Don’t neglect keywords

Charlie says: “Keywords do still matter.”

Why do some people not believe that to be the case?

“I don’t know if it’s off the back of AI developments, but a lot of people have been talking about keywords being dead and saying that you should stop focusing on them. They quite rightly mention entities, which should absolutely be at the forefront of any high-volume SEO strategy.

However, for lower-volume sectors where Google’s knowledge of those entities might be limited or when SEOs haven’t done much work building them, keywords are still hugely important. They should complement entity SEO, but they are still vital in terms of optimization and tracking the impact of your work.

Google has been building its knowledge graph of information, and acquiring non-knowledge graph information, for the last 10-15 years. It’s been trying to rank ‘things, not strings’ and concepts. Historically, a way to do that would have been by using keywords.

That’s less the case now as Google is building this entity database. For high-volume terms where these entities are pretty established and Google has a good concept of them, then you should be looking to do optimization and make sure that they are using those correctly. However, the keyword aspect of that is still important.

People have misused keywords, which has led Google to change the algorithm, but for local SEO and areas where Google doesn’t understand what those entities are, the quickest way to succeed is still by using keywords in the right way.

I’m not talking about sticking every single keyword from Google Keyword Planner onto every page. I’m talking about looking at clustering and search intent and using keywords to inform your content strategy, and that still needs to be keyword-led for a huge number of verticals.”

Is it not reasonable to write naturally and then Google will understand your content and display it for associated terms as long as you build up that entity authority?

“A lot of the time they do, and there are a lot of cases where they don’t quite do what they’re supposed to do. That is often due to the ambiguity of entities.

I’m working with an ambiguity at the moment, which is ‘workwear’ compared to ‘work wear’. One of them is industrial workwear and the other is things that people would wear to work. If you type that term into Google, you’ll see a very mixed SERP of people wearing high-vis jackets, and then brands like Next. Google still can’t work out the intent behind those searches, and that’s due to the entities that people have been building.

I wouldn’t always trust that Google will understand your content. If you structure it properly, using things like structured data and keyword data and looking at search intent, then it should do. For high-volume terms, if you can link your entities properly, then Google should understand that. However, the best way to rank is still to create engaging content on that theme, and a lot of that will be led by keyword data.”

By publishing great content around keywords, are you also educating Google about what your entity represents?

“That’s exactly it. By creating excellent content and creating something new about an entity, you’re training Google about that theme, which Google will reward you for. That’s especially true when you use structured data properly and things like sameAs schema to link it with Wikipedia, Wikimedia, or other sources. If you help Google understand that entity, they tend to reward you.

If you can create the best, most compelling and engaging content about a topic, and you’ve got decent authority, you should rank. If you’re helping Google understand that topic in more detail, then it’s a reciprocal process.”

What process do you go through to discover the optimum keyword to use in your content?

“I do use a few AI tools to help me cluster and research based on what Google’s actually returning rather than just a list from a keyword research tool. I don’t obsess too much about the volumes. They are obviously important, but you want to look at the intent behind those searches and what people are searching for.

I use a fairly traditional keyword research process of collating the data from as many different sources as possible. I don’t just use Google Keyword Planner. I look at things like People Also Ask terms to find emerging queries and get as much data from as many places as I can. Then, I cluster that and see what the search intent is and how Google is returning those keywords from an entity perspective.”

Do you cluster keywords by meaning or by intent?

“It’s a bit of both.

Firstly, you can cluster them based on what Google is returning. If Google is saying that, for a given keyword, all these other keywords are semantically linked, that’s a cluster. Therefore, that is probably worthy of its own landing page.

Then, they can be labelled as transactional, informational, or navigational, which will let you know whether or not you need a fairly succinct landing page or a more informational, long-form piece of content.

You can use software to label the keywords by intent, then you do a bit of a sense check. I use Keyword Insights to cluster my keywords for me. The way it groups keywords is based on what Google is currently returning in the SERPs. It’s not just stating that the terms feel like they’re related. Then, it will tell you whether they’re transactional, informational, etc., which is a good indicator of the form of content that you need.

It can also help you attack any keyword cannibalisation that you’ve currently got. If you already have a site that’s optimized for those keywords, it can tell you if you’ve got competing content. It kills 3 birds with 1 stone.

That’s a good way to take a traditional approach to keyword research and make it suitable for 2025. It follows the latest algorithm, and it has machine learning elements to it as well.”

Where do keyword volume and keyword difficulty come in?

“A lot of clustering tools will give you a combined keyword volume for the cluster, not just per keyword. That takes you away from just creating pages for every single keyword and allows you to create a landing page for a cluster.

Volumes are important but, depending on your average order value and the commercials behind each keyword, I wouldn’t dismiss the lower-volume ones. Cumulatively, within that cluster, they can all contribute to a worthwhile combined search volume.

Keep an eye on volume, but don’t obsess too much about going for only the higher volume terms. If you’ve got a cluster of keywords that are quite similar, and you combine them, it can still very much be worth going for those lower-volume ones.

They can also give you an average keyword difficulty score, which you can weigh up as well. You can see your average rank for those terms and then the average competition score to prioritise which of those clusters you should attack first.”

How often do you do keyword research?

“I’ll do a pretty extensive piece right at the start of the project, as most of us do in the auditing phase. That gives you the backbone, then you can start feeding those keywords into your tools – whether that’s Keyword Insights or rank tracking tools.

I tend to refresh my zero click terms and emerging trends at least quarterly, but preferably more often than that. It all depends on the sector. I’ve worked in fast fashion where trends and demand can change almost daily whereas, in other sectors, you can refresh your keyword research a year later and it won’t be that much different. There’s a bit of common sense there.

If you know you’re working in a sector that is quite volatile, then refresh it as often as you can. I use AlsoAsked for my People Also Ask data. Chuck that through ChatGPT and it can quite quickly tell you if anything has changed thematically in terms of what people are asking for. That can give you emerging trends and keywords that weren’t there a couple of months ago.

It can be seasonality, or it can be changes in demand. The higher-volume stuff tends not to change as much, but the medium to long tail can evolve quite quickly.”

How do you decide on the style of the piece of content that needs to be produced to target a keyword phrase?

“That will usually come down to whether it’s more of an informational or commercial keyword. Commercial keywords are tricky because you need to bear CRO and UX in mind, and you can’t have too much content. You need to be selective about where that content is going and how much of it there is. For commercial terms, it’s about focusing on some head terms, looking at your entities, and considering conversion.

Informational terms will probably be longer-form pieces. It’ll be about demonstrating expertise and looking at your EEAT to show that you are the authority in your niche. It will be an informational long-form style versus your commercial, more succinct but still keyword-rich approach.”

How do you track the progress of your keyword optimization project?

“One of the main reasons why keywords are not dead, and are still hugely important in 2025, is because keyword tracking is still the best way to gauge the impact of the work that you’re doing.

The holy grail KPI is obviously traffic. If you see your organic clicks going up for a certain term or certain terms, then you know it’s doing the job. Until those clicks start to come through, the best way of tracking the impact of this work will still be through rank tracking.

I’ve used Advanced Web Ranking since 2011. The tool has changed quite a bit, and I get all sorts of features in there, which I’m sure other tools have as well. You’ve got traditional rank tracking, and you can benchmark against competitors.

For each group of keywords, you can identify who your competitors are per category. You can see the market share changing, and they release all sorts of other features all the time. Essentially, it allows you to gauge the impact of the work.

If you’re optimizing for entities, creating new landing pages, merging content, or getting rid of cannibalisation, the best way to see the impact of that is still to track keyword performance.”

Do you look at what the user does on the site as well?

“I like to do that as much as I can, as long as it doesn’t tread on the toes of the CRO or UX teams that the client may have. We don’t know exactly how Google determines engagement, but we do know that ranking is not just based on the traditional SEO KPIs, but however Google deems a page to be engaging.

Conversions are important. They say that traffic is vanity, and revenue is sanity. If that traffic is not converting, then it’s still not quite doing its job. Although there are other channels involved in that, good quality, organic, non-branded traffic tends to convert well – provided no on-site issues are preventing that from happening.”

If an SEO is struggling for time, what should they stop doing right now so they can spend more time doing what you suggest in 2025?

“Stop trying to get robots to write content for your user. Get it to help with some of the nitty gritty and laborious tasks that used to take us away from strategy. Start focusing on entities and how traditional keyword research and keyword optimization can complement that for success.

Stop trying to game the system and focus on what your users are looking for.

Don’t get robots to write your content for you. You can use it to write your briefs and collate data. At a push, you could get it to help you with your first drafts to save some time. However, I’ve still not seen much AI content that has learned how to resonate with a particular demographic or end user, if any.

It might help you optimize for keywords, but it still needs a human touch. It needs to understand your audience, which I haven’t seen it do in the wild yet.”

Charlie Whitworth is Lead Consultant at Whitworth SEO, and you can find him over at WhitworthSEO.com.