Please ensure Javascript is enabled for purposes of website accessibility Qual360 Session: Understanding the Consumer World at the Speed of Online Reviews
Book a Meeting

Qual 360 Session: Understanding the Consumer World at the Speed of Online Reviews

Laura Kegley: Hi, everyone. Can you hear me okay? All right, Ido, do you want to start off by introducing yourself?

Ido Ramati: You first. Ladies first.

Laura Kegley: Um, hi, everyone. I’m Laura. I’m CRO of Revuze. We’re just out front at the end of this and we’ll give you a little bit of information before we, uh, end the presentation. Um, but I’ve been at this company for a year. Prior to that, I was in social listening for eight years, so I’ve spent a lot of time in the consumer analytics world. Um, I’m very excited to talk through product reviews and moving at the pace of your consumer.

Ido Ramati: Well, I’ve been at the company 12 years because I’m the founder, so probably the oldest one in the company. I founded the company back in 2011, something like that, and I’m very happy to be here with you guys tonight, er, this evening. And then we go to a nice presentation that demonstrates the capabilities of online reviews, what we’ve been trying to do in the last 10 years.

All right, so, uh, what I kind of want to start out with is why reviews are so important. One of the things that we’re hearing a lot from retailers and from brands is how to be a data organization, but it’s changing to how do I be a customer-led organization utilizing data. Historically, reviews data has been thought of as a small data set. Uh, you could only get the overall rating, you know, you have a one through a five on your product rating, and you might be able to comb through a couple of comments to understand general sentiment. But for the most part, you haven’t really been able to dig into that information and truly understand it.

Laura Kegley: So, quick question for everyone. Raise your hand if you read product reviews before you buy something. Okay, so, uh, what percentage—does anyone guess what percentage of people do read reviews before they buy a product? Anyone know? All right, 92%. So, when you’re starting to think through being a customer-led organization, understand what your customers are saying. You now have a data set that is the closest to revenue because 92% of your customers are reading these product reviews before they’re buying the product. This is a huge number. For those of you who are shopping on Amazon, that’s probably where it all sort of started, right?

So, what we want to do is we want to talk through why it’s so important to understand the difference between that four-star rating and that five-star rating. And what I mean by that is, you know, here you have, uh, one sentence, one review, but every review has about three times the opinions, right? So, when you’re looking at something overall and you’re seeing this is a general 4.5 rating, it’s the individual sentences that you want to understand the sentiment behind them. What’s the opinion that’s being said that’s making someone give it a three or a four and not that full five? And the other reason why this is so important is that these are the people that are actually buying your product. So, up until now, a lot of you might have been making decisions on opinions of your product and brand, but you don’t actually know if they purchased it. But the review—you have to buy the product in order to leave it, so it’s a really important thing for you to be understanding and breaking down. And at 92%, know that everyone is really looking at these.

So, why does it matter when it comes to speed? And, uh, I’m going to have Ido talk through a little bit about that because, again, it seems like it’s a small data set that people only go online and shop every once in a while, but in reality, that bubbling up of information—what do you see in aggregate that’s making something a four-star rating or dipping into this three-star rating—that’s happening on a daily basis. So, I’ll let Ido take it on why, um, speed is so important when it comes to the consumer world.

Ido Ramati: Well, yes. So, first of all, we can identify two main dynamics in the consumer world. The first one is how fast markets are changing and evolving, how many new products are launched every year. We used to be in a world that it’s all about physical shelves, right? It takes time to put a product on the physical shelf, on Walmart, on Target, and if you’re a brand trying to understand your competition, all you needed to do is to go walk into Walmart or Target, whatever it is, as an example, see who’s standing next to you on the shelf, and boom, you have the answer who is competing with you. But today, it’s all about digital shelves, right? Putting products on digital shelves takes minutes from anywhere in the world at any age. I’ve heard about children that are doing that or selling online, and as brands now trying to understand who competes with me, who is the next big thing in my category, now I must find ways to control the entire digital shelf to understand who is the next Dollar Shave Club of my product category, right? So, it’s all about how fast I’m able to understand who is on the shelf next to me, right, on the online sources. And that’s why it’s very, very critical to be able to understand these changes as fast as they happen, right? If I’m only going to hear about it three, four months after that, that’s going to be too late for me as a brand to compete with.

The second dynamic is how fast consumers are changing their minds, their likes, their dislikes. I’ve been around online reviews for more than 10 years, as I said before, and when we started off, we used to see a lot of comments inside reviews that are about, “I’m a third generation for this detergent,” “I’m a second generation for that product,” “My grandfather used this razor,” “My dad used this razor, so do I.” Obviously, you do not see this kind of comments anymore. In the last few years, no one is loyal to the brand too much. In fact, 54% of us will change the brand they’re using after just one negative experience, and that’s how loyal they are to us. They’re changing their minds on a daily basis, and this is where speed comes in again. So, we must understand the likes, the dislikes, what do they like, which smells, which color, which texture we should use in the product, which claims we should put in our marketing campaigns. It’s all changing on a daily basis, so this is where speed comes in again because we must understand all of those dynamics as fast as they happen. So, this is why speed is very, very critical, and online reviews support exactly these two motions. Product goes to sell online, here I have it, I know that it’s there. So, this is the first challenge. The second is that immediately as it goes up, online reviews will start to be written on that product. All the experiences, the good things, the bad things, what do they like about it, what they didn’t like about it. So, if you know how to read those reviews, we’ll talk about it later, how we’re analyzing them, you have been able to work and act in the speed of new products coming into the market and how fast likes and dislikes are actually changing. Okay, so this is why speed gets very, very critical for the consumer world.

Laura Kegley: And in addition to the speed, one of the things that you want to make sure that you’re able to do is benchmark, right? Insights come from comparison, and a lot of times with reviews data, you’ve been able to get your own but maybe not your competitors and certainly not by source, being able to compare the two. So, a lot of times what you might see is, say you sell mac and cheese, and mac and cheese has a really high sentiment. Everyone loves mac and cheese, right? So, what does that actually mean if everyone loves it? How do you know that it’s something that you’re doing differently? When people are shopping for mac and cheese in the store, they might be comparing you to other brands, but what happens before that when they’re picking the meal that they want for lunch or dinner? How do you know about the benchmark between yourself and the total category of meals or frozen foods? It’s that ability to pull from the category so that you can begin to really do that comparison. So, Ido, talk to us a little bit about why that’s so important to understand your competitors in the market.

Ido Ramati: Well, yes, so category insights—that’s the next level of online reviews analytics. That’s the next generation of what is out there today. It’s very easy, quote-unquote, I might say, to find products on Amazon or Best Buy or Walmart or Target or Sephora and read the reviews and then understand a little bit what consumers are feeling towards the product. It’s very easy, I would say, to do that. You can even use some technology to analyze it automatically for you, putting aside the quality of that. But it’s something that can be done. You go one product after the other when you do the analysis and get a sense of what consumers are feeling. But this has very big limitations as well. The first one, we can only do it for products that we are aware of. Can I search for a product that I do not know exists, right? So, this is already here with some limitations. The second limitation is, once I’ve been able to identify the products, what happens if I want to go one level up? Assuming that I want to find the real satisfaction around the camera of the iPhone 15. Well, hundreds of them are sold online all over the place. You have many sellers that are selling iPhone 15s, and if I pick one or two or three of them by going through Amazon or Best Buy and looking and reading reviews, will I get the accurate picture? Obviously not. I need to analyze everything.

So, here’s the problem. If I need to point to the relevant SKUs that I want to analyze, I can never get a complete picture of what consumers really feel. And this is where category insights are coming in. Category insights is the way of collecting data, not analyzing data. It means that you’re collecting all smartphones that are out there, all shampoos, all running shoes, every single one that is sold online. You’re able to collect it and bring it to your database for analysis.

Now, having this capability unlocks a lot of research capabilities that do not exist today. The first one is to create a category benchmark. If you want to understand what is the average satisfaction from a camera of a smartphone, how do I do it today? If I need to select one product after the other, will I check only the leading four brands and see the satisfaction? We have thousands of different SKUs out there from different brands, different products, different countries. How do we get to the true value for that if you’re only able to do it one product after the other? Category insights will allow you to get a whole category level view because you can analyze the thousands of different SKUs and group them together to understand, “Okay, here is the real average satisfaction from a camera of smartphones.”

The second capability that it brings us is to ask questions that researchers today cannot ask. Which headphone has the best noise cancellation feature? How can we answer this, right? If you’re going to do any way of survey or focus group or give consumers to try, you’re always getting two types of answers that we always get. If you ask the regular consumer which headphone has the best noise cancellation feature, they will probably name a product belonging to one of the biggest brands, the top five, the top seven, right? This is what they buy. But if you ask a musician, they will give you a brand that you never heard of, right? Because they know they’re the expert. They’re getting to the brands that are more high quality. That’s where it gets confusing. Okay, who do I listen to, to the musicians that understand audio or to regular consumers? Once again, with category insight, you’re able to mix them together. You’re able to measure both the ones that are for professional use and the ones that are for regular use. Then you can apply some weights, and here you have the answer. Which headphone has the best noise cancellation feature? Which running shoe has the best foot support? How can we analyze it on hundreds of thousands of category insights? It’s exactly for this.

Laura Kegley: Um, and one of the things that we hear a lot about, which I’m sure you guys heard a lot this week, is generative AI, right? And I think that what people want out of generative AI is a recommendation engine, a conclusion of what you should be doing with the data. There’s definitely ways to be presenting you with a summary, presenting you with insights, but in reality, what you really want is something that gives you enough insight that you can start your internal process as the expert in your brand and in your company to begin to understand, do we take action on this or not? If you’re asking a tool to do that for now, it’s really not going to give you the right answer, right? It might give you something that you’ve already thought of or isn’t going to work at all. So, having that category insight is knowing what are the signals in the category, what are the strengths and weaknesses of my competitor, what are the strengths and weaknesses of my own product. So, being able to handle that truth of what your customer is saying is really important to make those decisions until that recommendation engine and conclusion is coming up.

One of the things we also hear is how difficult it is to take this kind of data and implement it into your everyday. So, I wanted you to talk a little bit about the beauty brand that used our reviews platform in order to, uh, implement that and make some really cool decisions.

Ido Ramati: Yeah, so actually, that’s a very nice example as well. I’ve been working with a—our company was working with one of the biggest hair coloring manufacturers that is out there. Unfortunately, I’m not going to use it for too long, that’s what life is all about. But we’ve been working with them for the last few years, again, one of the biggest brands. And they did a relaunch to one of their most successful top-selling products. I don’t know why they decided to do a relaunch for it, right? Great decision. Totally a disaster. Sales went down, reputation went down, complaints got much bigger, complaints, a big buzz around that. The relaunch of that product was a total failure.

Now, what did they immediately do? This brand immediately panicked. That means they hired research firms, they hired strategic firms, they started to do surveys and focus groups and interviews and give this product for people to try, to ask questions, and they started this whole process of understanding how did this fall and to understand what was wrong in order to fix that. Lucky for this brand is that they had one team, I think it was four or five people only, monitoring social media and online reviews back then. The same four people used to do both. Today, with brands, you have huge teams that are doing only this or only that because it became so popular. But back then, there were only four of them. Social media didn’t reveal anything because social media was mainly about hashtags, the brand name, and how many times it was mentioned, good or bad. It didn’t say anything about that specific SKU, cannot tie to specific SKU in social media. But they did find a lot of online reviews on those specific products. Two weeks after the relaunch, that immediately revealed to them what went wrong. Immediately, they took this information internally and built the whole strategy on how to fix the problem. I’ll tell you later what the problem was, and they came up with a whole plan, including TV ads, on how to fix this. Fast-forwarding three and a half months, all the research came back with exactly the same answers that this team gave, but now they were already ready with the plan on how to fix that. Okay, so this gap—two weeks versus four months—this is exactly going back to the speed thing, is only because online reviews were able to populate exactly what was wrong with this relaunch.

And just not to leave you waiting with that, the problem was with the instructions on how to use the product. It required different types of mixing before you apply it on your hair, and it just didn’t communicate it well, and everyone that used it got a disaster. So, that was the problem that came up in the online reviews. And this is why online reviews is the place where you can get the fastest responses to whatever is going on with your product, or your competition, by the way. It doesn’t have to be only on your own. So, that was a great example of how online reviews were used in order to find faster solutions.

Laura Kegley: Um, and I think one of the things that’s so interesting about that too is that a lot of times when you think of reviews data, you think of it as a lagging indicator, but in reality, it can be a leading indicator. Um, and so, we have some interesting stories about how you can use this kind of analysis to start to plan out how to avoid being the person that falls behind in the market. And it goes back to how quickly getting that information, because if you’re using focus groups, NPS, any sort of survey forum, it’s going to take you a long time, but it’s also usually at that point going to be a lagging indicator of the decisions that you can make. So, if you can go through how you can use this analysis to lead instead of let it lead us.

Ido Ramati: Yeah, so it’s all about leading the witness, right? So, when it comes to online reviews, pretty much it’s the, I would say, even the only objective view of consumers. You don’t lead the witness. You don’t ask questions. You get answers only on where your flashlight points at. They talk about everything, anything. One example I can share with you—toilet paper. Quite a basic product. You’re laughing. How many topics of discussions do you think exist on toilet papers?

You sure you want to share with us this? Okay. Anyone can guess how many topics of discussions are inside online reviews on toilet papers? No, not how many reviews, how many topics? Sorry. Forty-two discussion topics on toilet paper, right? If I gave you half an hour, you would probably come up with 10 or 15, right? Forty-two discussion topics from texture, size, perforations, flushable, can I use it for pets as well? I don’t know what consumers are actually talking about. We found some crazy stuff. But forty-two discussion topics, most of them we couldn’t expect that consumers care about them, right? And that’s exactly where online reviews is that no one is leading the witness. They will write about anything that they want. And again, 10 years in this business, I did more than 900 different product categories. Believe me, vacuums are used for things that you haven’t even thought of. So, this is what I think is most fascinating about online reviews, is that you can expect everything, right? Assuming that you have the right technology and you know how to surface it up.

Another great example—do you guys know what’s the most important thing for consumers with smart TVs? Yes, any guess? No, you don’t watch TV too much. Okay, it’s the remote control, guys. That’s what’s important to consumers. Who would’ve thought? Not the picture quality, not the audio. It’s probably more or less the same around all television sets. Remote control is the number one most discussed topic in online reviews around televisions. Imagine now, as marketeers, as researchers, as product guys, what we can do with it. And this is a good example of how the data can lead us instead of us leading the data. And if you know how to measure trends over time, then you’re getting right to the point.

Laura Kegley: All right, going to go back to quiz time. So, when people are on the page, they’re about to add a product to the cart. Something to keep in mind is that they’re sharing that very precious real estate on that product description page with reviews data, right? So, they might not remember in that moment what was posted on social, what happened on TikTok, what influencer was talking about your brand, but what they do see is a really negative review. So, question to you all: what percentage do you think of people who actually consider reviews in order to purchase the product, to add it to the cart and make the conversion? Any guesses? Who actually take that in? It’s 88%. So, what that means is out of 100 people, if you have 92% that are reading the reviews and 88% of those people are actually making a decision based off of that, you could lose 81 customers just from one review. That’s a lot of customers. That’s where it’s directly tied to the revenue. This is why you need to make sure that you’re paying attention to this reviews data. People are taking it in the same way that they get a recommendation from a family member or a friend. And one of the ways that we’ve been seeing a lot of brands make the most of this real estate is by syndication and incentivizing. That makes data really dirty because you don’t know if the sentiment is because they got it for free or if the sentiment is because they truly, organically, want that. And you also don’t usually know if that great review that was on Target was just syndicated to Walmart to make the product look better. So, just want to talk through why it’s so important to have something that can break up unique, uh, organic, syndicated, or incentivized review, and what the challenges of that.

Ido Ramati: Well, dealing with online reviews, if you’re already convinced that it’s a very powerful source of information, you must have the ability to cut through the noise, right? Cutting through the noise with online reviews requires you to do two main things. First of all, you must know how to handle a lot of data manipulations that are done on websites. The more online reviews the website has, the more it’s going to sell. So, they’re boosting the numbers by duplicating reviews, taking the same review and copy-pasting it on different products. You need to know how to de-duplicate that, otherwise, you’re counting the same review multiple times, although it’s just one person. The second one is review sharing. So, those of you that are buying lipsticks online on Amazon, as an example, you have different colors. All the reviews for the red color are exactly the same for the pink one. Okay, you need to know how to divide them into the right variants. Otherwise, you’re reading the wrong reviews. And for researchers, it’s very critical because we all know that data integrity is the key to accuracy. If your data is not organized well, you’re missing there. So, that’s the first one, how to overcome data manipulations websites are doing in order to clean the data. The second one is how to handle different types of reviews, right? We want to be able to flag out reviews that were incentivized, those that received something in exchange for writing their reviews. They would tend to be by 15% more positive. If we are doing real research, we should remove them out. We should be able to understand syndicated reviews, those that are written on one website and copy-pasted to different ones. If you want to understand the real sentiment of Walmart shoppers versus the ones that shop on Target, we need to remove all reviews that are written elsewhere and be left only with those that are really written by real shoppers on Target or Sephora or Walmart or whatever it is. So, this will be another noise that you need to cut. There’s a lot of noise there, a lot of mess around online reviews, and in order to get real research to the right answers, cutting through the noise is one of the key challenges that needs to be taken.

I feel like that just scratches the surface on a technology that is specifically built to organize and analyze product reviews data because this is a bit of a newer tech that you’re seeing that’s outside of your typical consumer analytics space. So, I just want you to talk through why you need something specific for online reviews.

So, usually when I speak with brands or prospects that are trying to build their internal tools for online reviews or using some sort of a vendor, the first slide that I’m showing them, the analysis part, the sentiment and topic classification, that’s only half of the challenge. The second half is arranging and organizing the data in a way for the analysis to work. And that means I must be able to merge different taxonomies into one. If I want to understand what consumers feel towards shampoos, I must understand where shampoos are located on hundreds of different websites and collect all of them to bring them to one place. And then, God forbid, I shouldn’t have any conditioners sneaking inside this group of products because that will skew the results as well. So, we must know how to clean out all the conditioners, the one-in-twos, the three-in-ones—they have four-in-one already—a four-in-one shampoo or something, I don’t know. Usually, whatever. So, we need to filter them out and put them in their own categories, right? So, cleaning the data, cleansing the data, organizing it, that’s the first step before you try to analyze it. Only then do you get into the analysis part, usually done with AI and machine learning.

If you have some experience with AI, you know that AI is not really artificial intelligence; it’s armed with individuals training machines how to do the work, which is basically there as smart as the people who train it. So, you must have technology that knows how to do it automatically as well in order to get to the real sentiment, to those 42 topics that we’ve mentioned. And only then, after you’ve been able to do that, now we want to start to do research on top of that, right? We want to do SWOT analysis, you want to do driver analysis, you want to do other types of research.

So, even if I got an answer that 80% love the camera of iPhone 15, what does it tell me? Basically nothing if I do not know the benchmark. Is it good or bad, 80%? That’s where research comes in. So, you must find the ways also to automate that step as well. So, from data collection to final visuals, that’s what online reviews can offer you.

Laura Kegley: So, one of the things that we also start to hear from our brands and from the retailers that we work with is that the two main challenges are understanding their customers’ journey. How do they go about the awareness of the product, then the education of the product, and then actually purchasing the product? And the second one being, how do you get repeat consumers? It’s known that two out of three consumers are one-time buyers. So, really understanding why your consumers are buying in the first place and maybe not buying a second time is becoming extremely critical.

What people have used as a tactic prior has been syndication, incentivizing: give someone a free product, get them to leave a really good review. But the good news is that you can actually improve your reviews and your ratings by just flat out listening to your customers and making those changes. So, one of the quotes that I showed earlier, you can do things like update your product description page to talk about the benefits that people are speaking of, or you can update your product and make it cover all of the things that people are saying.

In our tool, you can actually see, “Are people saying I’m going to shop again?” or “I’ve recommended it to a friend.” So, you can really start to understand where they are in that journey and also, are they going to return to your store or to your product and buy again?

So, this is where we come in. What exactly is it that we do? We’ve probably covered it for the most part, but essentially we are collecting, we are, uh, uh, we’re collecting, we’re cleansing, and we are analyzing those product reviews. And you’re able to zoom in and out. So, everywhere from the category, the brand, the store, and the SKU level. And out of this, you start to get that true, that truth from your customer of why they’re buying your product and why they continue to buy your product.

So, something like this is a quote that is around losing hair, is the topic. And one of the things that’s so important about this is that we do not have customers come to us and say, “These are the brands that we want to follow, that we’re up against. Here’s the terms that we need to be watching for, or here are the SKUs that need to be uploaded to your tool.” We say, “What’s the category and what sources do you want to pull reviews from?” The tool will tell you what terms you need to be paying attention to. So, there’s no query building, there’s no needing to think through that biased information, “What do I need to be watching?” It will tell you through breaking down a review like this one.

There’s, like I said, multiple opinions within the review. And overall, this product actually had a 4.7 rating. So, what you want to know is an aggregate. Is this one review going to cause people to not buy? But how big of a problem is this, really? Do you need to start spending money to improve the product or work on some sort of potential crisis? And when we’re talking about those discussed topics, we’re going to break it down between the negative and the positive sentiment, but we’re also going to break down every opinion within it. We’re going to tell you what the benchmark is to the category or to select brands that you want to compare it to, and we’re going to tell you what’s driving the one, two, three, four, and five-star rating. You’ll know the exact topics that are causing reviews and bad reviews in aggregate.

And so, you can get down to that SKU level. When you’re looking at reviews, yes, you want to see the category, but if you’re looking at two different earbuds and you’re trying to figure out why one is better than the other, you can do a SWOT analysis very quickly to actually understand what the strengths and weaknesses are of each of these products. So, getting down to that SKU level is really important. You can also get down to that SKU level at every single store it’s sold at because maybe that product is doing better at Target than it is at Apple. It’s important for you to understand that, and you can begin to compare across multiple products.

So, this is a lot on one page, but basically what you’re seeing is all of these competitors and you’re seeing the industry. So, immediately you can benchmark yourself against the industry and your top competitors. You can see overall sentiment, the volume of reviews, and the star rating, but you start to see the topics. And again, these are not topics that are created by your team. These are topics that are being generated from the reviews data, and we’re telling you to pay attention to these topics. And you can start to see how does that topic compare for you versus the competitor versus the overall category. And you could break it down by source because maybe your overall satisfaction, again, at Target is much better than Walmart. So, how do you focus in on your partner with Walmart and make sure that it becomes better? What kind of moves can you make in product description, like with your example? How do you improve your instructions, or do you go back internally and say, “We’ve got to redo this product.” So, there’s a lot of different options that you can do with it.

Um, there’s even more things that you can see, slice and dice the information. It’s important to us that you start at the category, but you can zoom into the SKU, you can zoom out, um, and you can slice and dice the information as many ways as you want to, to really be able to get to that customer truth. If you guys have any other questions, um, we’ll take them, but also know that we can give demos and we’re just right outside at Booth Five.

Host: Great, thank you so much. We have time for one question. I saw your hand first.

Bridget: Oh gosh, the pressure. Um, hi, my name is Bridget, and Warner Brothers Discovery. Really interested in your topic today, and I was wondering maybe how you balance out niche communities versus general communities. For example, Goodreads having a much harder space in terms of ratings versus like Amazon Books or Letterboxd versus Google, where you have maybe a deeper or harsher critique for certain communities, and the fandoms that they’re really in.

Laura Kegley: I mean, I think that when it comes to the different types of sources, one of the things we want you to be able to do is analyze within the source and then analyze all of the sources. So, what you would want to do is first see overall on Goodreads what’s the positive and negative sentiment towards this category, and how does that compare to Amazon? How does that compare to these other sources? And you can either dig into that one source to really understand why it’s more negative, or you can begin to look at all of them. So, there’s not necessarily a weighting system as much as that ability to really understand what exactly is happening on this source and how can we fix it there. It might be anywhere from just harsher critics. It’s not necessarily that you need to be making changes, but you do need to know, here’s the sentiment we usually get from Goodreads, and if as long as we don’t dip further, we’re probably okay. And so, it’s more about having that ability to compare the sources to know where I am to begin with and where do I go if it’s going up and down. Do I need to pay attention to it? So, that’s where an alerting system would be really good because you can say, if on this source sentiment drops 10%, let me know. We already know that it’s low, but now it’s really low, so we want to pay attention to it. Did that answer your question?

Host: Wonderful, thank you so much, Lauren, and Ido we’re going to call Ronnie up now. Thanks so much.

The World’s Most Influential Brands Trust Us

  • Char Broil
  • COTY
  • Dolby
  • H&S
  • Hoover
  • J&J
  • LEVIS
  • Loreal
  • Lubrizol
  • MC
  • P&G