Name: From Search Signals to Business Insights: A Live Walkthrough
Uploaded: 2026-11-11T16:15:11.957Z
Duration: 1 h 13 min 24 s
Description: From Search Signals to Business Insights: A Live Walkthrough

Transcript for "From Search Signals to Business Insights: A Live Walkthrough": Great. Hello, everyone. Hope you can all hear me loud and clear. My name is Alastair, and I'm in the solutions team here at Cision based in London. But actually today, I'm in Copenhagen with the team here in our offices and, delighted to be doing the second session on the search signal advantage with you following on from the session my colleague, Maria, did with Danny last week. And so in today's session, we've just got thirty minutes together, and I'll make sure that there's time for q and a. And, just to mention, if you've got any questions, either already in your mind from listening maybe to the recording of the the last one if you missed it and weren't able to do the live q and a then or you get questions, from what I'm gonna present now, just type them in the q and a as we go through and I'll I'll make sure we come to that afterwards. And if we run out of time, we can always respond to questions, offline afterwards. And so in today's session, I'm going to be going into a bit more of a deep dive on the way we actually practically get into some of these insights using some of the the the tools that we have here in Cision and Brandwatch. And so mainly gonna be looking at two examples. I'm actually gonna sneak a third in, following on from some of the things that were mentioned, in the previous chat. And I just wanna show in the actual platforms themselves what some of this data looks like. And if we were to start from scratch, a little bit of, advice on how I would maybe tackle some of these types of research questions. And so the key things we wanna look at is how do we uncover, let's say, consumer needs. So we we have a topic that we're interested in, and we want to go in and, get the data. How do we validate it and then act on it. So I'm gonna come back to the slide at the end. So I'm gonna stop sharing my slides and then dive straight into the platform. Just make sure I get the right window open here. Okay. Just one moment. Okay. Alright. I think I found the wrong one there. Okay. Hopefully, you can all see that. Okay. So this was mentioned in the last webinar, the trends of Jabai chocolate compared against Matcha. And so the data that was shown on the screen comes from this environment. This is the Trajan platform. This is the the the youngest member of the Cision family. This is our search intelligence platform where we can look at Google search trends as well as l and m responses. And what we're doing here is we're focusing on Google search volumes, and we've got a tracker over a number of years looking at search volumes for Dubai chocolates against Matcha. And you can see there's this massive spike, in searches for Dubai chocolates. So 8,000,000 searches for Dubai chocolates just in one month, November 2024. So two years ago. It's got got really big, but then it's really kind of faded off. Whereas with Matcha, we see that continued growth over time. And so my colleagues, my colleague Maria and Danny talked a bit about what that kinda looks like and why that might have happened. But I just wanted to say that when you've got a big dataset like this with millions of search volumes, quite often, wait till we're gonna send what's going on is to to break down the many different keywords that comprise these 8,000,000, most... These multiple millions of search keywords, every month and look at it in terms of themes. So if we take, for example, Dubai chocolates, and we have lots of different specific things that people search for relation to to to buy chocolate. But then what we can do is we can categorize those into, different things that people might be wanting to look at, maybe the ingredients or ideas for recipes, or also, very importantly, where to buy it. So looking for real solid search intent. And what you can see, for example, with the... To buy chocolate is that there was this huge search. A lot of this was really just people searching for the basic keyword to buy chocolate or just trying to figure out what it was. They've heard about it. They wanted to to see what it was all about, all the fuss. You know, maybe it's... That's something they'd seen on Instagram, and they wanted to investigate it a bit more. That's what the breakdown looks like. And if you compare that with Matcha, it's a very different story. So this is the matcha search trend, and you can see that same colored bar, the red, is the where to buy. So with all of these search keywords, we're then getting a breakdown within those of different groupings of keywords. And these are all the different searches that are to do with actually buying match up, whereas we saw the Dubai chocolate. It was a very soft, part of the the bars in there. So that's one of the things just to add on what was discussed in the previous session where you want to dive into grouping all the different keywords and getting a... An insight into what's actually driving it. Are people just out of interest searching for it? They're not really wanting to buy it. Or as in the case with Matcha, we've seen this consistent growth. And I would say, if you want to do a comparison, you really need to be looking at the intention to buy and focusing on those. Whereas if you're looking just at, the raw keywords, it's it's probably more just a general interest piece. So I'm just gonna talk a bit about how we might actually get started with the search line. So I've just got a blank project here, and I'm gonna go start... Get started with a category search. And this is where we can get the Google keyword engine as well as some of our own AI capabilities to really help us jump start a discovery project. And so I'm just gonna go and look at Google search volume. So the same sort of process that we would have done with that, to buy chocolate match for example, and I'm gonna go into a a sleep supplement store in a moment. I'm just gonna pick a couple of markets, and then I'm gonna pick this... This is gonna be to do GLP one, products. Do that search. And now I can get... Start getting suggest... Suggestions, and I can use Google's own engine to help me get get started. So I can type in some keywords like Ozempic and just do a search for that. And so what I'm getting back from Google are all related related searches on Google to to do with this. I think I can also add in, for example, a Govee or, GOP one, like that. And then Google is... We're we're fetching this from from Google. We're getting loads of related searches because I don't I don't just want the the the single keywords at the top. I want all the different variations. And so I've got... I I could add these 1,700 keywords into my environment, and I can potentially expand on those. I can I can add a whole bunch of secrets and then get Google to populate that? Once you've got that, the key is to then think about how to segment those. So I'm just gonna show another example that we've mentioned the last time. This is one that we pre pre... Precreated on sleep supplements. So we talked before about the shift in people wanting to search for melatonin versus magnesium glycine. So this is what the data looks like. We've got, 2,000 seed keywords in here created in a similar way to that process I just showed. We started with some very basic seed queries, and then we used the Google keyword searches to expand on that and get us from just a couple of dozen seed searches to a few 100 or a few thousands, keywords to actually track, in Google. And, again, I'm just doing this in English. We can do this in multiple languages. And then I want to start thinking about how I might want to categorize these. And so what we want to think about is specific benefits that might be... People might be looking for, specific pain points they might be searching for. So if someone's searching for sleep supplements, why are they searching for it? And I'm gonna, in this environment, start to try and use tagging of keywords to get a bit of a breakdown in that. And so when we first... When I first set this up, I wasn't really sure what I was gonna get out of it. I just wanted to see what was going on. And I know that, this is, you know, for a piece of client research trying to understand one of their products that has one of these ingredients in it, what trends there are more broadly in the category. And so I've got lots of keywords in here. And, again, I I don't really know what's gonna be shifting of time, but I can use Google search to get me started on that. And, again, I've got these categorizations. So when I go in here, I've got a bunch of different, ways of segmenting things, so I can look for specific topics. So I could see, you know, most of the... There's a lot of conversations about the sleep supplements, so I can take that out. And then I can see what's changing over time. I could... There's... There are AI capabilities in here to produce reports, but I can see, for example, there's there's a growing search interest in red light therapy. Wearables actually on a slight decline, so that's been a thing for a while, but, actually, relatively, people are maybe less interested in getting, apps and wearables to track their sleep. But I can also see some growth in infrared sauna and heat therapy. So just within these segmentations, I can see a few shifts, and this is me to... I'm not really using anything particularly sophisticated today. I'm just using groupings of those, 2,000 keywords, and I can see what's trending over time. But by far and away, most of the search activity is around the actual ingredients in the sleep supplements, whereas, you know, the the conversations... The search one is around the tech and all the other things is much smaller. So I wanna get a breakdown of those ingredients. So here I've got those categorized, and you can see the yellow bar at the bottom is all the searches to do with melatonin. And for a long time, melatonin has been the core ingredients that people think of in relation to sleep supplements. But, actually, that is on a little bit of a gradual decline over the last four years. We we get the Google data for four years. And I can see a big shift in this growth here for magnesium glycine that's gone from a very small volume in February 2025 up to over a million searches a month typically, in the last year. And I can also see there was a massive jump in people searching for melatonin in November 2025. And then from then, there's there's been... You know, so there was a bit of a growth magnesium glycine. It dipped off, and then I've seen this big spike in melatonin. And then from there, steady growth of searches for magnesium glycine. And some of these other ones, if I just take this out, there's been a bit of growth there for some of these other supplements like theanine, and this is magnesium in relation to sleep, whereas, ashwagandha's maybe on a little bit of a decline. So I've got a signal here. Okay. I'm looking at the data. I've haven't done anything particularly sophisticated today. I've just gone to Google, put some core seed queries in, got a load of keywords back, grouped them into ingredients in a category like technology. I've also got some pain points and other things, and I can see something shifted. So I'm looking at this as a consumer analyst, and I wanna know why is this happening? What... Why is why is melatonin declining? Why is magnesium glycine going up? And I can get... I can I can? I could dig into this a bit more, and there are some clues actually in the data. But as we talked about the last session, we want to look at how we can bring search search data alongside social and news to try and get a more full picture of what's actually happening. Why are people searching something that probably happened in the real world that's triggered this massive spike in searches? So this is where I need to create a search query in BrownWatch, and this is actually a query, excuse me, that we created using our AI query writer tool. So I just logged in to BrownWatch, and I said, you know, create a query on, sleep supplements ingredients, and I actually just listed some of the key terms from search. So I I looked at the Trademark platform. I I saw the key things that were being searched for, and I just went to the branch and said create a query with all of these things in there, and it's created this query for me. So I don't have to write it myself. And it's it's really wants to make sure it gets conversations about all of those core ingredients. The query writer actually added some extra ones on top. And I also want to segment my data. This is another thing that you can do with the AI and Brandwatch is I can get the platform to help me create categories to break down that query because that's quite a large query. There's a lot of conversations going on about this topic, and I need to break it down. And so... And I can also, you know, use some of the clues that I've got from the search data to try and get an idea of the sorts of things I wanna look for. So I wanna look for things like core ingredients, health benefits. And then I put all of that into a dashboard, and I wanna start getting an idea of why this shift has happened. Why were people talking about... Why were people searching for melatonin November last year, and why are people now talking more about magnesium glycimates? So this is a a dashboard. It's actually a pretty standard dashboard, but, just using a template that I've got. But all of these text overlays are written by the Brownwatch, Iris AI feature. So it's it's looking at the data in each of the components, telling you, a bit about what's going on. And I can see, for example, that there... This is the AI's picked out that there was a significant peak in volume for sleep supplements occurring in November 2025. This search was driven by discussions around a study suggesting a link between melatonin and increased heart failure risk, with one major enhancing study suggesting link between melatonin use and increased heart failure risk. K. So that's what's happened. So this is my spike here, and I can click on that and I can get the, summarize that. I'm not gonna actually do that. But, this is where I I I start to join the dots. I can see in my search a massive spike, and I go to PowerWatch, and I can see, right, well, there was a lot of news, stories about this. And then, obviously, that triggered a big social conversation, and then that's resulted in people on the on the month that I happened trying to search this up and see more about this topic. And as a result, you can see that their general interest in melatonin has dried off died off since then. And instead, now people are looking for alternatives. So there was already some interest in magnesium glycine before this happens, but probably driven part by this, negative story about melatonin, the main supplement that people were using. And people are now starting to talk about taking magnesium glycinate instead. And so this is one of the good things as well that I can start to get real verbatims from people talking about their experience of using magnesium glycine. For many people, it's the first time they've started using it, and they're switching from melatonin to this alternative, core ingredient, and they're taking various different products with it. So I can start to look at the magnesium glycine conversation in more detail, and I can see, all these different topics that come up in that data. It's in the intro. I I... I'm a I'm a big data nerd, so as you can probably tell, I could spend a long time talking about all of this, but just wanna show you the sort of thought process that I have when I'm starting to dig into this sort of thing. And you can overlay. So... And I've got the search and news conversation overlaid against the search conversation. So this is a search volume each month for magnesium glycine. You can see there was that rise, and then it dipped off, and then it's been that steady, gradual increase. And so this November, that was the, I think, the catalyst, that that negative struggle in monotone that's been driven. So there's been this gradual steady growth. We've been talking about it and searching about it. So these two things are really going hand in hand. And And so I've got this dashboard. I've got a whole bunch of tabs in here. There's really too much data for me to read as a human. So this is where the AI can really help me is to start to highlight the spikes in the conversation. And if there's big words in the word cloud, it can highlight this for me. And it can also pull up verbatims, particularly, like, positive stories from people sharing their experience of taking my piece of glass. And I can open up my, AI sidebar, and I can use this now to talk with the dash... Or talk with the AI, and the AI can then look at everything in the dashboard. And I can have these sorts of, conversations now, or I can just jump to the top, and I can say, why are consumers choosing magnesium glass in it over melatonin? And it's saying, you know, melatonin has significantly more mentions, but this is the growth people are talking about it. And it's saying that melatonin is often talked about in relation to heart health, and I can say, you know, consumers are talking about the benefits of magnesium glycine. And then one of the things that we mentioned in the last conversation was AI, conversation. So what we've also done is we've gone and used Trajan to capture LLM responses, answering questions, because people might well search for something and also go to an AI chat tool like ChatPT or Claude and say, hey. Should I, stop taking melatonin? Should I take magnesium glass instead of what should I take? So we wanna start getting some of the answers in from the LLMs, and we can do that in the dashboard as well. So we can see here some of the actual AR responses. So down here, I've got a response from Gemini, response from Grok answering these sorts of questions. You know, what are the main benefits of taking magnesium glycinate? And I've got then the answer from Grok to that question, and I've got loads of other answers on those same topics. And I can analyze the AI's response alongside the armed discussions that we're having on social. And so you can see, for example, here, magnesium glycinate's role. This is now summarizing the AI's answers, and I can compare that. So I can see here, this is this is how the LLMs are talking about these ingredients, and I can then categorize those by different, health benefits, and I can compare that to the conversation on social media and on news websites. So I can see that the way that the AIs are actually talking about the pros and cons of these ingredients is actually a little bit different to how the conversation is panning out on social news. And it's really good to able to see this side by side and also then get the, AI capabilities in branch to highlight that as well. And I just wanna quickly... And I want... I promise to leave time for questions, and I think I've put a little bit too long in that first section. But, this is then a a project we've got set up to look at automotive. And we we showed, this data in the last conversation. This is search volumes around questions people have about buying a car. So best hybrid car, and best... This is in a number of languages. And you can see the overall volume here was gradually growing, and then it's really dropped off from, around, March, April 2025. Now that's not because people don't wanna buy cars as much or they're not wanting to do research buying a car. That's because they're starting a lot of these, discovery journeys directly in an NLM. People will still go to Google. You still see there's very high search volume through Google, and often they will get an error response in Google. But quite often, people are instead going to chat with you directly and having that conversation with with OpenAI's chatbot or in cold complicated expensive purchasing decision, and people wouldn't take time to research it thoroughly. And so we wanna then start to think about tracking those questions in an LLM as well as a whole lot of other questions. So we've got here an extensive tracker looking at the ways in which the AIs are talking about these different brands and, seeing, for example, if I filter this and I look for... I could... I can look at segments. So I can look at, for example, specifically conversations around, luxury and premium. So if somebody... So it's going and asking specifically they want a luxury or a premium car. Overall, which brands are being recommended? So overall, it's it's BMW, followed by Lexus, and I can then filter this to see what's being talked about, like, in terms of different, attributes or body types. So I can look at maybe just SUVs. So these are all questions that people will be putting in different languages to LLMs asking for, you know, the best luxury SUV. And then it could be, you know, electric or hybrid or it could be, you know, for a family. So we've got thousands of answers here. We've got 31,000 responses talking about... Oh, sorry. A 100... 13,000 responses talking about BMW. And so that's my overall tracker, but I can look at this over time. And I can also look at this in in the Brouch platform and compare that with other sources of data. So I want to look at the shifts in search interest compared with social conversation, news conversation, and then l and response option. I'm gonna jump to another project briefly. Bear with me one second. I'll share screen in a moment. Okay. So here's another project I've got. So that right. So this is now looking at that conversation really around premium SUVs. And what I wanna do is to compare the LLM narrative against search interest and then the way that the brands are being talked about in news and social. And, really, again, what I'm really interested in is differences in the way that the narrative might be panning out. And so I've got the element responses on the left. I'm interested in overall brand share of voice because, you know, I'm asking here for, you know, the best. So this is generally speaking positive sentiment. Okay? And I can see, for example, on the left, there are some similarities. So I can see, like, you know, BMW's talked a lot... Talked about a lot by the element. There's a lot of search interest in BMW SUVs, and then there's also a lot of conversations around BMW's SUVs on news and social. But where I see a big difference, though, is, for example, with Tesla. So Tesla is one of the most discussed brands on news and social, but its share of search is much lower. And also then the way that it's being talked about by the elements is much lower still. So you see, when I'm in the conversation around best SUVs, the elements really are not talking about Tesla a whole lot in comparison to its share of conversation on news and social. And, conversely, if I look at, let's say, Genesis, the the premium, Korean brand, the conversational news and social, and this is again looking at, I was actually basically focusing with English language content here. It's quite small. There's a bigger share of search interest from Genesys. But, actually, LLMs are talking about Genesys quite a bit. So that's one of the things that you might think, well, how is it that the LLMs are maybe talking about this particular model or this particular topic proportionally quite a bit more than maybe what you'd expect based upon the broader social and news conversation? It could be just the way that the framing of those, conversations are is really giving very strong signals to the elements that this is actually a really good model. Or it could also just be that the the places that that coverage is happening is happening in really the the websites that the LMC has really authoritative on the topic. And therefore, the LMC are reading one of that content at scale, and that's really what's shaping and framing their conversation. So I know I've covered quite a lot, here very quickly. I hope I haven't bamboozled you, listening to that. And I know some of you are, you know, users of these tools, and this might be new to others, but I just hope that it's helpful just to give an overview of, practically speaking, how we go about doing some of the things and getting some of the insights that we shared in our in our previous webinar. And so I think we've got some time for some questions, and I must make sure I go to the right place to get these questions. Okay. Right. I can see one here. Okay. There's a question here from Abby, which I can bring up. Are the social media questions filtered by sentiment platform and verified accounts so you can verify the topic you are searching related to the conversation? Social media questions. Well, they they can be. So one of the things that, I would recommend is using social listening to inform the prompts that you track in an NLM. So people might well have a question, and some people will maybe go to social media to ask that question. They might go post a question on PlayStat Reddit, and others maybe take that same sort of question and just go directly to the chatbots, and we don't... You know, they don't they don't bother to ask that question online. But nonetheless, especially if it's a sit... In a suitable... Suitably large volume of people asking the the... A question on social, we would definitely want to take that same question and put it to an NLM. So, we could we could look at both positive and negative questions, and look at that by platform. So the way that people might ask a question on Reddit, because often people will be posting anonymously, maybe quite different to how they would maybe post on Facebook or x, maybe where they're using a real profile, especially when it's in a health topic or something personal, like, personal finances. So depending on the the the... For some charity that you're in, the anonymous, places online are gonna be a better source to get the the really juicy questions, whereas other places, if it's more like a general interest question, people will be happy to post those using the real profiles. So, yeah, depending on which category you're in as a as a brand, certain networks will be more useful for you than others. And verified accounts, yeah, I mean, you you could filter to verified accounts. I mean, I think on x... I mean, if you're going for verified, normally, the people who bothered to buy the blue tick are kinda more more right meaning anyway. So you you just be mindful that if you limited yourself to verified accounts, you'll probably get, you know, it'll be maybe more of a a political filter perhaps that you get. I don't think it necessarily improves the quality of of filtering out bots because a lot of bots have the root tick anyway. So I don't necessarily think you need to to worry too much about the verified accounts. But, yeah, social is a great source to get the questions that you wanna track in another. Okay. I hope I answered your question. Reply back to me if if I haven't answered you properly, Abby. Another question here from Paul. On the car brand example beyond seeing elements of what does that tell you? What takeaways would you have that you can then apply? I'm wondering how you take the action based on an interpretation. Okay. Yeah. So one of the questions that I think people have is when you see the other answers is like, you know, so what? And I think one of the main, you know, things that you wanna try to determine is whether or not the ways in which the AIs are talking about you is really just a raw distillation of the coverage that you've got. So if, you know, the general kind of aggregate of all of the online coverage is, you know, the brand a is best followed by brand b, brand c, and then the LLMs just, you know, say the same thing, then you kinda see, right, okay. Well, like, that kinda makes sense. But especially for a complicated purchase or an expensive purchase, a lot of people are taking what the elements are giving them in those answers, and they're making their shortlist based upon the shortlist they give. You know? So we really wanna understand what's driving that. So we need to try to get a sense of whether the the AI narrative is actually different to maybe what we might be seeing elsewhere and then trying to figure out maybe what might be causing that. You know? Is it that the the coverage that is being generated on specific websites, is is we shifting the element of answers? You know, we actually need to see the the frame of the language. I mean, I I didn't wanted to do it on that session, but I would want to dive into some of those responses and see about why. What is the AI saying about this particular brand? Why why are they giving it, you know, a higher profile maybe than what we'd expect based upon that brand's general search share share share of search or share of voice on social? You know, it's it's getting a disproportionate, promotion by the elements. Is there certain qualities that just the AIs seem to be latching onto? And maybe then we should think about the framing of how we describe our own brands in that conversation and maybe maybe target different websites because maybe that particular brand is getting coverage on certain websites that are seen as more authoritative. So it's a combination of the language that's being used as well as the places that the coverage is happening that, would be the things I would focus on with that sort of analysis. Maybe, that would be something that you you do differently with maybe your peer activation or your social media campaigns because, you know, a lot of this is about getting that critical mass in order to, get the the AIs to kinda spot that... Those patterns. I think we've got time for... One question was, can you compare different markets or languages within the same project, and how do you account for differences in the way people search? Okay. Yes. So you you can track across different languages. What I would suggest is to use the categories and tags to try and have some, equivalency. So we might have lots of... Like in that, ingredients search. You know, people... You know, the the the way that people search for melatonin will be different across different countries. So I I need to have the keywords in the native language to actually get those search volumes back. But then the label that I stick on top, groups all of those together. So when I look at the label, I'm looking maybe at the English language label because that's my language. Or I could look at, you know, particular pain points or risk or concern or benefits. And the label that I have is English, but the actual keywords that comprise the label or or have the label are all done in the native language. So I need to make sure that any rules that I have to categorize things are done in the native language. And that's how then I group that together, and it also means I compare mark... Markets. When I can maybe see in one market, there's a very different set of needs and pain points compared to another. So, yeah, very important to have, native language rules and then looking at that together. And the second part of the question was, the ways... Different ways people search. So, you know, Google volume is the largest. You can also include, Bing and that, and you can also bring in then retail search into the dataset also. And then some people might be doing so... Social search alongside that. So you want to look for things like the volume of of hashtags on social media to get your clue. And then for LLMs, we don't get, at the moment, all of the raw prompts from the LLM providers themselves. So we have to kind of either use, the Google search volumes as a proxy or the social conversation as a proxy. And especially if you're in a a more niche software or in b two b, that's where you might need to do a bit more, what's what's called synthetic search where you actually just decide on prompts that you add in and, and then track based on that. And I think we're at the half hour point. So I tried to pack in as much as I could in the thirty minutes, just to give some examples, but, we'd love to continue the conversation. Do please connect with me on LinkedIn. You can ask questions via LinkedIn if you missed the chance on a live q and a, and we'll be happy to get back to you. So, yeah, thought it'd be good just to follow on the conversation that Maria and Danny had before and, lift the the hood a bit and show you the engine underneath some of the findings that we talked about in the last session. Thank you for your attention, and have a lovely rest of the day wherever you are.