Survey for Success: Growing Your Business with Customer Insights

1. Introduction: I would say that there's always a need for a survey no matter how silly or how serious it is. Hi, I'm Sarah Cho. I'm Director of Research at Survey Monkey, and today's class you'll be learning about how to conduct your own surveys. I like to say that surveys are a conversation that you're having with a very specific purpose, and that you can really efficiently have at scale. If you're interested in learning about how you can utilize surveys to grow your business, or how you can utilize surveys just in everyday life, I think everyone probably has one or two questions that they're super curious about learning. In the class today, what we're really going to be covering is the survey process from soup to nuts. First, we'll go over how to think about planning your survey, then we'll move into how do you think about the questions that you're going to ask, who you're going to send the survey to, how do you think about how to distribute your survey, how many people should you get in your survey, how should you analyze your survey, and then once you have the data, how do you act and make a decision upon your survey. I hope you'll leave this class with the confidence that you really can conduct your own survey. It's not that hard. It is part art and part science, but really it is manageable process that you can guide yourself through. If you're following along, share your survey plan in the project gallery and feel free to ask questions. As a professional question asker myself, I always love to ask questions and answer them. So feel free to do so and follow along in the discussion below. I'm super excited you're joining the class. So let's get started. 2. The Power of Surveys: So before we get into the really nitty-gritty details, let's actually step back for a second. So what actually is a survey? A survey according to the dictionary is really a close examination of someone or something. So you can think of for example, a land survey it's when someone is going out there and actually really digging deep and seeing what's in the landscape, how far is this one distance to another. Well, a survey is actually the same thing but you're asking questions and you're having a conversation with a purpose at a massive scale. So instead of going and asking one person or 10 people or a thousand people the same questions which would take you forever to do in person, you can really do so super efficiently through a survey. So I'm sure everyone has had some sort of experience with a survey, whether it's responding to a survey or even designing one of your own surveys. There's many different types of surveys, so the first and most common type of survey that we see is a customer satisfaction survey. Basically, we've all seen those, and those are the ones that you get a receipt at Walmart or CVS, one of those really long receipts, and use the little code at the bottom and you scan the code and then it takes you to a survey. Or maybe you are just staying at a hotel and they ask you to rate your stay. Those are all different kinds of customer satisfaction surveys, and that's probably the most common type of survey, which is why we've made that the example that we're going to cover today in the course. Another popular type of example is an employee satisfaction or employee experience survey. What that is is basically, if you've worked in a traditional office in environment, often people will want to know, what is it like to work here, are you happy with your job, you might have responded to one of those. Another common type of survey is a market research surveys. So people who are market researchers are in that industry, are utilizing surveys for all different kinds of things, like how much should I price my product for or I have two different products that I'm thinking about bringing to market, which one should I bring in which one is more relevant for my target market? Probably the biggest survey that people are familiar with is the US Census, and this is obviously done every 10 years, they survey every single person who lives in the US. What they do is they make really big decisions based on this one survey that's done every 10 years. Another fund type of survey is maybe the ones that you are doing in your personal life. So I use surveys for example when I go camping and I want to know what kind of dietary restrictions do people have, I want to know what are people thinking about brains, so we don't get to the camp site and have 15 tents and no cooking utensils. So that's an easy way to organize, is through a survey. Finally, you've probably seen a lot of surveys even on social media. So for example for those of you who use Instagram, those poles stickers on Instagram stories is another form of the survey. So there's lots of uses and applications for surveys, so I'm sure in everyday life you've come across one of those. So you're probably wondering why should I do a survey? What are the benefits for me to do a survey? There's a couple of clear benefits. So the first thing is your own sales data or your own information probably doesn't always show the full picture. It doesn't show where you can potentially improve, and where you could potentially see opportunities that you might not have seen. A good example that was done by Ford actually, so it's a really large company but this applies to any individual person or any small company. Is that, right before the recession when the auto market taint, what they did was if you look at the sales for SUVs it was all on the rise. Based on sales data, any auto company might have invested a little bit more in creating more SUVs, more trucks. But what Ford actually did, was they actually went and surveyed people and their customers, and they found that actually people just purchased an SUV so they're not looking to purchase another SUV. But the next car they're looking for something a little bit more fuel efficient, that doesn't causal gas. So what they did was they invested a little bit more heavily and they're more fuel efficient line of cars, and that was one thing that really helped them whether that sort of the Auto Crisis a little bit better than other companies. So another benefit is that one question in a survey can clearly actually make really, really radical changes in an experience. So a really good example of that is Netflix. So around 10 years ago, when Netflix started their streaming service, I don't know if anyone remembers that, but everything was all under one account. Say you would log in and say someone else in your household maybe it was your roommate, your partner, your kids, all would be sharing one account and you would see all your recommendations in one place. So if you are watching documentaries and you are really interested in documentaries, but your kids are watching Sponge Bob or whatever cartoons, what you would see is recommendations for documentaries but also recommendations for the kids cartoons which you probably aren't really helpful to you. But then Netflix got really smart, they ask just one question which was, who's watching and they set up profiles. So now your recommendations are tailored to you. So that's an easy way to show like how one question can make a huge difference to the experience for the end user and also probably for Netflix as well, because they can have a little bit more information on how to be more helpful to you. Now going more deeply into the benefits of a customer satisfaction survey. So why should you do one? Again, it identifies opportunities where number one you can see where should you be improving. Number two, you might be looking at Yelp reviews or maybe reviews on your Etsy store or maybe testimonials, but that really only gives you one side of the picture. It only gives you the picture who are volunteering to give you a review. So usually those end up being two sides of the coin. People who really, really are enthusiastic about your product or service and people who really didn't like your product or service. You don't get that picture from the people in the middle, who maybe were "eh" about it and these are the people that you really probably could have as repeat customers, because you just need to push them more into that really excited about your product or service, they're a lot easier to convince than those people who are on the opposite side of the spectrum, who are really not looking to return to you as a customer. Another benefit of customer satisfaction surveys, is they might be able to identify opportunities for you to expand that you may not have ever thought of, but your customers are thinking about. The final benefit is that if you know who your happiest customers are, you could actually tap them for things like testimonials on your website. Also, if you who your most unhappy customers are obviously you want to address that issue first, so there's lots of benefits about customer satisfaction surveys. So who can use a survey? Well, really anyone can, from small business owners and entrepreneurs, all the way to those big large corporations. We've actually even have little kids do their own surveys, so really anyone can do their own survey. Hopefully by the end of this course, you'll have the confidence to do one on your own. Because of the online survey tools available today, you don't actually need to be a mathematician to know how to analyze your survey data, you don't need to be a social psychologists to know how to ask the questions, really a lot of that is in the product to help you along. So you really can have the confidence that you can do this on your own. So today, we'll be working through the survey process from soup to nuts or from nose to tail, if you're a mediator, and really we'll be starting first with, how do you plan the fundamentals of your survey to make sure your survey is successful? Then we'll move on to, how do you write your questions so that you make sure that you're eliminating any bias that your questions might be unwillingly providing? Next we'll move on to, how do you think about logic and programming and designing your survey, so that you make sure that the design and the logic are customized to your respondents, to your brand, and also that you're making sure that they're not answering any questions that they don't need to be? Then we'll move on to thinking about, who to send your survey to and how many responses you might need. Then finally the last step is analyzing your data. So the example that we will be utilizing today is a customer satisfaction survey which is one of the most popular use cases. So while the example that we'll be utilizing today is for an Etsy store owner, you can really apply these principles for any type of business. So for the Etsy store owner, we will really be focusing on someone who does a digital and paper-based invitation business, and looking at how happy and satisfied their customers are and seeing what they can improve with their product. So even though we are utilizing customer satisfaction as the example, this survey principles are really hold true regardless of whether maybe you're serving your friends, or maybe you're serving your target market, or maybe you're surveying, if you have employees, your employee base. So really the principles remain the same, even if you are not doing a customer satisfaction survey in your own everyday lives. So what do you need to get started? So number one, you obviously need access to a computer and an online survey platform. We will obviously be utilizing Survey Monkey today, but there's lots of other options out there and you can figure out what's best for your needs. What's also helpful is to have some sort of word processing software like Google Docs or Word, so that you can really document what the procedures that you're following are, so that if you ever need to go back to what you have done, that's all there and recorded. Okay, now let's get started, and let's think about how do you actually plan for your survey. 3. Setting Goals: We've all taken a really bad survey before, whether it be too long, too confusing or worst, both. Usually, that happens when the survey creator is not really paying attention and not properly planning the survey before they actually write the survey. So how you can avoid that is actually creating a plan before you get started. So we're going to walk through the steps of how you can think about doing a plan, so that the survey is clear and concise for your respondents. So there are three key parts of creating a plan for your survey. First is setting a goal for your survey. Second is thinking about your analysis plan. Third, thinking about who you're sending out the respondents to, and we'll go through all three steps in detail next. So first thing that you want to do is to create a goal for your survey. The most important thing about having a goal is that you don't want to try to have your survey accomplished 10 things at once, because that gets into that territory where it's going to be too long or too confusing for your respondents. So really try to make your survey as concise as possible, and how you can do that is by really only having one or two goals that your survey wants to accomplish. How you can think about formulating a goal is think ahead. Think ahead to the decision that you want to make, and maybe that can become your goals. It could be as generalized, "How happy are my customers?" but it could be more specific like, "With this data, I want to learn how to improve my product to make my customers happier." So that could also be a goal. You can get more specific like that. The more specific your goal is, the better the data you will get in the end. What I would do is once you have that goal, is write it down, put it on a post-it, write it at the top of the document, and then reference it as you're moving along. If the question doesn't help you move towards your goal, then that's probably something that you should cut. We'll go through length a little bit later but shorter, in general, is always better. So that's why always referencing your goal and making sure that all the questions in your survey kind of lead up and feed into that goal is always the best. Once you've established your goal, the next thing that you want to do is actually think about your subgoals, and what your subgoals are is just one level deeper than your overall goal. So if your overall goal is to learn about how happier customers are, a subgoal might be how happy are they with the quality of your product or how happy are they with the price of your product. So that all rolls up into the overarching goal of how happy are they with your product, but they're a little bit more specific. Those more specific subgoals are the things that are actually going to become your questions in your survey. We'll talk a little bit later on about how to write those questions effectively, but first you need to have your goal, and then your subgoal. If you're actually not quite sure what your goal is or maybe you don't have a good sense, what you can actually do is do a small survey beforehand to actually narrow it down. So going back to the customer satisfaction example, say you know you want to know generally about customer satisfaction, but you're not sure how you're going to use the data in the end. You could do a quick customer satisfaction survey and ask maybe a small proportion of your customers, a general question about how happy they were with your product, and that might inspire some ideas to narrow down what your goal might be and get into ideas around the quality of the product or maybe it's the price point. Then you kind of figure out, "Okay, it's the price point I really want to narrow down on, because people seemed a little bit unhappy about that." Then that becomes really the goal for your survey is, "How do I more effectively price my product for my customers?" The second thing that you want to think about after you've established your goal is who are you surveying. So your goal will probably already help you out with that. So in a customer satisfaction survey, it's pretty obvious. It's pretty easy. It's just your customers, and that's your audience. But in a market research survey, that might be a little bit more complicated because usually, with market research surveys, it's your target market. But maybe you don't know what your target market is, so you'll have to maybe do a bit more research and digging to figure out who your audiences, depending on the type of survey that you're doing. But it's really important to know who your audiences before you actually get started, otherwise you might be surveying people who aren't relevant, again, and then that goes back to the first piece of planning your survey is. You don't want a bad survey experience for your respondents. So if you are sending out your survey to the wrong audience, that actually might make that poor respondent experience. So this is another way that you can avoid that, by making sure you clearly know who you're going to send your survey to. Finally, the last piece in planning your survey is actually thinking ahead to the analysis stage and thinking about your analysis plan. So often, what happens is we're so focused on our goal, we're so focused on the questions that we might forget to add in something that you might want to cut the data by. What I mean by cut the data by is actually analyze your survey responses by some sort of subgroups. So for example, you might be really focused on making sure that you're asking the right questions about whether your product is high quality or whether your price point is the right price point. Then, once you actually collect the data, you've forgotten, "Oh, my gosh. I wanted to know how men might differ from women on this." So that's really important to think ahead to the analysis stage because what if you forgot to ask a general question on your survey? What you might want to think about is time-bounding your audience. What really that means is, depending on who you're surveying and what you're surveying them about, you might want to think about restricting people who've purchased more recently from you or people who've used your service more recently. So a good example is, if you're a realtor for example, people generally remember the whole entire process about buying their home because it's such a large investment in terms of time and money that you're not really going to forget about what it was like to buy your house. So if you're a realtor, you probably can just ask all your previous customers about whether or not they were satisfied with the service that you provided to them. But say you are a masseuse, masseuses generally, if someone is getting massage, they might be getting a massage on a regular basis. So they might have gotten another massage since you saw them. So you maybe want to time-bound your customers, instead of saying anyone who's gotten a massage from you anytime. Because someone's probably not going to remember a massage from three years ago, but they might remember how the service was from your massage from like six months ago. So that's why you should think about the timing around it and maybe only survey your customers who had been more recent. If your product is something that people utilize more on a regular basis, as opposed to if your product or your service is something that is only purchased or utilized very infrequently, then you probably can think about having a longer time frame. So that's why it's important to think about how you're analyzing the data and the different groups that you might want to dig into in your analysis phase ahead of time so that you also have those questions represented in your survey. So that means that you'll take your subgoals and your analysis plan, and those two together will formulate the questions that you'll ask in your survey. So planning for your survey doesn't require a lot of time investment, but the payoff at the end is really great. So really take the time to invest maybe an hour, two hours. I mean, depending on the project, it could be a couple of days or a couple of weeks. But just take that time to invest the time on thinking and planning for your survey, because ultimately at the end, that will really help the results that come out, and also, it will help the respondents who are taking your survey. Okay. So now, let's walk through this survey example with our Etsy store owner. So as you can see here in this document, I've already laid out the goal, sub-goals, audience, and analysis plan. Again, you don't have to do it in a Google Doc. If a post-it note is a little bit better for you, you can write it on a post-it note or you can write it in your notebook. But the idea is just to think about the planning process ahead of time before you actually get started. So you can see here, for the Etsy store owner, she's really looking to see how satisfied her customers are with her invitations and identify opportunities for improvement. That's the main goal. The sub-goals that she has is, "Do people like the quality of the invitations?", "Do they think they're getting good value for their money?", "Should she consider changing her shipping provider?" Some of the feedback that she's gotten into her Etsy reviews have said that the shipping is maybe poor, so she wants to dig in a little bit deeper on that. So those will become your individual questions. So you're probably wondering why the subgoals seemed to be very specific and may not seem directly related to satisfaction. But satisfaction with your product or service is often multifaceted. So that's why things like shipping, the quality of your product, the price of your product are all components of satisfaction. So that's why those pieces will become your individual question. Of course, you want the overall like, "How satisfied or dissatisfied are you with my product?" But you really need to dig down into what are the factors that are driving that satisfaction, and so that's why you need to think about what are the components that might be at play here. Next is thinking about the audience. Here, we're going to really think about the audience and maybe a time-bound format. Because particularly for an Etsy store example, you might not remember a purchase that you made a year ago, but you might remember a purchase that you made six months ago. So here, we're really looking at customers who've purchased an invitation within the past six months. Finally, the analysis plan. So the first thing that we want to do is look at the topline results. The topline results are basically your results in the raw percentage format with no filters, no deep dives attached to it. So it's just literally the percent of people who think you're product is high quality or the percent of people who think that your product offers really good value. So that's the first thing that I want to do in my analysis plan. It should be the first thing that anyone wants to do in their analysis plan. The second is I want to dig deeper into those people who have purchased multiple invites versus those people who have only purchased one type of invite. Say someone who's just purchased for their baby shower versus someone who's purchased an invite for their baby shower, their wedding, their anniversary party, all their birthdays. They might have different experiences with the product than people who've only purchased one time. I also want to look at, say the Etsy store owner has just now expanded their business to actually start targeting corporate events. So you might want to look at, "Do customers who purchase invites for corporate events have higher or lower satisfaction rates than people who purchase for personal events like a wedding or a baby shower?" So if you're following along, right now is a great time for you to take a couple minutes to write down the goal, your subtopic, your audience, and your analysis plan in a document, or on a post-it, or on a piece of paper and submit that to the project gallery. Or if you're not doing a customer satisfaction survey, you can also take the time to do it for any other type of survey, because the principle still apply regardless of the topic of your survey. So now that we've planned your survey out, next step is actually getting down and dirty in writing your questions. 4. Writing Questions: Okay. So now really, let's get into, How do you actually formulate your questions and how do you write the words that are in your question? So a couple of tips of advice. The first thing is, make it as user-friendly as possible. That means you should try to utilize more close-ended, multiple-choice questions and avoid those open-ended questions because of the reasons that we talked about earlier. The second thing that you really want to do, is make sure that your questions are clear and concise. There's four tips that you can think about to make sure that they're clear and concise. Number one, make sure that you are using language that is approachable. So things like acronyms or jargon or abbreviations, are really no-go's because that might not be vocabulary that, even if you're using every day, your respondents aren't using every day. A great example of this and my favorite example of this is, if you're in the restaurant industry or a retail industry, you really know what a P.O.S is, right? It stands for point of sale. But if you're not in the restaurant or a retail industry, you might think P.O.S is something very different. So if you utilize that acronym, you might be getting yourself into a little bit of hot water there. Instead, you should really ask, "What is the checkout process efficient or how long did you have to wait to checkout or did they process your credit card very quickly?" Those are the kinds of things that you might want to ask about instead of how would you evaluate the P.O.S system? The next on how to create a very clear and concise question is making sure you're only asking about one concept at a time. Sometimes, two things go together so you think that they really intermix, like hamburgers and hotdogs. So when you go to a barbecue, almost always there's both hamburgers and hotdogs. But if you asked in a survey question, "Do you like hamburgers and hotdogs?", what if the respondent loves hamburgers but hates hotdogs or vice versa? They would have a very hard time answering that question. So it's best to only ask about one concept at a time. Third, when you're using open-ended questions, make sure they're very concise and precise. So a really good example of a question that's not concise or precise, is just asking an open-ended question of, "Where do you live?" When I've asked that before, people answer, "In an apartment or in New York City or in a bedroom." Those are all really valid answers of where people live, but you might be asking more specifically about, "Do you live in an urban-rural or suburban environment or what state do you live in?" Then the last step is around answer options. So if you're utilizing five answer options at the beginning of the survey, make sure that the number of options are consistent throughout. This really only applies to answer options that are in scales like excellent, very good, good, fair, poor, or on a scale of one to five or something from extremely important to not at all important. Those are examples of scales where there's an order to in which they increase or decrease. So if you're using five points at the beginning of your survey, you want to make sure that you're consistently using five points and not switching to four points later on. Because what that means is, in the analysis phase you can really compare question to question to question, which you'll see in the analysis section of this course. Another type is, you don't need to reinvent the wheel. Often people have already invented the wheel for you. So what you can do is you can actually go and Google for example. So if you know you're doing a customer satisfaction survey, you can actually Google customer satisfaction survey examples, and there's a plethora of examples that you really can draw from. So you don't need to go through and think about, How do I really word this customer's satisfaction question? A lot of that work is already done for you. So you can go and search for that. Survey monkey, for example, has a lot of templates that are already available so that you can actually go in and have pre-written entire surveys if you need, or just questions that you can drag and drop from our question make. But those are just ideas where, again, you don't need to reinvent the rule, you can get inspiration from other places. So if you search for a question and you find a great question that you want to utilize, not only do they just invent the wheel for you, but you can also utilize the responses and that provides you valuable benchmark data or comparison data that you can use in the analysis phase. So one of the most popular questions I get is, how many questions can be on my survey or how long is too long of a survey? Really, unfortunately, the answer is, it depends. It depends on two things. Number one, the topic of your survey and number two, who you're surveying. So if you're thinking about sending say, 100 questions survey about Ariana Grande, most people would not get through 100 questions about Ariana Grande. So that's the topic, it's Ariana Grande, but the audience? Most people wouldn't get through it, but say you were sending it to her fan club, I bet you that her fan club would get through 100 questions. So those are the two factors that you need to consider. You really want to make sure that your survey is concise because, for most of your customers, they're not ingrained in your product every day like you are. So you really want to ask maybe 10, 20, maximum 30 questions, while maybe your employees, they're really thinking about your product or your service all the time. So they might be able and willing to answer 100 questions. So you really have to vary it based on who you're trying to survey and what you're surveying them about. If you really need a hard and fast number, a good rule of thumb is try to keep it around 25 questions because generally, that's maybe a couple of minutes of someone's time. Can a survey ever be too short? Well, sometimes you can actually have too short of a survey because it's for you as the survey creator and the survey analyst. You're not going to have enough information to actually make a decision upon. So well, from the respondent perspective, a survey can never be too short. From the creator's perspective, it could potentially be too short. So that's why going back to an earlier lesson of making sure that you're planning your survey very thoughtfully, you want to do that so that you avoid, maybe, your survey being too short and not asking about something that in the analysis phase you go back and now you're all like, "Oh no! This is critical to ask." So really plan your survey well because sometimes a too short survey may be great for your respondent, but not so good for you. Now, going back to our survey example of the Etsy store owner who's surveying her customers about the satisfaction with her product. Again, you don't want to reinvent the wheel here. So for this, I'm utilizing the Survey Monkey customer satisfaction template as the basis for the survey and then customizing it to the Etsy store owner's name. Again, you don't need to use a Survey monkey template, you can just Google and find an example survey that's right for you. But even me, a professional survey question asker, I do this all the time where I'm borrowing from other people. So now, let's look at the actual questions. So the first question in the survey is the Net Promoter Score question, which is a very common methodology that people use in customer satisfaction surveys. The advantage of utilizing this question type is that so many people ask this question, you can easily have a benchmark available to see how you're doing compared to someone else. The next couple of questions are just general questions from the template about how satisfied people are with your product or your company, and what attributes they associate with, the invitations that the Etsy store owner provided, how well the invitations met their expectations. Then going down to the bottom we're asking a lot about things that are a little bit more specific, which is, "How would you rate the shipping process?" That's not available in the template because not everyone needs to ask about shipping, but it's one thing that the Etsy store owner really, really wanted to know, is dig a little bit deeper on that shipping. Then the other piece that is more customized is the different types of invitations people have purchased. So you can either ask this in a question or if I already had the information available, what I could do is append that information, when I send out the survey, so I don't need to ask it in the survey. So there's a couple of options here. Here I've gone with the easier way where I'm just asking it straight out, but you can always append it to your customer list if you prefer to do that. How did I decide what different question types for going in this survey? So number one going back to the couple of principles that we just talked about, I really wanted to limit the number of open ends. As you can see here, I'm only asking two open ends. That's really because I don't want people to have to think too hard or type too much in the survey, really you're asking them to give their feedback out of the goodness of their hearts and spend their time to do this, so you don't wanna take too much of their time. So you're probably wondering, can I use other types of question types? Absolutely. Some of it is just personal preference. So for example, a lot of these questions I could have used a star rating question, but I'm using the regular multiple-choice board question because that's just generally what I prefer. Now, thinking back to the sub-goals of the survey, one of the sub-goals that I had is, "How do people think about the quality of the invitations?" and the other thing was, "Do people perceive a good value for what money they spend on the invitation?" As you scroll down in the question below, you can see that there are two questions on that. How would you rate the quality of the invitations and then how would you rate the value for the money of the invitations? That's how I really mean making sure I'm meeting those sub-goals because again, those sub-goals feed into that overarching goal of like, "How satisfied are they with my product?" because quality and cost are often the two biggest components of whether someone is satisfied or not. One sub-goal doesn't necessarily mean it equals one question. It might mean that if it's a very large sub-goal, you might need multiple questions to really evaluate that or you might just need one. So now that you have seen us walk through this example survey, why don't you create your own questions. Go through maybe do a couple of Google searches and see if there's any questions out there that you might want to re-purpose. Then write some questions on your own. Next up, we're going to talk a little bit more about logic and how to utilize that to customize your survey for your respondents. 5. Customizing with Logic: So in this lesson, we're going to learn about logic, and logic can encompass a lot of different things when it comes to surveys. But for this particular lesson, when I talk about logic, we're going to focus on skip logic, and what does skip logic me? It really means that you're directing certain questions to only be seen by a certain proportion of your respondents based on a previous response that they've given. So logic can be a little bit tricky. So this is one of the trickiest pieces in the whole entire survey, but the benefits of logic is that you can really customize this survey so that's relevant to your respondent. So say for example, I came up to you and I started talking about Stradivarius violins. I used to be a semi-professional violinist, and for most people, they would be like what's a Stradivarius? Then if I added more on top and I asked, "Well, let's debate the merits between a Stradivarius and Guarneri. " You would be lost. You're never going to have a conversation with someone where it's not relevant because they don't have anything to contribute. The same thing goes for surveys. So if you think back to what a survey is a conversation with a purpose, you want to make sure that that conversation you're having at scale is relevant for the respondent, and how you do that is you add in survey logic. What survey logic does is it only routes the questions that are applicable to the respondent based on a previous response. So say for example in that customer satisfaction survey example, the Etsy store owner wanted to follow up with people who had only purchased wedding invitations from her asking, "Hey would you think about purchasing baby shower invitations from me in the future?" So you can actually do that utilizing logic rather than exposing everyone in your entire survey to that baby shower question, because maybe they just bought a baby shower invitation from you, so they don't need another one, or maybe they're there for their corporate events and holiday party, so you want to make sure that they're not seeing any weird recommendations for baby shower invitations, and you're only showing the most relevant questions to the right people. So as you're thinking about logic, you might be reviewing your survey in realizing "These one or two questions are only applicable to this one group of respondents." So you may realize that, "No I forgot to ask if these group of respondents actually fit into that category." Maybe realize that you forgot to ask like what types of invitations have you purchased from me in the past? That's a great opportunity to add that question back ends. So feel free to edit your survey. It's a live document. Don't worry about making sure that it's set in stone, you can always add it right before you're sending the survey if you need to make a few tweaks here and there. So the most common mistake people make is like not utilizing logic at all. So you're already getting a step ahead of the competition in a sense by even considering utilizing logic. The most common situation is, someone writes an entire survey assuming that the audience that they're sending it to is completely different than the one that they're getting. So sometimes you want to confirm that the audience that you're sending the survey tool is the audience that you're looking for. So thinking about the time-bound example from earlier, where we're asking customers for the satisfaction survey for the Etsy store owner who purchased within the last six months, maybe you want to confirm at the top of the survey that they indeed did purchase within the last six months if you're unsure about your records about your customers. So in the process, you'll see in the document that I have here is I have all the questions numbered, and it's important to keep them numbered because this is what's helpful for programming. So if you scroll down to question number eight, I'm asking about how would you rate the shipping process? Because remember back to one of the sub-goals that I had is actually, and some of the feedback that I got on my Etsy store is that some people were a little unhappy with a shipping. You can see in this survey for question that we've asked, how would you rate the shipping process? Excellent, above average, average, below average, or poor. So for people who've responded below average or poor, I really want to follow-up with them on why was it a bad shopping experience? So this is where logic comes into play. If they have had an excellent shipping experience, then I really don't need to ask them that follow-up question, and that makes that sure but just a little bit shorter for the respondents and also more relevant for them. So you can see here in the doc that I've actually written "Skip to question 10 for excellent above average and average. " The reason why I do that is so that I know when I'm programming the survey that these three answer choices are not going to get the next question, which is question nine, they're going to be skipped over it and moved on to question 10, which is the next question that's relevant for them. The reason why it's important to have logic is two-fold. Number one, you may not be the one programming the survey. So sometimes you might be collaborating with a colleague or maybe you're working with an agency to do this survey, and so you need to be very clear with them on how you want the logic to be set up. So that's why it's important to write it down. The second thing is going back, and if you ever need to review what you did. It's sometimes hard to think about even what you did last week. Let alone if you think about what you did last year. So that's why it's helpful for just retaining some of that memory about "What did I do in my last customer satisfaction survey?" If you've had the logic written down, it's easy to reference. So the other thing is it also reminds you when you're programming your survey to make sure you put it in that logic, and we'll go through some of that process in the next step when we show you how we might go about programming. 6. Creating Your Survey: So now let's actually go through the steps of programming your survey, and you'll see why I recommend you to write the logic in the document. So when you're in survey monkey, or the survey software of your choice, the features and functionality will be slightly different, but most of this will apply. How do you want to get started? I want to start from scratch because I already have my survey written. I'm putting in my survey name, and my questions are already written, and then what it does it it pops up this modal where you can copy and paste your questions. You can see on the small preview on the side just to make sure that everything looks good, and then add questions. So once you've copied and pasted your questions in, you still need to do at some slight modifications, because most of the question types it detects is the open-end multiple choice and the comment boxes, so you'll need to make a couple modifications here. So one of the first things is the NPS or Net Promoter Score question. This is a question on a scale from 0-10. Luckily in survey monkey, we already have a NPS question type. So all you need to do is you need to find the question type, and then you just need to drag and add. Once you've added the question, you just need to customize the name, and the name of this company is called Paper Boutique, and then press save. You can see it automatically populates that NPS question, has all 11 points on that scale, and now you can delete that placeholder that you have for that question. Moving forward, you can see that everything else looks good with the second question, but the third question should be a checkbox question. So all you need to do is select "Edit," and then you just need to change the question type here. So you need to change the question type from multiple choice two check-boxes. That's all you need to do and press "Save." But here, as I'm looking at the questions, you can see that sometimes you might want to make an adjustment on the fly, and I'm realizing here that actually I forgot a "None of the Above" option here. So that's one of the things that commonly happens with a checkbox question type, is that you put all the answer options that people might potentially check off, but then you forget something like "Does not apply," or "None of the above," and that's exactly what I did here. So I'm just going to press "Edit," and now I need to just add in an option that says "None of the above," and now I'm getting into a little bit advanced here. So this is totally optional. But what I recommend is that you go into "Options," and then you actually randomize or flip the choices, and then you click "Do not randomize the last choice." When you have a lot of options like this, you can see that there are quite a fair number of options. You want to minimize the chance that there might be some response bias here, and what I mean by response bias is that people might be more likely to read the top four answer options, or maybe the bottom four answer options, and not read anything in the middle because you've had too many. So in order to get rid of that order bias, you want to randomize the way that they're displayed. So you minimize that chance that there may be only reading the top couple, or the bottom few, and so that's what I just did there. The next couple questions look good because they're all Radio Button questions, and a Radio Button is just another word for this button here where you see this circle. It means it's a single select as opposed to this checkbox, which is in a square button. So a square button means that you can check multiple ones of those options. So that's what really what as checkbox question is, and in this multiple choice question, you'll see that this radio button, sometimes though to call in the industry, is as circular which indicates you can only check one option. Now we are at question eight. So you can see here, this is where I've added in logic and I got it in Skip to question 10, skip to question 10, skip to question 10. But I now have some of these page breaks, and the reason why I've written in page breaks is just to remind myself, hey I wanted a page break here. So what I do is I add the page break, I delete this page break placeholder, I'll add another one here after question nine, and I'll delete this page break placeholder. Going back to question eight where I have the logic. So now I'm going to click "Logic," and then have these people skip to question 10. So page three, question 10, and then press "Save." One thing I'll do now is just edit this reminder out for me, again these are just the reminders, so that you know when your program in the survey, you don't really have to go back and forth between the Word document, and the Survey software. So now that is all that logic is saved. The one thing I want to do here is, this is only a text-box, I want them to write a little bit more. So basically the rule of thumb is, the more you want someone to write the bigger the box should be. So it gives them a visual clue of how much detail you want them to add in your survey. So I'm going to change that from a single text box, to a comment box and press "Save." Then here, question 10 is another one of those where it should have been checkbox, I'm just going to quickly change it to check boxes, and then here I've written "Other, please specify," but it doesn't give us specify box. So what I'll do, is we actually have that as a default option here, where I just need to click that, and I'll auto populates other "Please specify" in order so that they're not duplicated, it just need to delete it from the answer, and press "Save". The rest look good, but wait, I need just this one to be Comment box, and that's it. You're probably wondering why we should have some page breaks in here, and the page breaks are great to use in two instances. The first is to utilize it when you're utilizing logic, otherwise, even if you utilize logic and skip them over a certain question, someone who wasn't supposed to see that previous question could technically scroll up and still view it. The second reason is to give the respondents a mental breather. So if you think about any website, say you're reading the news, and you're reading this really long form article, they're really engaging but sometimes you just need a quick mental breather because it's hard to sit through the entire thing. It's the same thing for a survey. So often what we recommend is to add an a page break in the middle of the survey if it's pretty long, so say more than ten questions, so that you give the respondent a chance as the page reloads, to just reset their mind, give them a chance to just rest for a second, and then move on with the rest of your survey. You should put a page break when you're switching topics. Say you have two different kinds of products, and many of your customers have purchased both products. So a good rule of thumb is, maybe asking about product one on one page, add a page break and then ask them about product two, because then it gives them that time to pause, regroup, and then move on to the next section. Also, one last thing about page breaks, it's a little bit of a good practice to have if you have a longer survey, because what happens on the back-end and the technology side is that that's when the responses for your respondents are recorded. So if you have 50 questions survey, and it's all on one page, and someone gets to 49 questions, but doesn't actually get to the last question and hit "Submit Survey," you're going to lose all of that data. But say you had a page break after question 25, then at least you have part of that data from questions one through 25 recorded, and you only lost the other half. So now it's your turn. So go through your survey, add any logic that you think you might need, and then go over and program it into your survey platform of choice. Now, next step once you've programmed your survey, it's all about branding and testing your survey. 7. Branding Your Survey: Now that you've programmed your survey, it's time to actually brand and test your survey. When you're thinking about branding your survey, it's always good to just have the colors that people are familiar with, potentially your logo, potentially any images you want to add in your survey. It's always helpful to have some visual cues for your respondents so they are remembering that they're doing this survey for your brand. But unfortunately there is a fine balance between having the right amount of brand and too much brand. So what do I mean by that? So think about the logo, for example. If you're utilizing the logo, make sure it's very small, it doesn't take up a lot of screen instead. It's just really up at the top just to remind them, again, you're doing this survey for my company, not a different company. Another thing to remember is when you're utilizing images. Images can really distract away from a survey when really you're asking the respondents for their responses to the questions. So you want to make sure that the images that you use aren't maybe too pretty because maybe if they're too pretty you're going to distract them from the survey itself. So you really want to strike a fine balance there. The other thing about visuals is the questions themselves and the words that you use. If you find that there are huge blocks of text in your survey, that probably means that people aren't really going to get through your survey because what happens, for example when you get a long e-mail, you say, TL DR, Too Long, Don't Read. When a respondent sees a huge block of text, they're either going to skim through it quickly or maybe they may not read through it at all because it's just really overwhelming to see all of that. So the best rule of thumb is just take a visual scan of your survey. Do you have too much text in your questions or too much text in your responses? If so, try to trim it down, try to be more concise without losing any of the meaning behind your questions. So when you running your survey, your thinking about things like your preferred colors, fonts, font sizes, images. But the other thing you have to keep in mind is make sure you keep your respondent in mind. That's really key to the whole time during the survey process as you are creating the survey to keep your respondents in mind. Not every one of your survey respondents will be able to read very small size font or maybe you will be able to see certain colors. You want to make sure that the colors and fonts you're using are accessible by any type of respondent that you might come across. The best rule of thumb in terms of fonts and colors is actually to use the default. So usually companies have already thought about this. So unless the default is red or green, you should just use the default. If it is red or green, consider some other color because that's the most common type of colorblindness as red-green colorblind. Then, when you think about fonts, you want to make sure you're using a simple font. So while those fonts with like lots of cursive or lots of blocks might look cool, they're much harder for people to read. So just use a very simple Ariel or Helvetica or whatever font, just use the default, make sure it's simple and readable. So now let's think about branding your survey. You can see here now that we're in a Survey Monkey, you have an option to add your logo. We're not going to add one today, but you have that option if you'd like to. But you can see here that our default color unfortunately is green. We talked about how it's not super accessible. So let's actually change our default color. So you can see here I have in my theme section an appearance. I can edit the theme that I have. Let's change the colors and I can just easily toggle to a different color. So blue is a nice, safe color. You can see everything gets changed to blue. That's a nice way to make sure that your survey is fully accessible. So the other things that you can do in addition to branding your survey, is changing what is displayed in the survey. So say you don't want to display the title, that says Customer Survey or maybe you want to be a bit more descriptive. You can always be more descriptive, maybe Customer Satisfaction Survey and maybe put Paper Boutique. You can see here, this is asking for a page title but maybe you don't need a page title. So there you can turn off the page title. Another thing that I like to do, and again, this is more personal preference and this is not required, is actually remove the question numbers because it looks a little bit cleaner. So I always go and remove the question numbers. So again, that last step is a little bit more optional, but the point is is that you can really change the look and feel according to what your needs are. Just make sure that in the next step you test it and make sure it looks good on all your devices. So there are clearly a lots of other themes that you can utilize here on the left-hand side that are maybe a little bit more fancy like this work day theme. So we can change that and see the nice image of a computer in the background and the survey is on the right-hand side. It depends on what your needs are. If you think that this is what you're looking for, I tend to like the little more simpler version but it's really a balance of what you're looking for in the look and feel of your survey. So obviously, branding your survey and designing your survey is all up to you. You can use a bit more complicated or more visually appealing design, but I prefer the more simple white background, blue font, or black font that I know and is guaranteed can be excessively viewed by everyone. 8. Testing Your Survey: So after you've branded your survey, now it's time to test out your survey. This is a really critical step for a couple of reasons. Number one, it's a chance that you can actually see how your survey looks on a mobile device. So everything you've been doing up until now has been on a computer. You've created your survey on a computer, you've written your survey on a Word processing document, but that doesn't necessarily mean all of your respondents are going to be responding on a computer. So it's really important that you test your survey on a mobile device because what we find is that, about one-third of respondents in the US will actually respond on a mobile device. A mobile device can be anything from a tablet, like an iPad, to a smartphone like a Samsung Galaxy or an iPhone. If you're doing surveys in international countries or have international customers, that number actually goes from one-third to close to half. So it's even more important to make sure your survey looks great on a small screen. So it's important that you try and test your survey out not only on a smartphone or a small screen but also on your friends because you've probably been knee-deep in writing the survey and seeing these things, and you might have overlooked a very critical typo that you've just seen so many times that you think is actually correct. In addition to making sure that you don't have any typos lingering around in your survey, having a friend or a colleague test out your survey can also make sure that your survey questions are clear to all your respondents, and also that you've programmed your logic properly. Now, let's actually test our examples survey. There's a couple of ways that you can test it. Number one, you can do it in the SurveyMonkey product in the preview and score section. Number two, I also recommend not only going through our preview and score section but also just sending the link to yourself and making sure that it looks great on your mobile device and also your desktop and laptop. So let's walk through those steps. So let's first look at our preview and score sections. So when I move into that section, it does a couple of things. Number one, you can toggle between a couple of different views. So here, I'm viewing what the survey would look like on a desktop, versus what the survey would look like on a tablet, versus what the survey would look like on a mobile device. You can see it allows you to scroll and interact just like it would on that device. We also have something called SurveyMonkey Genius, which looks through your survey and utilizing the data that we've collected over many years and millions of responses. We can actually give you an estimated completion rate, an estimated completion time, and recommendations on how to fix your survey. So that's one of the things that you can look for. Here luckily, it says there's not much to be done here. Many of the recommendations that are in SurveyMonkey Genius are recommendations that we've actually gone through already in this course. So after you've previewed and scored your survey, you can generate a test link so that you can actually test it on your device. So there's a couple of ways you can do that. You can generate it, copy it, e-mail it to yourself. You just type in the URL. You could actually even generate a QR code, and put your phone up to your computer and actually have it direct that way. So there's lots of ways, but the idea is, somehow get the link to yourself and test it on your phone. So let's test out our survey. Here's the survey. I'm going to roll through and see overall how satisfied I am, do I like all my things? It's reliable, it's unique, extremely well, very high-quality, excellent for value, extremely responsive. I'm going to test out my logic here. So if I press, ''Above average", I should not get the follow-up question about shipping. So I press "Next". You don't get the follow-up. That's great. Now, which of the following invitations? That's a lot of birthdays there. It's a lot. So as you can see, while we're testing this out, you can see in my screen right now, I only see lists of birthdays on the types of invitations, and I need to scroll further to get to wedding, baby shower, and holiday party. So this is a situation where you probably want to edit to my survey and just list birthdays or birthday invitations instead of first birthday, 10th birthday, 16th birthday, 18th birthday. So that's one added that I'll make after I'm done previewing the survey. Extremely likely, and then no. I love all your invitations. It's done, and we're all done. So now that we're done previewing, let's actually take a look and change our survey. So let's scroll down to the question that's all the way down towards the bottom and fix this birthday thing. Following invitations, let's remove all of these birthdays, and now we've simplified it from many answers to adding, accidentally deleted one, couple baby shower, holiday party, birthdays. Press ''Save,'' and there we go, much better. That edit that I made was pretty simple. It was pretty easy. But if you're making more complicated edits like at edits to logic, it's always best to retest after you've made those edits. I'm not going to retest since here all I was doing was really removing answer options and adding one, but it's always best practice to test your survey one more time to make sure that everything looks good. Now, it's your turn. So add branding to your survey and test out your survey and grab a friend, and recruit them to test it out too. Now that you've actually completed the design of your survey, let's move on to sending out your survey and determining your sample size. 9. Sending for Success: So now thinking about sending your survey. There is a lot of different ways that you can think about sending your survey. The most common way, especially for an online web survey, is of course email. Then there's multiple ways that you can send out the email survey. Number one, you can just generate a URL, and that's pretty simple. All you have to do is just generate the URL, place it in an email, and blast it out to all your customers. The second way that's a little bit more sophisticated, but what I recommend, is sending it via an email collector. So you can utilize a SurveyMonkey email collector or say you already have a CRM system like malechimp. You can use the SurveyMonkey integration with malechimp to actually send out the survey, so you can track the number of people who've opened up your survey invitation, the number of people whoever responded to your invitation, and then send reminders to only the people who haven't responded, so you're more effectively utilizing the emails and not spamming any of your respondents. There's a couple of other ways. Say maybe you don't have access to your customers email addresses, but you actually have access to their phone numbers. You can actually send them a link to the survey via SMS, or you can say maybe most of your customers are following you on Instagram or Facebook, you can actually post the link on Instagram or Facebook or even Twitter and have them respond in that way. We do also have a Facebook Messenger Collector. So if you are engaging with some of your customers via Facebook Messenger, you could also send out your survey via Facebook Messenger. One caveat there is that the format of the survey needs to, there are a little bit more restrictions, so I would read up on what those restrictions are, we won't go into that detail here, but that is also an option as well. So for those of you who are reaching your respondents in person, there's a couple of different ways that you can do this. So the first way is a QR code. So say you have a coffee shop and you're really looking to get feedback on your offerings at your coffee shop, or maybe it's just as simple as like is my bathroom clean and you want to put a QR code inside the dockers so that people can notify you when you need to clean the bathroom. So what you can do is actually generate a QR code directly in SurveyMonkey, post the QR code anywhere it where it's visible, and post I would do it fairly large, maybe taking up most of the sheet of paper, and then all people need to do is they need to wipe out their phones and point their phone's camera to it. If a respondent has the latest iOS, what it does is it automatically directs the respondent into the SurveyMonkey survey. If they have an android, they'll need a QR code reader or utilized Google lens, but it's a pretty seamless process. Sometimes you'll have people who are not familiar with technology, especially those who might be older or from different countries, they're just not familiar with how to respond to surveys on a digital device. So sometimes what you can also do is print out a PDF of the survey from survey monkey, and then have the people respondents respond on paper. You do have to go through the extra step of taking those paper responses and uploading them into SurveyMonkey so that does take time and I would plan for that if you do want to do that, but that's another option if you mostly deal with your customers in-person. So after you've decided how to send out your survey, you should also think about when to send out your survey, because when you send out your survey, actually does make a big difference. So for example, what we see on SurveyMonkey every single day is that people tend to respond in the morning right as they have gotten to work. So right around 09:00 AM. A lot of people take a break for lunch, so we see that response patterns actually dipped down at lunchtime, and then they come go up and back up in the afternoon, and then come way back down in the evening once people start coming home, once they start cooking dinner, eating dinner, and relaxing. That's what we see as a nerdy way to describe a bi-modal distribution like this, a little bit like a camel's hump. The next thing to keep in mind is that if you send your survey at night, you'll actually get more people responding on their mobile phone. So that's why thinking about when you send your survey what time of day is under survey can have a factor and whether it should be mobile optimized or not. Now, let's turn back to our example. So we're going to go back into SurveyMonkey and stand out our survey. So you can see here, we've done previewing score. We're in design currently. We've already previewed in score our survey, we're now going to move to collect responses. When I click on the collect responses tab, you'll see that there's a couple of different collector options that I have available. I have a web link, email, you can purchase respondents which for this particular customer satisfaction example is not relevant, you can post to social media, you can embed it on your website if you would like to. So if you have a website, it's probably not as relevant for an apps store owners since they're publishing most available there things on Etsy, but say you had your own retail site can actually embed this survey on your own retail site, or you can do manual data entry, which is if you are doing a paper survey, you want to enter in the results, Facebook Messenger, and then the newest collector which is SMS. So those are all options available to you. One of the things that I will walk through is the email collector. So I want to add a new collector and add an email collector. Then what you will mean in terms of having your email collector to set up is a couple of things. Number one, you'll need a list of emails of your customers. If you're running an Etsy store, you probably can export a list from Etsy as long as it's in CSV format, or you can even copy and paste it into SurveyMonkey, but it's CSV is nice because what you can do is you can upload additional attributes into SurveyMonkey that you can utilize to analyze your data. The other thing that you can do is you can customize the look and feel of your survey using custom HTML. So here's the default email template that we utilize and Survey Monkey, but what you can actually do is you can edit the message and use custom HTML. I recommend this if you are very design heavy and you've already designed a lot of your HTML. If you're comfortable designing a lot of HTML code for your emails, once the email has been designed, you can copy and paste HTML code into the SurveyMonkey email collector. There needs to be a couple of adjustments that you need to make in order for it to go through. So I will only would recommend this if you are a little bit more, if you're familiar with HTML. The other thing that I would do is you can see here, it says, "We're running a survey and would love your input. Please let us know what you would think below. Thanks for participating." This is where I would add in a one to two sentence introduction of why you're doing the survey and also personalize it. So one of the things that you can do here, especially if you already know information, is a, Hi, customer's name. So my name is Sarah. So if the email is coming to me, it would say, "Hi Sarah. Thanks for being a customer of Paper Boutique. Really want to know what you thought about our service, please take a minute to provide us some feedback." That's a way of making sure that the conversation again thinking that a survey is a conversation with a purpose, you want to make sure it's as engaging as possible. It can be a lot more engaging than saying "Hey, take my survey." So once I've entered in my recipient list, send it to my research team, and I'll press next, and you have a couple of options in which you can customize. You can customize the email address and from which you're sending it from, your own email address, I always recommend in the invitation tracking and then you can do anonymous responses as well. So I would utilize this is pure collecting particularly sensitive information, because sometimes people won't want to disclose that if they know that you might be tracking it back to their name or their email address, because they don't want that response tag back to them. But for most cases, that really isn't a problem. So I'm just going to click "Next". Then I can either send the email now or schedule it to send out a very specific time. I can also schedule to send my reminders as well. So you can do that all at one point in time and not have to worry about doing it in the future. So for now, I'm just going to send now and I'm set. After you send out your survey, you can expect that the majority of your responses will come in within 24 hours of sending that initial request out. You'll see a small bump between 24 and 48 and then that big decrease after 48 hours. So sometimes, you'll send your survey and you'll see that only maybe five people respond, and you're like, "Oh, no, what do I do?" There's a couple of strategies here. Number one is to do a reminder. So this is where our utilizing an email collector really, really helped. You can send the reminder to people who haven't responded, or you can, if you haven't utilized the email collector, you can just blast everyone one more time, and say, "Hey, please respond." So while the vast majority of responses will come in with your first email requests, you will see around a 10 percentage point boost in your response rate every time you send out a reminder. That's what we generally see in our research. Again, it really varies from customer to customer and from survey to survey, that's what we generally use as a guideline. Another way to boost response rates is to think about offering an incentive. An incentive is basically a thank you for taking your survey and also encourage someone to do so. So might be, "Hey, take my survey. If you do so I'll give you a $5 Amazon gift card." Or, "Take my survey. If you do so, a one person out of a 100 will get a $100 Amazon gift cards." So a sweepstakes opportunity. So there are pros and cons as to what types of incentives to offer, but the number one thing to think about is you want to offer an incentive that does not bias the type of people that you are getting into your survey. In terms of the most effective incentive, it's giving an incentive regardless of whether or not someone is going to complete the survey, that's what's called a universal incentive. That's much more effective than sweepstakes incentive where you enter in all the respondents and you pick one. In a universal incentive basically, you give them the incentive before they've even taken the survey in regardless of whether or not they take the survey. The other effective way is to pay upon completion, but just not as effective as a universal incentive. Explore some of the collector options that you might want. Do you want to send it via email, via text, or do you just want to post a QR code? It's all up to you, and when you're ready, send away. Up next, analyzing and learning from your responses. 10. Drawing Conclusions: Once you've determined that your data is high-quality enough to actually proceed on with analysis, there's a couple of ways that you can analyze your data. The first and most generic is the topline data. That is just the individual results, the raw percentages without any filters or any deep dives attached. So what that is is you can see what percent said your product is very high quality versus what percent said it's just somewhat high quality. You can see how at percent are very satisfied with your product versus what percent are very dissatisfied with your product. That is what a topline level analysis is. The second is you can do crosstabs, filtering, or compares. Basically, what that does is it drills down into one or two different groups. The next is one of the most advanced level where you're looking at correlations and regression analysis, where you're seeing how multiple factors are related to one another. That's something that you need to do in a statistical program like R, SPSS, or Stata or SAS, but those are all options if you want to take that extra level. But for most people, you don't really need that level of analysis, you really can do and make the most business decisions based on the topline level data and based on the filtering and crosstabs. So today, for our survey example, we'll focus on the topline level analysis and filtering by a couple of different variables. So once you've actually done all of your analysis and you really want to act upon your data, there's a couple of things that you should keep in mind. Number one, just because you find that two things are related, it doesn't mean that it's necessarily true, which is basically correlation doesn't always equal causation. So you just have to make sure that sometimes you'll see two things related and just caveat all your data with, sometimes two things that are related not because they actually are but just by chance. The other thing to keep in mind is watch your base sizes. What do I mean by that? Particularly, when you're filtering and when you're comparing data, you might get into really small number of people. So you find that men really love your product and so you're really going to target men for your next product roll-out. Well, if that's only based on two men, you probably want to rethink that before you act upon the data. But if it's based on 100 men, you probably want to go ahead and do that. So that's one example of make sure once you've filtered down into a specific subset of people, you have enough people to actually make actionable decisions upon. Then finally, the results can always change over time. So what was once true may not be true forever. Maybe, you've made some changes in your business, or maybe your competitors have made changes in their business, so those are big factors at play, are external factors at play that may change the relationships that you found today. So now, actually, let's take an example of what this analysis looks like in the SurveyMonkey product. So the first thing you want to do is, number one, refer back to the goals that you've had. Remember, you actually wrote down what your analysis plan will be, so that will actually be very helpful in this stage. The second thing we want to do as a reminder is just make sure that the quality of your data and the respondents to your survey are actually the respondents that you want to respond to your survey. Then the third thing is then to dig deep, take that analysis plan, and actually analyze the data. So let's work through that first piece. So what is our analysis plan? Just as a reminder, we said that we are, number one, examining the topline results; number two, examining differences between those who have purchased multiple invites; and then number three, do my customers who purchased invites for corporate events, like holiday parties, have higher or lower satisfaction rates than those who purchase for personal events. Great. Then the other goals that I had for my survey was, should I change my shipping provider? Do people think that they're getting good value for their money, and do people like the quality of the invitations? All of those factors really roll up into our overall goal of how satisfied are my customers, and what are the opportunities for me to improve my product? So now, let's do that first thing, which is let's make sure that the quality of data that we have or the respondent pool that we have is the right respondent pool. In this survey, we really only asked one question around that, and that's what type of invitations have you purchased? If I look back at the sales data for Paper Boutique, I see that around 70 percent of the customers have purchased a wedding invitation, around 15 percent have purchased an invitation for a holiday party, which tends to be more than its corporate events. So we ask that question, and if we scroll down below and we want to confirm that the people actually match that data. So here we see that of the respondents, 83 percent have purchased a wedding invitation and zero percent have actually purchased a holiday party invitation. So that means that we don't have anyone who's purchased a holiday party invitation represented in this data, so what we want to do is go back to that previous lesson of how do I get people to respond? Well, maybe send just the reminder to respond to people who purchased a holiday party invitation. So you know that they're represented in your sample. So now say you've sent out that reminder, and you want to take a look at whether your respondents actually looked like your customer base. Now that we look at this result, we see, great, 69 percent of my respondents have purchased a wedding invitation, and 17 percent have purchased a holiday party invitation. That's awesome. The reminder seem to really have worked in this case. That's a lot more in line with what my customer base looks like, where 70 percent have purchased a wedding invitation and 15 percent has passed purchased a holiday party invitation. Now, I know I can competently go forward and analyze my data and that the responses are high-quality. Next step is let's examine the top line level results and see if we can answer any of the goals of our survey. So to remind you, the goal of the survey was actually to make sure that my customers were satisfied with my product. Here, I can see the net promoter score is 54, which means that there are more promoters, people who responded nine or 10 on "How likely are you to recommend Paper Boutique to a friend or a family member?" Versus the only 15 percent who said 0-6. So that's pretty good. The vast majority are satisfied with the product, and you can see it's 67 percent who are satisfied and 15 percent who are somewhat satisfied. Sometimes, I might want to lump those together into percent satisfied, and what I can do is do that in SurveyMonkey I can just combine with. Someone I'm going to combine very satisfied in somewhat satisfied into satisfied. So now, I see 80 percent is satisfied, and then I can combine with the dissatisfied, so somewhat in very, and I see that only 13 percent are dissatisfied. So overwhelmingly most people are satisfied with the product. Sometimes it's just too much of a mouthful to say, "67 percent of our customers were very satisfied, and 15 percent were somewhat satisfied, five percent were neither satisfied nor dissatisfied, five percent were somewhat dissatisfied, and eight percent were very dissatisfied." It's a whole mouthful to say. That's why sometimes it's easier to combine the verys and the somewhats to make the story a bit more concise when you're explaining it. Now, moving forward, everyone associates unique and high quality, which is great, and then the third is good value for the money. So it seems the two things that we are looking for in terms of high-quality and good value are the top three attributes that people think about our invitations. But here, it looks like we could do a little bit better on value for money because it's only 28 percent versus 49 percent who say high quality. So we might want to think about doing something there. Most say the invites exceeded their expectations and the quality of the invitation similar to that attribute question seems like it's very high, at least according to the respondents. But one thing to note is when we go to the shipping question, you see that a lot of people have actually said the process was average or below average. Now, when we look at the shipping question, which is one of the key questions we wanted to look at, we see that 23 percent say the shipping process is below average or poor, which is much higher than the 13 percent overall who are dissatisfied with the product. So right now, we already concede because we used five-point scales in both the shipping question and the overall satisfaction question, we can make sure that they are a little bit more comparable, and here it shows us that shipping seems to be a really bad point for us. Now I can actually look and dig into those open-ended responses and see, "Invitations came a day later than expected," "Get tracking. It was frustrating not to know where my invitations were." "My invitations didn't come in a timeframe I expected which delayed my timeline for sending it out," "I need faster shipping options," "I ordered invitations from my company holiday party and they barely made it in time." So it's a lot about speed. So this is where some of the insights can come about analyzing, no I know that for Paper Boutique, one of the improvements they can make is actually improving the speed in which the invitations come to the customer. Next, I actually want to dig deeper into that other question I had, is people who purchase my holiday party invitations or more corporate customers, are they different from people who've purchased, say wedding invitations, who tend to purchase invitations for personal events? So I'm going to filter my data and compare between those two factors and see what I get. So now, I'm comparing those who've only purchase wedding invitations versus those who have purchased holiday party invitations, and already, right off the bat, I can see people who purchased a wedding invitation, their net promoter score is 88, versus people who have purchased a holiday party invitation, their net promoter score is negative 67. So that's really bad. There's a lot of improvement that I need to make with my corporate customers versus those who might be purchasing a wedding invitation. You can see that that trend continues with overall satisfaction with Paper Boutique, where people who have purchased wedding invites are very satisfied while you can see there's a high proportion who are dissatisfied on the holiday party side. You also see a very big difference on the shipping question here where close to the vast majority of people who have purchased holiday party invitation, say the shipping has been poor. So it seems like the shipping is a big factor for those who purchased holiday party invitation. So one thing to keep in mind is that there's only six corporate customers who've responded. So that's a little bit of a small sample size, so that's a little bit of a caveat that you want to keep in the back of your mind, but based on these six responses, all of those six can be very unhappy with my service. So they are potentially a group that I really want to improve the service to and my business to because they have the worst perception of Paper Boutique. That's an area where, number one, you could do a follow-up survey for, or number two, maybe you want to do an email campaign and trying to bring them back, that's more targeted to these corporate customers, or just number three, think about different ideas to engage them. So a couple of things that I can do to act upon this data. Number one, I think it's very clear. Overall, people really like Paper Boutique. They think that the quality is high, that the invitations are unique. I could do a little bit to improve the value that they're getting, but overall most people are satisfied with that. The biggest thing that I can do with Paper Boutique to improve that perception of our business is actually improve the shipping, particularly around the timeliness of the shipping. Now that we've walked through the entire survey process, from soup to nuts or from nose to tail, depending on analogy you prefer, you can really see going back to the goal that this survey really actually helps us answer that question which was, "Are people satisfied with my business?" Yes, and, "How can I improve my business?" Get better shipping. So really the good thing about this survey is it really provides good guidelines of how you can act next to be a better business. So now that we've walked through this example with Paper Boutique, it's now your turn. Hopefully, you have some results. You can play around with it, or even if you don't have any results, maybe now is a good opportunity to revisit that analysis plan since you have a little bit more tips and tricks on how to do that analysis. So get going. 11. Final Thoughts: Congratulations. You did it all. You planned your survey, wrote your questions, programmed in your logic, sent out your survey, collected all your responses, and analyze your data and actually made a decision based on that data. That's a lot that you just accomplished. So congratulations. But it doesn't actually stop there. There's actually a lot more you can think about doing after you've collected your data. There's a couple of tips. Number one, document retention. I know that sounds super boring, but it's actually really important because what often happens, is you want to look back at the numbers and you want to understand why, how you got to those numbers. The question is that you use, the numbers that actually, that the actual numbers numbers and sometimes not all of us have a photographic memories. So we actually need some documents to reference back. So make sure you save all your materials somewhere so you can always access them. Second, think about how regularly you want to send out a survey. In this example that we walk through, we did a customer satisfaction survey. You probably want to think about doing that on a regular basis, because just because you're customers are satisfied now, doesn't necessarily mean they'll be satisfied going to in the future, or maybe your customers are not so satisfied and you've made some changes, and you want to see how those changes are actually improving satisfaction over time or hopefully they are. You can only know that if you actually measure it. So now, think about doing or repeating your surveys on a regular basis. Lastly, based on some of the results that you found, maybe you want to do a follow-up survey. Like we mentioned in our example, a follow-up survey for me, could have been I could follow up with those people who are those holiday party invitation customers only, with a separate survey on how we can really make our offering better to them. So make sure you're also thinking about those three things in addition to wrapping up your survey. I hope you all feel more confident on how to conduct your next surveys, and also utilize it as a tool to connect more closely and improve your product or service for your customers. So if you've been following along, I love for you to share any examples that you might have, your projects, your surveys, your data, anything in the project gallery below. Make sure you ask any questions that you might have or join the discussion as well. Thanks for joining. It's been so much fun talking to you guys about surveys and everything that I do. So now it's your turn, go and spread the survey gospel and go and conductor your next survey. 12. Explore More Classes on Skillshare: