Advanced Keyword Research

So yeah, this is me. I am head of insight at LinkDex. For those of you who aren't familiar with LinkDex, we are a big data SEO keyword ranking market analysis platform, not dissimilar to Semrush in many ways. Our clients tend to be kind of the much bigger organizations with complex ecosystems and infrastructure, et cetera, et cetera. Yeah, there's a lot of overlap between the kinds of things we do and the kinds of ways we think about things like keyword rankings and market analysis. So it's great to be here talking to an audience that overlaps that world. So my— a lot of what I spend doing in my role is problem solving for clients and trying to understand what their challenges are in terms of day-to-day SEO, operational challenges, processes, tools, where the difficulties are, how the political landscape works, what the future of SEO looks like, and trying to solve a lot of those problems. And in order to keep that focused and make sure that we're dealing with the things that matter— because obviously there's a huge world of options there— do we spend our time focusing on the overlap between PR and SEO, or do we focus on more granular rank tracking capabilities. There's a million options. And the thing it always comes back to is this core concept that people search and that search is the gateway from whatever the consumer wants and whatever their needs are to the providers and platforms and the companies who are positioned to deliver that. So everything needs to come back and start with search. And search, the mechanism of search, certainly at the moment, is keywords. So you type in what you want. There's a degree of inference from Google or the search engine as to what that might mean based on context or scenario. And even if the landscape and the mechanism of search fundamentally changes, if we move to a world where the search results as we know it ceases to exist, if things like Google Now continue to proactively anticipate and interpret what you need before you even search for it, we're still in a world where there is a form of intent there which is transmitted to the search engine, which is then interpreted. And at the moment, That's keywords. And I stumbled on a really important word there, is keywords mean nothing in their own right. What keywords are is a mask for this concept of intent. It's my attempt to explain what I want and what I need, which is part of the process of trying to find something that fills that intent. So if I type pizza, it's not because I'm looking for things about the word pizza. It's because I'm hungry. And that kind of concept applies across all keywords. So, Really, good SEO isn't about getting the right keywords in the right place and finding those opportunities. It's about managing good intent alignment. It's about ensuring that the people who want a thing are able to find that thing, and that experience lines up. And that's a much more nuanced premise than, do I rank well for the keyword that— even necessarily for the keyword that they were looking for, or the keyword that brings me the most money, or the keyword with the highest search volume. It's much more about, did this person with this unique set of needs in this unique circumstance get the thing they were looking for? Was that the right experience for them? Beyond the right result, was it a good experience? Was their intent matched? So that's kind of where I start a lot of my thinking from, and that's really important, because the combination of understanding the consumer intent and marrying that with the business intent, which obviously should overlap with a good organization— the thing you're trying to sell should align with the things that people want, but it's not always quite that cut and dry— and then also injecting an understanding of competitor intent, so how do those lines overlap and what's the gap, really allows you to start to understand the marketplace. Only with that understanding can you then start to pick apart things like the strengths, weaknesses, opportunities for both you and your competitors to start to identify and analyse changing and emerging trends. And only when you bring all of that together are you really in a position where you can start thinking about strategy rather than just having a list of keywords you want to rank for, because there's so many things And it's so much more nuanced and complex than picking a list of keywords that you want to compete on. Obviously, things like competitor context and market changes and shifts and seasonality and those kinds of considerations need to be woven into that. So this is really the stack that you need to have in place in order to make sure that what you're doing is a strategy that's not just a wish list. So yeah, this— taking you back, therefore, having really good keyword research and understanding that that leads to intent, that leads to market understanding, et cetera, is what underpins a good wider strategy. However, this process is broken. I have a background working in and with large SEO agencies, with development agencies, consulting from small businesses to pan-global enterprises, and I don't think I've ever come across any organisation where the keyword research strategy, the selection of topics that they want to focus on, the prioritisation of tactical day-to-day actions, where any of that is truly sophisticated and takes all of this into consideration, where there's a real— where the user intent champions the process and where the structure's been built around that, there's always challenges and limitations, and I'll come on to some of what those are and how we can come across it. So the biggest challenge I see in the world is that stakeholders confuse the concept of keyword research with providing a wish list, and that often takes the form of something like our top 10 golden keywords. These are the kind of terminology that you quite often come across, the idea of maybe having 100 commercial priority terms or keywords that are classified as primary or secondary based on commercial metrics, or keywords which are attack or defend, or keywords where you're only interested in them this quarter because they're the current hot topic or they're product is currently seasonally favorable, et cetera, et cetera. And the reason the world works this way is because this kind of thinking makes it very easy to report, so the tools don't have to be particularly sophisticated in order to manage this. It's easy to understand and digest without a vast amount of experience or education in the sector, and it's also easy to challenge them. So the senior stakeholders in the business can wag their finger and say, why have we dropped 3 places for this priority keyword? And obviously, that leads to problems where they don't understand the wider context and the movement around that and the fact that Google, in many ways, is a zero-energy system. So people can't go up in rankings without other people going down and vice versa. And even if you work really hard and do really well, things happen around you that influence that. All of these factors come into play, and you get stuck in this world where you end up being culpable for things that are outside of your control, which, from a business perspective, from a strategic perspective, and from a personal perspective, isn't a great position to be in. Not massively enviable. So this is a massive trap, this kind of approach where organisations pick what keywords they want to rank for, and the senior stakeholders have a say in, we want to chase this keyword and not that keyword. They don't have access to the wealth of information and the context and those strategic pillars that they need in order to be able to make those decisions, good decisions, primarily because you don't know what you don't know. So, Let's say one of those top 10 golden keywords is car insurance, which is an example I'll come on to momentarily. What happens if it looks on paper like you're doing really, really well for the phrase car insurance, but you drop off the radar for the phrase women's car insurance? You are only looking with a very myopic focus at the things that somebody has fairly arbitrarily said are important, and therefore the movement around that becomes invisible until then they spot it. It becomes the new top 10 keyword, the priority shift, and you end up stuck in this pattern where you are forever chasing the things that accidentally move around. You really need to be in a position where you understand the entire ecosystem and can look at it from a really high view and then make strategic decisions. And of course, as a result of this, there's money left on the table. If you are chasing unrealistic keywords or you are constantly moving the barrier on priorities and focuses, then really you're missing out on tactical opportunities as well as strategic focus where there may be pools of keywords or opportunities or types of intents that aren't being met, but because they're not as sexy or exciting as the top 10 golden keywords or this month's focus, then they don't get the appreciation and the effort that they need to take advantage of them. So all of this together breeds really bad behavior. It breeds really bad relationships in organizations where really all you ever end up doing is firefighting and chasing your own tail and trying to make everybody happy whilst everybody is well aware that they're They're doing the wrong things and making the wrong decisions and desperately trying to make that work when the premise itself and the whole approach is flawed. So there's some stuff that underpins why this is the case. And I think number one is pretty insurmountable. So you've got to deal with this— is generally the people involved in these processes who sign off on what's good, what's bad, what they want to focus on, have a lack of education or experience or an awareness that there are alternatives. So that bit we can solve through process. Historically, and certainly in some organizations, the lack of the tech or tools or processes to— sorry, the lack of the tech or processes, et cetera, to make this more sophisticated. So in some cases, you will only track 100 keywords because that's what you've got budget for, that's what your tools are capable of, and so on and so forth, which is somewhere where both Semrush and LinkDex share the same kind of idea about making it very, very easy to see much larger keyword pools and to understand far beyond that. So that's certainly not a new thing, but a lot of tools and a lot of organisations are geared towards selecting the keywords they want to look at and monitoring them, rather than getting an understanding of everything that's out there and then focusing in on the important bits. This is the big one for me, and I think it's a problem that's pervasive across a lot of areas, that I've certainly come to believe that the primary responsibility for senior stakeholders in any organisation Is to be in control, and to— or at least give the appearance of being in control, because essentially that's what leadership is. And that manifests through stating objections, asking questions, pushing back, being resistant to new processes and ideas, not because they're necessarily negative people, but because— oh, my train of thought has just gone all over the place. Not because they're negative people, but because the world, certainly the high-end commercial world, rewards— Not making mistakes much more than it rewards risk-taking that leads to success. So these people have found their way into these roles and pushed these organisations the way they have in order to be quite defensive and not to take any risks. So as soon as you try and present different options and opportunities, their first reaction is naturally to be adverse to it. And that means that things like saying, we want to target 5,000 low-value keywords that will add up over time, and we want to look at the long tail, and we want to do a content strategy— all of these these things as a natural resistance to because they're different and they're unproven and it's difficult. The result of this tends to lead to the keyword research as a process very rarely resembles actual research. It tends to be a very truncated token effort at spotting some numbers that match patterns that we've all been conditioned to look for, which is find me the keywords that have a lot of people searching for them, which look like they're reasonably commercial but not too commercial, that look like there's some competition to suggest that there's interest, but they're not unachievable. And you will filter through a couple of thousand keywords and pick out the top 100 that match that criteria, and that becomes your golden list, your seasonal focus. But that doesn't feel like research. That feels just like a process and vetting down from Google AdWords data. Research involves the human element. It involves thinking about that core concept of intent matching, that these are human beings who are expressing a desire for something. How well is that met? Where are the gaps? Where are the opportunities? What's the expected experience? Research is a thing that in any other industry, in any other context, takes weeks, weeks and weeks. If you try and pitch the idea of keyword research taking 2 to 4 weeks to complete, you hit that CEO and senior stakeholder resistance because their experience is, I want it faster, I want it quicker, can't this be truncated, do I really have to pay for it, is this just an excuse to inflate the costs? Certainly in agency world and in-house with smaller teams, you don't have the luxury of spending 4 weeks undergoing a process to thoroughly understand your market and spot all the tactical opportunities. You've got to spot the hit list, because there are sales targets this quarter, and the numbers need to be met. So you need some, quote unquote, quick wins, even though typically the 20 other companies in your sector environment are having exactly the same problems in process. So we're all chasing each other across those top 20 golden terms. And you really end up in a position where there's a huge amount of money left on the table just a little bit below everybody's focus. But they're all very, very busy trying to compete against each other for the stuff that nobody can win. So I think a lot of this can be fixed without huge amounts of effort. I think certainly more effort than we have at the moment, and it will take longer than we take at the moment. But I think there's an educational curve and processes and tools that we can use to make this better. And it really boils down to these 5 steps, which is it's not enough to skim through and decide what keywords you want. You really need to get all all of the keywords, and by that I mean every possible phrase that anybody might ever use in any context to find the thing that you do or want to do, or in the case of moving into new markets, the words that they might use that don't necessarily exist yet. You need to have all of those in place so that you can look at them and say, yes, here is a tactical opportunity, here is an area that we should build strategy around. Of course, then the problem is that you have hundreds of thousands of keywords, and they're utterly impossible to consume and utilize, so you need to start to introduce a much more granular and powerful level of categorisation and classification than we typically do. So multiple taxonomies, multiple types of categories, multiple slices and segments are really important, and I'll come on to that. And then slicing and dicing by those to find opportunities, which again I'll talk in more detail. You are trying to change behaviours at this point, so the process of getting buy-in becomes challenging, so there's some thinking around how best to do that. And then monitor the results. I should point out at this point that This is a process that at LinkDex we are working with, with some of the world's largest companies, and it's been an interesting experience trying to change hearts and minds around how keyword focus and keyword research works. But the power that becomes evident when you can see the entire market, rather than saying, we've picked these 10 keywords, being able to say, here is a niche that none of your competitors are even thinking about, and look at all the opportunity, is really, really powerful. I'll hopefully persuade you that this is a good way of thinking as we progress through. So the process of getting all the keywords. I want to give you an example to start with about just how big a deal this is. So typical keyword research would suggest go to AdWords, go to Keyword Tool, go to Wikipedia, and so on and so forth, and start to collate all these keywords. And you'll get something like young driver's insurance, which Google AdWords says has 6,500 monthly searches. So that goes onto your list, and it's not as exciting or impressive as other keywords, so it tends to drop up off the bottom. But what becomes apparent is when you start to think about things like variants, plurals, and stems, and synonyms, and typos, and start to think, what about car insurance for young drivers? Young driver versus drivers, singulars, plurals, so on and so forth. You can start to go out and say, what are all of the possible phrases that have the same intent as this term? It might be insurance for people Bad example. There will be phrases which don't look and sound exactly like the same word that have the same meaning, that imply the same needs, that suggest the same audience, that don't necessarily just exist as typos and flaws. So there's a huge process to go through to find all of these, and that involves extensive research with Google AdWords, with things like Keyword Tool, with Semrush, looking at the competitors to understand which competitors have overlapping keywords. Semrush is great for this, and LinkDex too. Is go and find out what everybody else is ranking for in these areas, and then go out and expand those keywords through Google AdWords, through other tools, and start to build out this universe until it's really, really, really huge, and you've got a reasonable chance that you've caught all of it. So you'll end up with something like this, a table which gives you all of the keywords that you're looking at, and then a whole load of extra data like paid search information and who's ranking where. Not dissimilar to the kind of thing you would export from Semrush or LinkDex that you might then add some business data like conversion rates and so on and so forth onto, to give you a nice big table that you can pivot in 100 different ways and get a feel for what the opportunity is. So this is step 1. The thing is, there are 266— just in 20 minutes or so of research, I found over 260 variants of the phrase young drivers insurance. So if you're looking at it in terms of this table and it says young drivers insurance and it falls off the list, actually it's hugely unrepresentative. Because there are 65,000 monthly searches for the combination of those phrases. That's 10 times as much as there is for the phrase young drivers insurance. And when you start seeing those kinds of terms, this isn't there are 6,000 people searching for this word. It's there are 65,000 people who have this problem, who have this need, who have this intent. And the way you model that needs to become different to represent it. So of that, there's £320,000 of monthly CPC equivalent value there. That's a huge amount of money if you can take even just a small slice of it. What becomes really fascinating is from a brief bit of research, this is from about a month ago, the site ranking in position 2 for Google for the phrase young drivers car insurance and a whole bunch of combinations of those terms owns 13% of the market, gets about 8,000 visits per month, which is about £40,000 of media value per month. And it's got a page with 120 words with no links, no shares, and on a tiny website. This kind of scenario is really pervasive. You'll find scenarios where, because everybody who's competing in this space has been dismissive of any keyword that doesn't fall in their top 10, all of these opportunities are missed. And you'll frequently find that when you start to combine these intent groups, that the guys ranking in position 2, 3, 4, or 5 have absolutely no equity and are really easy to beat. So a screenshot from LinkDex just to demonstrate this. As you can see, the guy in position 2 has very, very, very little content, a tiny site, and over these 266 keywords, they're in second place in the market, which is an absolutely ludicrous position to be in. And the other lines highlighted in blue are brands that you've never heard of. They're small affiliate sites. They're little blogs. And these guys are generating a huge amount of revenue and certainly a huge amount of market share just by dint of being there. There's nobody competing in this space because the big organizations and the big brands are stuck in this world where they're looking at their golden keyword terms. So there's a huge amount of opportunity just below the surface. So it'd be easy to dismiss this as just a one-off, so definitely not the case. Same industry, same kind of problem, looking at the phrase one-day car insurance, which is hugely competitive. And if you look at the kinds of brands who are evidently competing on this, it's people like Confuse, Money Supermarket, Aviva, These are big brands with big budgets, with big search strategies, and they have decent pages with decent content. They've got some decent links. But again, that's not a single keyword. You look behind that, and there is one-day car insurance. There is day car insurance with another two and a half thousand. There is car insurance for a day with another two thousand. So on daily car insurance, one and a half thousand, and so on, etc., etc., etc., etc. There are 71,000 people searching for 1-day car insurance using a vast array and various language, but those are the same kinds of people with the same kind of need. Yes, within that there are probably subgroups with slightly different needs, but overall this is a whole topic area where people are competing very, very heavily. But the biggest brands ranking within that space, the Avivas, the Money Supermarkets, have less than 10% of the market share. And again, the guys in position 3, 5, 7 are absolutely nobodies, because nobody is looking at this middle ground space and grouping these phrases together. It's scary. So the core concept here is that keyword research really requires you to stop thinking about individual keywords and to start thinking about groups of topics and phrases and need groups. So who are these users? The language they use is irrelevant. The intent behind that language is what's really important. And as soon as you start clustering that and grouping it together, you can get a much better understanding of who are my consumers, what are they looking for, and how do I best react to that. What's really interesting, I think, is that certainly from my experience, some of the world's biggest brands and some of the world's most active, most sophisticated SEO strategies, some of the really big players, at most are tracking a couple of thousand keywords. Now, in conventional thinking, a couple of thousand keywords sounds like a huge volume, but when you think that there are 266 keywords just for young drivers car insurance. If you're competing across the entire insurance industry or the fashion industry or the tech industry, regardless of what it is, if individual keywords we're looking at, at a scale of 250 times too small, there's a huge amount here. And if the biggest brands on the planet are only tracking a few thousand, what are they missing? And how many need groups are they underrepresenting if they've only got a fraction of the keywords for pools of intent and people who have certain requirements, if they're only seeing a small amount of that, all of their estimations on how valuable those niches are, how well they're performing, what the opportunities are, are misrepresented. So we need to move to a world where we look at tracking many, many, many more types of keywords, many different scales of keywords, and start to think about those by need groups. So that's step 1, is that kind of slight shift in thinking away from individual keywords. Because really you want to get to a point where the process of keyword research changes from being about discovering a magical keyword that has loads of volume that nobody's thought about competing on to actually already having collected and classified and categorized all of the keywords that you might be interested in, and then pivoting them, exploring them, finding out where the interesting bits are for the here and now, and then taking tactical action off the back of that. So what do I mean? I've used the phrase categorization a few times, and it's something that we do a little bit of at the moment in the traditional keyword research process. So we might take a stab at, these are keywords to do with trousers, these are keywords to do with socks, these are keywords to do with men or women. But actually, we need to go much further than that. So the first step is to group by topic, and then by behavioral clues, and then by real-world context. So what do I mean? The biggest thing for me is horizontal classification, in that typically the keyword research set, the golden keyword list, will often reflect either a site's vertical hierarchy, so your category system and your product service group, or by extension, the business's hierarchy, so the way you operate and the products that you sell. It's very, very straightforward to say these are trousers keywords, these are dresses keywords, these are hardware keywords. What's really interesting is starting to say that these are red keywords, or these are long keywords, or these are cheap keywords. Even if those particular phrases and pieces of language aren't present in the individual keywords, you know by inference and by categorisation that those keywords are relevant to those categories. And you can start to say, again, categorise by intent. These are keywords which are likely to be used by somebody who has high buying intent. Or somebody who is asking a question, or somebody who's unsure about our product benefits, or is new to our industry. This kind of classification is much more powerful, because then when you get to the point of saying, how do we pivot this to explore and find an opportunity where our competitors aren't doing particularly well, or where there's a need group where the intent isn't being matched, being able to pivot by the types of questions and language people are using, and the horizontal classifications, and then say, okay, now show me where there's volume, versus opportunity or where we've already got a page is much more actionable than, yes, okay, we've found some competitive keywords where we're already trying and kind of struggling against competitors. You can find gaps very, very quickly with this. So what can you infer about visitor intent from keywords? There's so much you can do with this. There's a wonderful presentation by Carmen Mandeiros. There's a link in the deck to it. I thought I'd done a bit there, but clearly not. About analyzing your keywords and inferring bits of intent. So she's got— it's well worth looking at the whole deck. There's some wonderful sections like this talking about something. So quite often you get a very basic slice of brand keywords versus non-brand keywords. Brand even. How do you take it further? How do you do— okay, within brands there are people who are definitely bought-in evangelists who are definitely going to buy, but there's also people who are heard about us and associated us with a product but maybe don't look like they're gonna purchase. You've also then got support queries, which you probably don't want to be including in your tactical opportunities. You've got people whose use of language would suggest that they're repeat buyers and that they're coming back again rather than discovering you for the first time. Slicing and dicing and classifying in this way again allows you to get quite granular and clever with where you look for opportunities. So you can start to take this further. You can add in real-world context. So again, using something like Semrush or LinkDex, you can take all of that existing framework and layer on things like, OK, where are my competitors ranking? Because actually what looks like an opportunity might be misleading if it's a very— the space is saturated with people who I've nominated as my competitors. Or even more interesting, where are they outranking me? Where is my content in 4th and the 3 people in front of me are competitors. Because what might look like a tactical opportunity when you slice and dice and segment that way may in fact be unrealistic. And again, rather than having that on a golden keyword list where you expend disproportionate amounts of resource to try and move it when it's not realistically going to stay there or give the returns, you can start to say, okay, well, we'll try and maintain that position, but we're not going to be proactively chasing it, and we ought to look over here instead where the guys in front of us are little blogs or those kinds of affiliate sites and so on. You can start to get really sophisticated with this and start to think about things like, where am I being outranked by, for example, YouTube videos or eBay? And we had a pitch for a company recently, which was quite interesting, where part of their SEO strategy is to be outranked by retailers like eBay, like Amazon, because it's a really positive relationship for them. And they don't mind if the guys ranking in positions 1, 2, and 3 are YouTube videos of people engaging with their products. If they're ranking in position 4, that's great for them. That's media exposure. That's a much better testimonial system for them than putting it on their own sites. This kind of thinking gets really sophisticated. Again, when you take it back to the keyword list, you're going to start classifying as keywords where it's fine for neutral parties or vendors or retailers to outrank us as long as there's a relationship there. And you can start to manage that ecosystem. And this is really the key thing, is it's not about having a big list of keywords that you want to chase and rank for. It goes back to control and making sure that you are classifying keywords by individual strategies, by individual tactics. Also, looking inwards rather than outwards, where do you have executional advantage? If you can't produce content easily because you are, say, in the finance industry, there is no point trying to target keywords where people have questions. So any keyword which is who, what, where, why, when, is going to require you to produce FAQ-like content or blog posts or informational stuff that matches those intents and fulfils those queries. If you have content production restrictions, there's no point those kinds of keywords being on a strategy list, even if when you aggregate all of them together there's a huge amount of volume. You need an element of pragmatism and realism, which then gives you a strategy rather than just a tactical wishlist. That element of grounding it to your reality is really, really important. The problem is this is a lot of work. This is really, really resource intensive. I mean, those are a few fairly arbitrary examples, but if you're talking about thousands of keywords, you can imagine how long it would take to crawl through every single one of them and say, this is a question about red trousers that implies a new brand experience that is very cheap, and so on and so forth. You don't want to have to go through and do that all by hand. It's a huge amount of work. Hugely valuable, but again, going back to the real world, you don't have 4 weeks to do this typically, so you need a way of doing it faster. It's also ludicrously huge. So we recently categorised and mapped the entire mobile phones industry, and we reckon just looking at handsets and a couple of the big brands, there are upwards of 150,000 keywords, which you don't want to be crawling through one at a time and saying, this is this brand, this is this handset, This is a support query. This is about an app, so on and so forth. But you need that scale in order to spot the tactical opportunities. What you can start to do, which is really cool, as an example from something we're working on at the moment, is this is a visualisation of the keyword universe of the finance industry by the types of content that's informational. So this is fairly specific to a project we're working on at the moment, but you can start to model entire landscapes visually and say, who are the types of sites who are playing here? What's the relationships between the types of language being used? And how can we then take advantage of that, and where are there opportunities? So there's reams of data behind this, but you can start to play with using this kind of thing to get stakeholder buy-in, to explain the complex concepts around keyword universes and different slices and dices. You can use that term to get the message across. The other problem is these universes aren't static. They change frequently and continually, some more so than other industries, but going back to the mobile phone sector example, example, it's very, very device-centric, and there are constantly new devices being released and devices being retired, and the technology itself changes. So you've got new types of keywords and new phrases. The word phablet, as much as I despise it, is relatively new to the industry. If you only do this process once and only tag and classify all your keywords once, big bits of it start becoming irrelevant and incorrect in about 6 months, but you start missing opportunities, which puts you back to the original problem, is you've essentially got a big golden keyword list and you're not looking at the stuff that you don't know. So you've got to keep this up to date as well, which makes it an even bigger process, which puts you in a position where you spend all your time analyzing this keyword pool and you get that top-level 30,000 view, but you lose the agility and the ability to act tactically. So one of the things that we've been working on is making this process much easier. We've recognized the importance of it and the necessity of it if you really want to spot tactical opportunities. So we've built a tool which I'd like to share with everybody. So this is a free tool which is currently a prototype, and it's very, very much in beta, but it's powered by a whole lot of linkdex thinking. And the idea is when you get to that stage when you've got 5,000, 10,000, 20,000, 50,000 keywords, you can paste them all into a text box. You can click which industries they align to. So in the example on the screen, this is a huge amount of keywords relating to the car insurance industry, so I've selected cars and insurance. You can then run that. It takes a couple of seconds up to a couple of minutes depending on how much data you've put in, and it spits back something that looks like this. So it gives you each of your keywords back, and it's run against a whole load of logic rules and a whole load of language analysis and said, okay, we think the phrase For example, moneysavingexpert.com car insurance is about a brand. It's about a travel brand. It's a URL— excuse me. It's in the insurance niche. And it looks like that the combination of those factors would indicate that this is a navigational search behavior, which you would expect to, for example, have a different click-through rate, which you might want to factor into your keyword modeling. You would expect it to— You would expect the intent to be different, that you might expect a high conversion rate, so on and so forth. You've got really interesting things like cheapest flights online booking is about booking, it's online, it's also price-centric, so that's quite interesting. But the combination of those factors suggests that the user may be in the consideration set of the buying cycle and also that they've got high purchase intent. And when you extract this kind of information back into your tables or CSVs, you pivot against it, You can start to say very quickly, find me the people who are using questions who look like they're in the consideration set. Well, those are people that you might provide some informative content for, which means when they are further in their buying cycle, you're already in that consideration set, and it might mean that you, in fact, change SEO from being just a demand model where you are trying to align people's intent to what they're searching for to actually creating demand. And if you're putting yourself in higher up the buying cycle than you previously were, people's search behaviour might change, their brand association might mean that you end up being present later in the cycle as well, and you start to really take control of that ecosystem. This is very early stages for the tool. It performs better in some industries than others. It's a little bit quirky in places, but it's already phenomenally powerful. This has become part of our day-to-day processes at Lindex, is when you're looking at large volumes of keywords, You don't want to be scanning them all individually. It's not perfect. You still want to spend time skimming through the outputs and getting a feel for what's in there. But this is a phenomenally powerful assistant to really doing a lot of that classification. So go and have a play. I would love your feedback. I'd love to understand practical use cases and how it plays a role in your workflows. And feel free to go and jump in there. So yeah, even having done that, you need to go and eyeball it all still. So you need to spend time skimming through and making sure that you've not confused things. Like the word Boots is a huge pharmaceutical brand, it's also a type of clothing, and mostly the tool does a reasonable job of classifying that correctly, but sometimes it'll get it wrong, sometimes it'll get other things, false positives. And of course you've got a human element in here as well, yourself and the people reviewing stuff. You want to add extra context and change things and reinterpret. So that takes you back to this kind of place where you're at with this table of keyword data. You can now expand this with 30, 40, 50 rows of segments, of categories, of taxonomies, of colours, of shapes, of sizes, of locations, and it becomes a phenomenally powerful tool. And you can say, show me all the little opportunities. So things like, show me keyword phrases relating to our core strengths where visitors are early in their buying cycle where we don't have highly performing content. And that's quite a complex request, quite a complex statement, but actually when you've got a spreadsheet or database or tool configured in this way, then you can very easily say, okay, pivot by buying cycle classification, order by levels of content performance, and then you can spot gaps where you can say, okay, well, here we need some educational information, here we need more information about our proposition, maybe not hard-sell content because these people don't look like they're ready to buy, but if we can make them brand evangelists and interested at that stage, then great. I see examples of companies who do this really well. Bellroy in particular do phenomenally well at the moment. They produce very, very skinny wallets, so their product USP is their wallets and purses are designed to minimise the airflow and unnecessary folds so that you can have 30 credit cards and a wad of cash in your wallet and it not eat your pocket. So what these guys do very well is they understand that not everybody is going to buy immediately because it's a premium product, but there's a moment of discovery where they've created content that ranks for informational queries which are high in the buying cycle. So I want a nice new wallet. They know full well that a lot of people who Google that, who come on the site, who watch all their videos, read all their content, have their intents matched and aligned, won't necessarily buy right then. But they will come back in 6 months and search Bellroy wallet on a brand search that looks like it's got high purchase intent and then convert. So they've nailed this. There's a few other people who do as well, but this kind of process is really powerful. You might want to look at things like, show me topics where we are ranking well, where competitors are weak, and consumers are asking questions. Well, that's again really easy to pivot. Analyzing and assessing questions is something that the tool does really well at the moment, so you can spit that out. And then all you've got to do is find the existing content that you've got and extend it and expand it. This isn't a huge strategy. It's not high investment. It's not creating new— vast amounts of new content. There should be relatively few technical barriers if these are pages that already exist. This is really, really easy, low-hanging fruit, which allows you to get proof of concept, which allows you to show quick and demonstrable success because you're ranking all of these keywords. You can see the value that you're delivering. You can generate the extra conversions. This is really powerful. Another example, types of language which indicates high purchase intent where we're outranked by competitors who don't rank for longer tail queries. Again, this is about tactical disruption of the buying cycle. You can say, okay, well, if there are these kinds of scenarios, then where can we interject content that fulfills a need and matches an intent when people are still at the research phase or much higher up in the process? You've got to track and react to all of this over time, which means you need a tool or platform. Again, Semrush and Linkdex and others are very, very good at this, but this allows you to maintain and constantly refine an understanding of the entire market. You've got to top up the keywords as new things emerge. You want to be looking at using things like Google Trends and Wikipedia and keyword tools, Google AdWords obviously, in order to collect and harvest and input those new keywords. The nice thing about this is it's iterative. So if you add in another 2,000 keywords to your pool, you don't have to start again. You've already classified and tracked a whole load of them. You can inherit and extend classifications and tagging and so forth. So this is all nice, but it's a lot of work, which means that somebody senior in a business is going to say, no, just give me results now, rank for these 10 keywords. So you need a process of getting some buy-in as well. The thing is, when you start to segment down to this level, it is so, so easy to find little opportunities like that example of pages where I'm already ranking where people are asking questions. It's a day's work to go and craft a load of questions and answers, get them onto a website, and watch the rankings for that group jump from position 8 to position 6. And whilst that's not hugely exciting, when you consider that that key might be a group of keywords of 10, 100, 500 keywords, when they all move a little bit, it has a huge impact on levels of traffic, on values of traffic, and so on and so forth. So you've got to really start small but aim big. Identify the stuff that is fairly easy to move the needle, that won't change the world overnight, but if you push it and push it and push it further, then there's big payoff at the end. You do have to aim high enough in order to generate some kind of tangible result. It's not enough to make a token gesture at something. I see this so often with companies who will finally give buy-in to create a new content page or to target a type of keyword that isn't obviously commercial or transactional, And they will spend so little time and so little budget and give so little love to it that they doom it to fail immediately. So you get examples like this. This is one of my favourite things I've ever seen. This page has 150 words. It's trying to attract informational queries to a page about business finance and invoice factoring, which incidentally is an industry where the average cost per click is often up upwards of £20 or £30. This page has no links and it has no likes. You can get 150 words of content written on the open market for about $2. This page isn't going to achieve anything. So even if you've gone so far as to identify that there's an opportunity around stats for haulage within a certain niche where there are people asking specific questions and so on and so forth, you've got to do a good enough job to actually meet that intent. And this, this is the problem with the whole world of thinking about how with keyword rankings and keyword research is the process that's evidently been gone through is they've spotted that keyword, and they've done a good enough job to identify that there's a niche here, but they've not thought about the audience intent, and they've not aligned to that intent, and they've not met the query. They have ticked a box of, here is a keyword that we want to rank for that we think might be right for us, so it could work, but then just created a page, just written some content, but not thought, have we effectively met the intent of the user, which is to answer the questions that they have, to convince them that this is a useful resource, to hopefully get them into the buying cycle. I don't think it achieves any of that. Similarly, we find quite often people will sign off on a piece of creative. Exactly the same story on this is actually the concept of the history of money and the evolution of the concept of currency is quite interesting. Personally, I think you could do a huge amount with this kind of content in response to high buying cycle and research behavior in the finance sector, in personal finance and business banking, any of those areas, this is the kind of content that could answer a lot of questions. But again, executionally weak, underfunded, poorly conceived, and doesn't do a good job of matching that intent. So this kind of execution isn't going to get you the buy-in, isn't going to show you the kind of success that you're going to need to do more stuff. So really, pick an easy win. Pick a really ludicrously easy win that isn't commercially sensitive. So deliberately present some tactics to discover some keyword groups that aren't high commercial, that aren't highly commercial, that realistically very few people are ever going to convert from, but fulfils a need. Find people who are asking questions, who need guidance, who need help, who need support, and go look at some tactics there. So again, things like find pages which are already ranking for informational queries and work to improve those. Find places where consumers are getting a poor user experience, where the homepage is ranking for a specific query rather than a dedicated landing page, or where the only content returned on a page is a dead forum thread. Find these kinds of areas where you can make tactical tweaks. Define the success of these activities in terms of improved intent alignment, not in terms of revenue, not in terms of conversions and sales. Set out and say, I'm going to find places where there are people looking for something and finding us and having a bad experience. Fix those bits and measure success in terms of things like surveying and saying, what were you trying to achieve? Did you try and achieve it? How would you rate your experience? And then you'll have tangible metrics which say, previously we found this niche where we're not doing very well, it's not very commercial, but people were having a bad experience with our brand. Our brand equity, our reputation was suffering because of this content, because of this experience. We have demonstrable data which suggests that we have made these changes and this is no longer the case. We've improved it by X%. You're not going to make any sales because it's a bit woolly. It's going to be a little bit tentative, but if you can get buy-in on that and you can execute and you can get a small win, you can then take it and say, okay, we're going to do something commercial. We're going to find another small opportunity, but this time we're going to look for something that should drive some values, that should make some sales. We're going to find somewhere where there's a niche that's underrepresented, where people's needs aren't really being met, where they're searching for a colour that we don't stock, or they're searching for a product that we call something else, or a service which is a little bit ambiguous where we maybe ought to have a specific landing page. Find those little tactical options and then push for those and do one little project at a time that all it's designed to do is identify consumers who need something where you're not giving them an optimal experience, and there are always There's so many of those. Go back to young drivers car insurance. The site in number 2 has a page and terrible content and no likes and shares. It doesn't take a huge amount of work for anybody with any budget or brand or website platform to outperform that. The only reason the search landscape looks like that is because nobody's actively competing. Charlie did a really good webinar at SEMrush last week about how to produce that kind of content, how to match it up, go have a look at that. Really, really great. Implement a lot of that thinking. It lines up perfectly. Where have I gone? Got lost. Right, so just kind of wrapping up. So the process of turning that strategic. I have a link to another presentation that I put together a while ago, which is This is the kind of output of it, and the idea— I won't dwell too long on this, but the core concept is, in order to manage senior stakeholders' expectations and to get buy-in and to get agreement on what success looks like, which gives you further buy-in, which allows you to keep working doing this stuff, you need to have established a framework that articulates what the high-level business objectives are, what role the website and your activity plays in that, and which explicit KPIs you use to measure and define success. In each of those cases, you can take a process framework like this and integrate your keyword research and tactical activities off the back of it into. So I would definitely recommend going and having a look at this. In practical terms, you might expect to end up with a framework that looks something like this, which includes a KPI like What proportion of visitors are getting a representative experience where we meet their needs? What percentage— how are we mitigating having a high percentage of visits where they don't find what they want, what they're looking for? And again, tying that back to tactical activity. Definitely go have a look at the process for building this kind of framework and getting senior stakeholder buy-in. It's really powerful, and it enables you to do some really Secondly, one of the bits that does require some hard work is you— even if you say this small proof-of-concept piece of work isn't designed to deliver commercial return, you will often still be expected to model the big picture and the long game to get buy-in to even begin to do the small pieces of work. So you need to get to a place where you have a forecasting and trend model in place that allows you to say, if we did 100 of these, it would achieve us this much over this period of time. There is no easy way or shortcut to doing that. There is a process with the template that I'll share which follows something like this, which is essentially you need to look over several years of historical performance data, you need to process all of that and forecast what it looks like if it goes forward, compensate for seasonality, look at getting rid of any outliers from tactical activity and so on, and then extend that and plan and try and anticipate what your impact will achieve. So you will end up with something that looks like— this is a small snapshot of a large framework which says month 1, we anticipate getting this many visits. If we do a campaign, if we do this campaign, we expect it to increase traffic by 2%, to decrease conversion rate by 5%, to add whatever equity, etc., etc., etc. You can model that over time and plan out. If you're going to do one of these little opportunities every month, you know what the traffic volumes are. You can expect that increasing performance will give you whatever percentage of the market. You can model it back, and you can get to something like this that says, okay, if we only do one, it's not going to achieve really a lot. And in fact, if we only do 50 in a year, it's not going to achieve a lot. What happens is when we start doing lots and lots and lots of these over time, we start building up the effect. So you will need to get some kind of forecasting infrastructure in place. I will share a link to a cleaned-up version of this framework if you like. So wrapping up then, the key really for me is we've got to start looking at more keywords. It's not enough to have a top 10 list or a top 100 list or even a pool of 1,000. You really need to be in a position where you are looking at everything that's important. And that process, it's obviously not realistic for everybody to track a million keywords, certainly not even within Semrush and LinkDex, it's massively unrealistic. But you can make sure that you've looked at and evaluated up to, say, 1 million keywords if you've got time to burn. You certainly look at huge volumes and then discount and ignore large chunks of them and focus on the stuff that matters. You don't have to actively rank track or analyse and assess stuff that isn't going to be relevant or important to you. You can exclude things which aren't commercially viable, which you're not going to take action on. If there are keywords where if they dropped the radar overnight, you wouldn't do anything about, then there's probably not a lot of value in tracking them or assessing and analysing them. Categorise and classify the heck out of everything in order to be able to start thinking about topics and need groups and horizontal classifications rather than just vertical scales that reflect your site hierarchies. Because, again, off the back of that, you can start to identify tactical opportunities and demonstrate effectiveness. Which ultimately allows you to move from having wishlists into having a real strategy that empowers you to focus tactically on the things that matter and make a difference.