arguably, one of the best ways to collect feedback from stakeholders,
  • 00:02:26
    whether they're customers or employees.
  • 00:02:27
    So, it's, basically, the closest thing to sitting down and having a conversation
  • 00:02:32
    with someone to understand their experiences at key points in their journey.
  • 00:02:35
    So, there are two techniques that we use to do this.
  • 00:02:39
    The first is through sentiment analysis,
  • 00:02:41
    where you basically capture whether a response is positive, neutral, or negative,
  • 00:02:46
    and the other is through topic detection, and that's basically where responses are grouped,
  • 00:02:50
    or categorized, into certain themes or buckets that are important to your agency.
  • 00:02:55
    So, Qualtrics, in recognizing in which the way business evolves
  • 00:02:59
    and how much experience matters, it's not just enough to capture what experiences are,
  • 00:03:05
    whether they're positive or negative, but it's also important
  • 00:03:07
    to understand the what drives those feelings and how they can be addressed.
  • 00:03:11
    So, understanding Text iQ.
  • 00:03:15
    So, what is Text iQ?
  • 00:03:18
    Text iQ is basically a feature that Qualtrics developed
  • 00:03:22
    to support and capture exactly that, right?
  • 00:03:24
    So not only the what but they why.
  • 00:03:27
    So, it's a purpose-built feature that helps its users surface insights while removing the manual
  • 00:03:32
    and time-consuming process that frequently follows it.
  • 00:03:35
    So, because it's cumbersome, organizations often look at samples of written feedback,
  • 00:03:40
    but you risk losing some of the really rich content that you can receive.
  • 00:03:43
    So, Text iQ allows you to look at all
  • 00:03:46
    of that information automatically and make it easier for the user.
  • 00:03:50
    So, this feature, basically, allows you just to instantly understand a collection
  • 00:03:55
    of open-ended feedback and allows
  • 00:03:57
    for data-driven decision-making based on what you collect.
  • 00:04:01
    So, with that, through open ended questions, you can collect feedback in free-text form,
  • 00:04:05
    automate ticketing based off comments to quickly address changes or challenges,
  • 00:04:09
    and use dashboards to visualize consumable results through customizable widgets.
  • 00:04:14
    So, being proficient in using Text iQ, as you can imagine, can help your organization move
  • 00:04:19
    from being a more reactive to proactive organization
  • 00:04:22
    and help you not only understand the experiences, but help you predict the things
  • 00:04:26
    that matter most to your customers or your employees.
  • 00:04:28
    So, in their own words, and you let them define, basically, what is most important to them.
  • 00:04:38
    Okay, so I know the burning question
  • 00:04:40
    that everyone has is what's the benefit of Text iQ, right?
  • 00:04:43
    So we talked a little bit about this in the slide prior,
  • 00:04:46
    but I want to really hit home on why it matters.
  • 00:04:49
    So, firstly, the Text iQ helps you automatically uncover insights
  • 00:04:53
    to understand the why behind the qualitative data, right?
  • 00:04:56
    So, quickly and accurately.
  • 00:04:59
    So, the AI built into the feature finds patterns and trends in open text
  • 00:05:03
    and automatically suggests relevant topics that they align to.
  • 00:05:07
    And unlike other text analytics tools that require a lot of complex modeling and training,
  • 00:05:12
    this specific feature uses predictive text technology
  • 00:05:14
    to automatically produce high-quality topics for feedback to aligned to.
  • 00:05:19
    Secondly, in a second box here, Text iQ eliminates the manual labor that comes
  • 00:05:24
    from reading and tagging comments individually, which you can imagine,
  • 00:05:27
    gets it's pretty difficult as the population you're serving increases in size.
  • 00:05:32
    So, it automatically sorts the feedback and instantly updates with a click of a button.
  • 00:05:37
    So, is fully integrated with all of Qualtrics's reporting tools.
  • 00:05:40
    So, it can also be filtered, analyzed, and shared pretty easily.
  • 00:05:43
    Lastly, Text iQ automatically classifies sentiment of open text,
  • 00:05:50
    giving you a quick macro-level understanding
  • 00:05:53
    of how the population you're serving feels about a certain topic.
  • 00:05:56
    So, you can easily drill down into the specifics of the why
  • 00:05:59
    and read verbatim around what the comments are.
  • 00:06:02
    So, Text iQ quantifies the qualitative data so that you can see what areas
  • 00:06:10
    of an organization people are talking about most often, positively or negatively.
  • 00:06:14
    It also allows you to take a data-driven approach to identifying the area of focus
  • 00:06:20
    that will have the most and largest data-driven --
  • 00:06:22
    or, sorry, the largest effect on your population.
  • 00:06:25
    So, it makes your data more accessible, and if you already know that there's an area you want
  • 00:06:29
    to learn about, you can certainly drill down into them and understand,
  • 00:06:34
    or get a better understanding, of what people are feeling or thinking and how to address them.
  • 00:06:40
    One thing to note here, and that I really want to emphasize, though, is that it's important
  • 00:06:44
    to collect all this useful information judiciously, and we say that because we want
  • 00:06:48
    to ensure not only the quality of the data, as much as the quantity of the data.
  • 00:06:52
    So, to just throw a statistic out at you, based on our studies and research,
  • 00:06:56
    open-ended questions are particularly taxing on respondents,
  • 00:06:59
    and about 10% of respondents drop off for every open-ended question that you ask.
  • 00:07:03
    So, the best practice is to use a maximum of three or so questions
  • 00:07:07
    to optimize responses by asking for specific feedback.
  • 00:07:11
    So, there are a list of best practices towards the end of the deck.
  • 00:07:15
    So, a lot of that information, you'll be able to find as a reference.
  • 00:07:23
    Okay, so, I want to hop into the process for performing text analysis,
  • 00:07:28
    just to give you guys some insights on how it works on the back end.
  • 00:07:32
    So, what is the process for performing text analytics using Text iQ?
  • 00:07:37
    So, there are four different components here that I want to highlight,
  • 00:07:39
    and then we'll go into them deeper once we happen to the demo.
  • 00:07:42
    So, one, getting the right comments so, really asking, what is it that is being said.
  • 00:07:48
    Secondly, creating text topics, and that's really centered around, you know,
  • 00:07:52
    what do they mean by what they said?
  • 00:07:54
    And then, thirdly, using sentiment analysis,
  • 00:07:57
    which is really asking how did the experience make them feel when they responded?
  • 00:08:01
    The fourth one is around taking action on text insights,
  • 00:08:04
    and we'll walk through that a little bit later in the training.
  • 00:08:07
    So, if you look at the graphic to the right, I'll provide a very general example here.
  • 00:08:12
    So, say you've created a survey with an open-text question centered
  • 00:08:16
    around the customer's experience purchasing an item.
  • 00:08:18
    So, how does that work?
  • 00:08:20
    So, first, you develop an open-ended question that collects the response
  • 00:08:23
    around a specific experience, which you'll want to balance
  • 00:08:26
    with prompting some specific feedback, but also, give room for them to respond, right?
  • 00:08:31
    So, instead of asking a question like how was your experience with x, you'd want to, instead,
  • 00:08:35
    asked something like what did you like most about your experience with x?
  • 00:08:39
    Or what changes would most improve your experience with x?
  • 00:08:42
    That way, you can kind of target the type of information you're getting in response.
  • 00:08:45
    So, if you'll see here on the example on the right, say we asked something
  • 00:08:51
    like what do you most like about your experience in this retail store?
  • 00:08:55
    Right? And so, the customer says that "This is my second pair of boot-cut jeans.
  • 00:08:59
    It was wonderful to wear.
  • 00:09:00
    I highly recommend it, and it has great value."
  • 00:09:03
    So, based on that response, and off of the topics that you've created, either automatically
  • 00:09:08
    or manually, the tool then aligns that comment to a topic it resonates with.
  • 00:09:12
    So, you'll see the topics checked off over to the right there are apparel and value,
  • 00:09:18
    and that's because value and a piece of apparel was referenced in the comment.
  • 00:09:21
    So, similarly, if the customer said something like the employee
  • 00:09:25
    that helped me find a product was very helpful and created a great experience,
  • 00:09:29
    then something like effort or helpfulness would have been selected.
  • 00:09:33
    Further, you'll see on that last column there,
  • 00:09:35
    the sentiment is marked and rated as positive, right?
  • 00:09:38
    And so, that's determined based on keywords such as wonderful or or highly
  • 00:09:41
    or recommend or good value, right?
  • 00:09:43
    So, you'll see that the machine learning and the automated nature of the tool allows all
  • 00:09:48
    that stuff to be done kind of instantly.
  • 00:09:50
    So, I hope that introduction provided some just good background around text analytics,
  • 00:09:56
    and I'm happy to hop into the demo.
  • 00:10:07
    Okay, so, for this portion of the training, I'll do a demonstration of how you would use Text iQ
  • 00:10:14
    to analyze data captured within open text responses from a micro level, and then,
  • 00:10:18
    talk through why it matters at macrolevel and what types
  • 00:10:21
    of insights you can surface using the feature.
  • 00:10:24
    So, let's start with accessing and navigating Text iQ.
  • 00:10:27
    This is kind of the most important part to use in the tool.
  • 00:10:32
    The easiest way to access Text iQ is by clicking on your project.
  • 00:10:37
    Then hovering to the third icon here on the right-hand side that looks like a scatterplot,
  • 00:10:42
    and it's outlined in red here on the slide for easy reference.
  • 00:10:46
    But this is the Text iQ icon.
  • 00:10:48
    So, if you hover over it, there should be some text that pops up that says Analyze in Text iQ.
  • 00:10:53
    So, you want to click that, and it'll bring you to the Text iQ pane, which looks like this.
  • 00:10:59
    So, the Text iQ pane has two modes that you can use it in.
  • 00:11:04
    There is an editing mode where -- which is what you would use to add or modify topics
  • 00:11:09
    and widgets, and then, there's the analysis mode, which you can use to view visualizations
  • 00:11:14
    that depict trends in your data, which is what you see here.
  • 00:11:21
    So, before you can start adding topics, or widgets, you'll have to put Text iQ
  • 00:11:25
    into editing mode, and editing mode can be accessed by clicking that blue pencil icon,
  • 00:11:30
    also wrapped in the red outline there on the upper-right-hand corner.
  • 00:11:33
    So, after you're done editing, you click on the blue checkmark, which basically replaces
  • 00:11:38
    that pencil when you're in editing mode, and a summary of your edits will follow to apply
  • 00:11:43
    or discard changes that you've made.
  • 00:11:49
    Okay, so, now that you're on the Text iQ pane, a couple of components here I want to point
  • 00:11:58
    out before we jump into the live demo.
  • 00:12:00
    So, four specific components here that I want to talk to.
  • 00:12:05
    The first is the Fields drop down.
  • 00:12:07
    So, this is, basically, used to select which text responses you want
  • 00:12:10
    to analyze based on the questions that you asked.
  • 00:12:13
    So, every time you log in, Text iQ will remember what field you were on,
  • 00:12:17
    the last time you accessed it, and it makes it easy for you to pick up where you left off.
  • 00:12:22
    Secondly, there is a menu to the left, which allows you to select different topics.
  • 00:12:27
    So, you can view the responses tagged to them or change the topic's criteria, things like that.
  • 00:12:32
    There's also a section where you can click on the topics that are tagged or untagged,
  • 00:12:39
    and when you check off the untagged ones, you can see all of the ones
  • 00:12:43
    that still require topics to be aligned to them.
  • 00:12:46
    Third, there's also a search bar up at the top that can be used to search through responses,
  • 00:12:52
    build sophisticated queries, or create new topics,
  • 00:12:56
    we'll go into more depth on that a little bit later.
  • 00:12:59
    And then, fourth, there is a list of responses.
  • 00:13:02
    So, when you first log in, they will all display, and here,
  • 00:13:07
    you can take a look at a couple of things.
  • 00:13:08
    Every response appears with a sentiment on the left.
  • 00:13:13
    So, it's represented as a score there, the number eight, the topic that it aligns to,
  • 00:13:18
    and then, any additional details around it.
  • 00:13:20
    And then, lastly, I don't know if this is highlighted well.
  • 00:13:23
    It might be cut off a bit, but there's a sidebar on the right that contains topic recommendations
  • 00:13:28
    for topic building, and this is, basically,
  • 00:13:31
    something that Text iQ automatically generates for you to use as topics.
  • 00:13:38
    So, once you feel comfortable navigating this particular pane, you can start to add questions
  • 00:13:42
    that have open-text responses aligned to them, and I will demonstrate how we do that next.
  • 00:14:10
    Okay, I just want to make sure everyone can see this window here.
  • 00:14:13
    This should be, basically, the Text iQ pane that I referenced earlier.
  • 00:14:22
    Okay, so, now that you understand how to access the feature itself and are familiar
  • 00:14:27
    with the Text iQ pane, I'll start with a blank Text iQ pane to demonstrate how to add questions
  • 00:14:33
    and walk through the initial process for creating text topics.
  • 00:14:36
    So, regarding adding a question, if you go to this drop-down,
  • 00:14:41
    you're able to add questions that have text responses, right?
  • 00:14:44
    So, I've preloaded this one, and this question is, basically,
  • 00:14:49
    what could this company do better to help you be successful?
  • 00:14:52
    So, this is centered around the company that's trying to get a good gauge
  • 00:14:56
    of how their employees are -- what their employee experience is.
  • 00:15:08
    Okay, so, for this example, let's walk through how we add that question, right?
  • 00:15:13
    So, a link will appear at the bottom that prompts you to add a question.
  • 00:15:18
    Once you do that, it will automatically load all
  • 00:15:20
    of the responses here, oops, all of the responses here.
  • 00:15:24
    You'll see that each of the responses has this icon here, and a legend at the top that just,
  • 00:15:29
    basically, helps you designate and understand what it represents, right?
  • 00:15:32
    Additionally, there is an opportunity for you to edit them, if you feel like they're not,
  • 00:15:37
    you know, particularly accurate, or if you feel like the sentiment is more positive
  • 00:15:43
    or more negative than automatically told there.
  • 00:15:48
    So, please note that if there are any response that comes in, that there'll be a green button
  • 00:15:52
    at the top that will appear, and you'll just have to click it
  • 00:15:54
    to update your list of questions.
  • 00:15:56
    So, once all of that is loaded in, you'll want to create a text topic, right?
  • 00:16:01
    So, how you want to do that is there is an icon here -- I'm sorry.
  • 00:16:07
    There's a search box here where you would then enter a specific word, right,
  • 00:16:12
    that you think aligns well with the sentiment.
  • 00:16:14
    So, basically, how that works is if we're thinking about what a company could do
  • 00:16:19
    to better help use be successful?
  • 00:16:21
    What are some of the things that, you know, you might consider?
  • 00:16:25
    When I think about it, I think, may be a promotion would be impactful to my career
  • 00:16:28
    or training or flexibility or leadership or something like that, right?
  • 00:16:32
    So, I would type that word into the search bar to create a topic, let's say promotion,
  • 00:16:38
    and here, it'll tell you how many results aligned to that, right?
  • 00:16:41
    So, if we think that promotion is one of the topics that a lot of folks are commenting about,
  • 00:16:46
    we'll want to create that topic, and then, click add topic,
  • 00:16:50
    and it'll be included here on the left.
  • 00:16:52
    One of the other ones that I mentioned was maybe flexibility or leadership.
  • 00:16:57
    So, let's type in leadership here.
  • 00:16:59
    It's showing 30 results.
  • 00:17:03
    So, that means 30 of these responses that people have submitted have to do with leadership.
  • 00:17:08
    So, I would want to create that topic, as well.
  • 00:17:13
    So, once we search that term, we see, basically, every variation of that specific word
  • 00:17:17
    and the sentiment automatically aligned to it, and then,
  • 00:17:19
    we bring it over to the left-hand side.
  • 00:17:22
    So, once we have all of those tagged, I want to walk through a couple
  • 00:17:26
    of other ways that you can also tag topics.
  • 00:17:29
    So, you can also use a query builder to help you build topics, and how you would do
  • 00:17:34
    that is pretty similar to how you've manually kind of search for them.
  • 00:17:39
    So, say I wanted to go in a little bit deeper and think about training.
  • 00:17:44
    I click on training, and then I click on the word itself, it gives me a bunch of kind
  • 00:17:49
    of related text terms that people discuss that center around training, right?
  • 00:17:54
    So, say I want to add skill or train or operation or course or practice altogether,
  • 00:18:00
    and if I click enter,, it'll show me
  • 00:18:02
    that 65 results are aligned to that particular sentiment.
  • 00:18:06
    So, I can create that topic and rename it, say, training, and add that over, as well.
  • 00:18:14
    So, this is kind of an iterative process that you would go through until you can get through,
  • 00:18:18
    I would say I would say about 80% of your comments responses, right?
  • 00:18:22
    And then, lastly, what Text iQ offers is the recommended terms based
  • 00:18:29
    on frequently discussed terms automatically generated by Text iQ.
  • 00:18:33
    So, if you see here on the right, there is a number of keywords here.
  • 00:18:38
    So, pay, for example, is one of the ones that has the most sentiments aligned to it.
  • 00:18:43
    So, this is a topic that you might want to consider including.
  • 00:18:46
    So, it gives you a count of how many folks actually responded with a sentiment
  • 00:18:52
    that aligns there, and what percentage of the total population discussed them.
  • 00:18:56
    So, if I wanted to add that, I would just simply click there, create the topic, rename it as pay
  • 00:19:02
    or something relevant, and add that over.
  • 00:19:05
    So, this is kind of the process of creating text topics.
  • 00:19:08
    So, I see that there are some questions in the chat box here.
  • 00:19:11
    So, let me make sure that I am looking at these.
  • 00:19:18
    [mumbling] Yeah, so, David, I see your question here on the right here.
  • 00:19:21
    So, if you wanted to go about that in a more automated way,
  • 00:19:25
    you would go about that using this here on the right-hand side,
  • 00:19:28
    right, the topic recommendations.
  • 00:19:29
    You can pull in a bunch of those kind of topics based on the keywords and send them all
  • 00:19:35
    over to make them more automated.
  • 00:19:37
    We recommend that you go through kind of like a manual process, just so that you can see some
  • 00:19:42
    of the comments that are coming in, but absolutely,
  • 00:19:46
    you can go towards the topic recommendations to make it quicker and easier.
  • 00:19:51
    For this, once all of your comments are loaded, it just takes a bit of time
  • 00:19:57
    for Text iQ to automatically generate that.
  • 00:20:00
    So, it should, in the next couple of minutes.
  • 00:20:02
    If it doesn't, feel free to let me know.
  • 00:20:11
    Okay, so, once all of this is in place, and you have all of your items here
  • 00:20:16
    and topics generated, you would click the checkbox and apply these changes, and then,
  • 00:20:25
    kind of wait for all of them to be uploaded.
  • 00:20:29
    So, for purposes of this demo, I'll actually go through sentiment analysis
  • 00:20:32
    on a pre-populated one, so we don't have to wait here.
  • 00:20:41
    Okay, so, this is basically what it looks like once everything has been loaded.
  • 00:20:45
    Everything is kind of automatically updated here, and these widgets are put
  • 00:20:49
    in place to do all the analysis for you.
  • 00:20:52
    So, I'll kind of walk through this really quickly, just so you have a quick understanding
  • 00:20:56
    of what all of these widgets represent.
  • 00:20:58
    So, at the very top here is the bubble widget, and that shows, basically,
  • 00:21:03
    how many topics you have and how represented they are with the comments, right?
  • 00:21:07
    So, the bigger the circle, the more comments tagged to the topic.
  • 00:21:11
    And so, if you hover over a bubble, you'll see a drop-down appear,
  • 00:21:15
    and it works as basically a filtering system, right?
  • 00:21:17
    All right, so, if you click on a parent topic,
  • 00:21:19
    you can drill down into the child topics and see what they're made up of.
  • 00:21:23
    So, I'll kind of show that here.
  • 00:21:26
    So, here, if I hover over this claim ticket topic, you'll see that there is a number
  • 00:21:32
    of different types of responses, in terms of negative to positive responses,
  • 00:21:38
    and then the percentage around what the make-up is.
  • 00:21:46
    You also have, let me scroll down here, the topic change widget,
  • 00:21:52
    and what this represents here is, basically, how frequently that topic has grown
  • 00:21:58
    or shrunk, in terms of their reference.
  • 00:22:01
    On the right-hand side, there is a comments section, as well,
  • 00:22:06
    and that' there so that you can quickly understand what people are actually saying
  • 00:22:10
    and keep them in the right context.
  • 00:22:11
    You can also go back to your edit pane and expand the comments section there,
  • 00:22:16
    and like I mentioned earlier, adjust sentiment scores if you feel like they are not as positive
  • 00:22:20
    or negative as it was automatically tagged.
  • 00:22:27
    Okay? Scrolling down here, there's also a word cloud here that exists,
  • 00:22:33
    and what this really does is create a visual representation
  • 00:22:36
    of how frequently specific topics are being referenced.
  • 00:22:39
    The larger the word, the more frequently it's being used,
  • 00:22:42
    and you drill down using the graphic, as well, by clicking on any of those words.
  • 00:22:46
    So, let's explore these widgets a little bit further.
  • 00:22:52
    So, say I'm looking at the bubble chart here at the top,
  • 00:22:56
    and the claim ticket bubble is large relative to others.
  • 00:23:00
    So if we click that, you can kind of drill down to all the kind of child topics that kind
  • 00:23:07
    of align to that larger parent topic, right?
  • 00:23:09
    So, when we're thinking about claim tickets, people were talking about the assignment,
  • 00:23:13
    the steps for completing the process, instructions that are given,
  • 00:23:18
    and then things like that, right?
  • 00:23:20
    So, you'll be able to, once you drill down, look at specific responses aligned
  • 00:23:25
    to that particular topic, just to get a better understanding of what people's sentiment are.
  • 00:23:29
    So, you know, one of the responses here that is relatively negative is
  • 00:23:33
    that there's a lot of turnover in agents.
  • 00:23:34
    And so, for this example, specifically, this --
  • 00:23:37
    all of these responses are centered around an insurance agency.
  • 00:23:41
    You'll see that there are also some positive comments, as well.
  • 00:23:47
    You know, everyone's extremely helpful, and the claims process, just was it was explained to me.
  • 00:23:52
    So, you'll get a good handle for, you know, what people's sentiments are
  • 00:23:55
    and what people are feeling when you're looking at this comment box.
  • 00:23:58
    And then, if you just want to go back up to the original page,
  • 00:24:02
    you could just breadcrumb here and go back to all topics.
  • 00:24:08
    So, now that we have a better handle on how to access Text iQ and add questions
  • 00:24:13
    and create topics, that's really all you need to know to, basically,
  • 00:24:17
    have this automatic analysis page generated for you.
  • 00:24:20
    Okay, so, you know, kind of pivoting a little bit on the taking actions on text and insights.
  • 00:24:30
    Why is text analysis and Text iQ important at a macro level?
  • 00:24:33
    It makes it so that you can take action and mobilize others in the organization
  • 00:24:37
    to work towards key findings from Text iQ.
  • 00:24:40
    So, what you did well might reflect where there are best practices that can be leveraged,
  • 00:24:44
    while understanding where there are gaps can help you prioritize what pressing issues your
  • 00:24:49
    organization will need to address first.
  • 00:24:51
    You can also use these to create graphs and tables with text information inside a report
  • 00:24:57
    or on a dashboard, and you can look at what your organization did right,
  • 00:25:01
    but also look at where there are opportunities
  • 00:25:03
    to improve your experience with your stakeholders.
  • 00:25:05
    So, we can go back to the larger dashboard here to see how that works, right?
  • 00:25:12
    So, a lot of the widgets that are in that analysis page can be, also,
  • 00:25:18
    put into larger dashboards with quantitative data, and you can do that just
  • 00:25:25
    by adding a widget here at the bottom, which some of you may be familiar with.
  • 00:25:31
    There are some that are dedicated towards Text iQ, but you can also pull Text iQ data
  • 00:25:38
    or open text comments into other flexible charts and things like that.
  • 00:25:44
    So, any questions there?
  • 00:25:47
    I see things here on the right.
  • 00:25:50
    Julia, were you able to get those recommended text responses or topics?
  • 00:26:02
    Okay, so, Julie, I can follow up with you individually on what's going on there.
  • 00:26:15
    >> Actually, I have a quick question for you.
  • 00:26:16
    >> Sure.
  • 00:26:17
    >> So, is it safe to say that this Text iQ feature,
  • 00:26:21
    is this more of like a keyword analysis rather than a thematic analysis of qualitative data?
  • 00:26:28
    >> So, do I think it's a keyword analysis?
  • 00:26:32
    Definitely that.
  • 00:26:33
    I think that you can, very much, use this tool to do kind of like a quick, but thorough,
  • 00:26:39
    analysis of what's being said, and then kind of getting some quantitative data out of it, right?
  • 00:26:46
    So, getting sentiment scores aligned to it, getting a good feel for kind
  • 00:26:52
    of like what specific topics they aligned to, things like that.
  • 00:26:57
    >> That's okay, thank you.
  • 00:26:59
    >> Yep.
  • 00:27:04
    No problem at all there.
  • 00:27:06
    Okay, let me bring you back to the slide deck really quickly.
  • 00:27:24
    Okay, are there any other specific questions around the demo
  • 00:27:27
    or things that you want to walk through?
  • 00:27:29
    Otherwise, I can kind of point you towards some resources that I think will be helpful.
  • 00:27:34
    The processes, actually, generally, like they're relatively simple, right?
  • 00:27:37
    Once you have all of your open text comments uploaded in there,
  • 00:27:42
    it's just really aligning them to choose specific topics, either manually
  • 00:27:46
    or automatically, to then get a good feel for and breakdown of, you know,
  • 00:27:51
    how they align to certain topics and how you can drill down to them
  • 00:27:54
    to better understand what the sentiments are.
  • 00:28:03
    Okay, I will hop over to some resources, then.
  • 00:28:05
    So, let me pull up quick here.
  • 00:28:08
    So, just a quick plug here for some of the best resources
  • 00:28:13
    that I think would be helpful for your use during Text iQ.
  • 00:28:17
    First is the XM Basecamp website.
  • 00:28:20
    So, all of the Qualtrics users have this accessible for free.
  • 00:28:25
    You can easily click into on-demand training for Text iQ or hop on a webinar as they pop up,
  • 00:28:31
    and there are videos and guided tutorials and the links to access those are located
  • 00:28:35
    at the bottom of the slide there.
  • 00:28:37
    Additionally, you can navigate to the Qualtrics support page.
  • 00:28:41
    There's a link there at the bottom, as well, that takes you to Text iQ-specific support pages
  • 00:28:48
    and have step-by-step instructions for using the features.
  • 00:28:50
    So, you can quickly search for a specific question that you might have or scroll
  • 00:28:54
    through the table of contents on the left there.
  • 00:28:59
    And then, lastly, there's some best practice I put together for everyone that aligned
  • 00:29:04
    to the steps for just preparing your information and your data for Text iQ,
  • 00:29:08
    and you can kind of look through those in your own time as you're preparing for, you know,
  • 00:29:16
    developing your or using Text iQ.
  • 00:29:20
    Okay, so, I know that was a kind like a quick and brief introduction into Text iQ
  • 00:29:27
    and text analysis and using the feature itself.
  • 00:29:30
    Are there any questions, I just want to open it up to anyone
  • 00:29:33
    that has specific questions that I can answer for you all?
  • 00:29:38
    >> Karen, there are a couple of questions in the chat.
  • 00:29:41
    >> Okay, perfect, let me take a look here.
  • 00:29:45
    Okay, "You mentioned new features at the beginning.
  • 00:29:47
    Are you able to expand on what is new?"
  • 00:29:50
    >> Okay, yeah, so, a lot of this is actually on the backend, right?
  • 00:29:54
    So, nothing that you'll see on the front-end, but a lot of it is in how they're computing
  • 00:29:58
    and kind of generating topics, based on the comments that are put in there.
  • 00:30:04
    So, moving from a academics, like a heuristic algorithm approach
  • 00:30:08
    to kind of a machine learning one.
  • 00:30:10
    So, the more that you use the tool and make updates or manually suggest things,
  • 00:30:15
    the more that it continues to learn and grow and be more and more accurate as time kind
  • 00:30:21
    of progresses, does that kind of makes sense there?
  • 00:30:24
    >> Does that mean if I've got models that are already set up,
  • 00:30:27
    topics that I've already created, is it going to go in and reclassify those issues,
  • 00:30:34
    or those entries, or is this just changing how the recommendations are going to be made,
  • 00:30:39
    or is that going to affect something that I already have in place?
  • 00:30:43
    >> So, no. It won't affect anything that you have in place.
  • 00:30:45
    It's more that the machine is learning what you're putting
  • 00:30:50
    in to make future once more accurate on the backend, if that makes sense.
  • 00:30:54
    >> So, are you talking about the recommendations then, the recommendations will be different?
  • 00:30:59
    >> Right, that's correct,
  • 00:31:00
    >> Okay, all right, thank you.
  • 00:31:02
    >> Yep, no problem.
  • 00:31:12
    >> Katie also had a question about whether Text iQ analysis could be exported to other tools
  • 00:31:18
    like Tableau or other analysis tools outside of Qualtrics?
  • 00:31:22
    >> Okay, Yep, I see that question there.
  • 00:31:24
    From what I understand, I believe that you can, but let me --
  • 00:31:28
    I will need to check in on that and confirm.
  • 00:31:37
    So, let me take that for action there.
  • 00:31:49
    So, I'll take a list of questions that I'll follow up on, and then,
  • 00:31:52
    I'll send them over to you, Camille, as you can share that with the group.
  • 00:31:57
    >> Sounds great.
  • 00:31:58
    Thank you.
  • 00:31:58
    >> All right, perfect.
  • 00:32:00
    Let me see if there are any other questions here.
  • 00:32:04
    >> I've got one more question.
  • 00:32:06
    Something that we've noticed with the sentiment analysis is that if somebody has a complaint
  • 00:32:10
    but they phrase it really, really nicely, it comes up as being more positive than,
  • 00:32:15
    just because they said it really nice.
  • 00:32:19
    With the new sentiment analysis, does it break that out any differently, or how does that work
  • 00:32:32
    like where -- our stuff comes across as overly positive sometimes,
  • 00:32:39
    just because the people answering are very polite.
  • 00:32:42
    >> Yeah, definitely, and that's why we kind of mentioned doing some
  • 00:32:45
    of that manual analysis up front, right?
  • 00:32:47
    Because when you go into that sidebar on the top-left corner,
  • 00:32:51
    it tells you what's been classified and what hasn't, and sometimes,
  • 00:32:54
    that'll show up is not being classified, and you'll be able to click on those and then kind
  • 00:32:59
    of manually kind of adjust those, if necessary.
  • 00:33:03
    So, I will say that, you know, with the updates to Text iQ, that it'll be more accurate,
  • 00:33:08
    but for those kind of like one offs, I think it would be challenging.
  • 00:33:12
    You would probably have to take a look at that.
  • 00:33:18
    >> Okay, I mean, overall relative, you know, we can see our relative trends, you know,
  • 00:33:22
    and see if we're doing better or worse, but yeah, we have noticed that, yeah,
  • 00:33:27
    if people are nice that it comes across more positive, yeah.
  • 00:33:32
    >> Yeah, absolutely, and then, when you recognize that that is the case,
  • 00:33:35
    you can manually kind of override the rating or the sentiment score that's assigned to it
  • 00:33:40
    or aligned to it, and hopefully, that will make the tool increasingly more accurate
  • 00:33:46
    as it continues to filter through all the comments.
  • 00:33:51
    >> Thank you, yeah.
  • 00:33:52
    >> Yep, no problem.
  • 00:33:53
    >> Karen, if we're doing multi-year analysis on Text iQ, are we able to add filters
  • 00:34:03
    to see the different years or to see any kind of trends, as well?
  • 00:34:08
    >> So, there are -- here, let me pull this back up here and take a look at that for you.
  • 00:34:15
    So, there are specific widgets that actually trend over time.
  • 00:34:19
    So, you are able to capture that.
  • 00:34:21
    Let me check really quickly to see if you can filter on those.
  • 00:34:40
    Okay, so, we're in the Text iQ field here.
  • 00:34:43
    So, you can see a topic change over time, I believe that you can create a widget
  • 00:34:51
    and have the -- one of the axes be time.
  • 00:34:54
    And then, you'll be able to look at trend over time.
  • 00:34:58
    >> Thanks.
  • 00:34:58
    >> If that would be an issue here.
  • 00:35:09
    Okay, and then I see, Christy, that you had a question earlier
  • 00:35:11
    that "Will this slide deck be distributed to participants after the presentation?"
  • 00:35:15
    Yes, I sent this over to Camille and Dion [assumed spelling].
  • 00:35:18
    So, they should be able to provide the deck to you all after the training.
  • 00:35:32
    Okay, any other questions you all have?
  • 00:35:45
    Okay, so, if you just go through the deck and referenced the resources
  • 00:35:49
    that I have listed there on the slide, it should be able to provide similar details, if not more,
  • 00:35:55
    kind of advanced kind of information around how to use the tool.
  • 00:36:02
    So, if you need that, feel free to reference that, and if not, happy to also take questions
  • 00:36:06
    as necessary or connect one on one where necessary if things are working
  • 00:36:11
    or if you aren't getting things to populate properly.
  • 00:36:14
    I can absolutely support that, as well.
  • 00:36:33
    Okay, so, we'll pass it back over to you, Dion [assumed spelling] or Camille.
  • 00:36:37
    If there's nothing else, I'm happy to close up early, but also,
  • 00:36:40
    happy to take any additional questions if you have them, as well.
  • 00:36:48
    >> Any final questions for Karen?
  • 00:36:58
    >> Okay, perfect.
  • 00:36:58
    So, if there are any kind of issues that you all are consistently seeing across the board
  • 00:37:04
    or anything, I can hop back into one of your recurring meetings and do kind
  • 00:37:09
    of a more thorough overview for any specific area
  • 00:37:11
    that you guys might be encountering challenges.
  • 00:37:13
    So, just let me know how I can be helpful there, and I'll be there.
  • 00:37:16
    >> Great, thank you so much, Karen.
  • 00:37:21
    As a reminder, this is a standing meet at noon Eastern Time.
  • 00:37:24
    This is our office hour.
  • 00:37:27
    So, if you need one-on-one support with anything Qualtrics related, please feel free to drop in,
  • 00:37:30
    and if, like Karen said, if we are hearing more folks --
  • 00:37:33
    that you want to hear more about Text iQ,
  • 00:37:35
    we'll have Karen come back and talk with us more about it.
  • 00:37:39
    So, with that, I'll give everybody a lot of time back in their day.
  • 00:37:42
    Thanks a lot for joining us.
  • 00:37:43
    Have a great afternoon.
  • 00:37:45
    >> Thank you.
  • 00:37:46
    >> Thank you.
  • 00:37:48
    >> Thank you.
الوسوم
  • Text iQ
  • Qualtrics
  • text analysis
  • sentiment analysis
  • topic detection
  • machine learning
  • qualitative data
  • feedback
  • data-driven decision-making
  • best practices