SEO 101 says the best answers finish top of search results. Search engines need to surface the most relevant content in response to users’ queries. That’s how they increase their users, the ads they show, and ultimately their revenue. It naturally follows then, that to increase performance in search, we must show search engines that content we produce is the best at answering relevant user queries.
With this in mind, there’s nothing particularly mind-blowing about an SEO roadmap based on understanding the intent of keywords. To answer a question well we need to know what a user is looking to find when searching it.
However, what’s more exciting – and what I’ll focus on here – is how aggregating intent data from a category can inform business decisions. Moving from a mind-set of “people searching this term want that outcome”, to one of “this category has these characteristics”, opens up a whole host of new opportunities.
The process I’ll explore today answers these questions:
- What are our potential customers searching?
- How do we categorise the intent of these searches?
- What can we learn from search data per intent area?
- Which topics should we target, when, and with what sort of content?
By understanding these areas in detail, we’ll be able to create a truly customer intent-lead strategy.
What are our potential customers searching?
I’m not going to spend too long here going through how to do keyword research as I’m sure many of you are familiar with the topic. In a nutshell the task is to generate a comprehensive list of terms which customers are searching in your category. The list should cover branded terms, products or services and any closely-related topics to them.
Whilst the likes of Google Keyword Planner will provide the most highly searched terms, it’s also useful to pick up more peripheral, question-led queries from tools like Answer the Public. We’ll look to create a keyword set which covers not only what people search at the point of purchase, but also the broader, less direct terms that come up through the various stages of their conversion funnel. Below is an Answer the Public Keyword visualisation for “Running Shoes”. This gives a lot of the useful “long tail” terms which customers are likely searching in their decision-making process.
What is the intent behind customer searches?
Once we’ve identified relevant customer search terms, the process of categorising intent begins. There are various different models in place for this, with Google’s Moments one of the most popular classification systems. However their nomenclature might vary, these systems have the same goal at heart – define the distinct ways people search around a category and identify what they want when typing these queries.
A few examples of intent categories we might distinguish are below, with running category terms used as examples. Notice that categories are ordered by propensity to purchase. Terms in category 1 are the furthest from an actual purchase, whilst those in 6 and 7 are closest. Categorising in this way allows us to understand users’ journey towards conversion, and how frequently searches appear at points in this journey.
- Seeking to understand a category (e.g “marathon distance”)
- Addressing a need or challenge (e.g “knee ache running”)
- Seeking broad product information (e.g “running shoes for bad knees”)
- Comparing brands or products (e.g “Nike vs Adidas running shoes”)
- Researching features of a product (e.g “Nike Zoom X foam”)
- Looking to visit a specific website (e.g “Nike”)
- Looking to make a transaction (e.g “Buy Nike Vaporfly”)
- Aftercare or product add-ons (e.g “replace Nike laces”)
Segmenting keywords by intent like this isn’t easy, and there is no golden rule that works every time. Depending on how many terms we’re looking at, sometimes it’s a manual job and sometimes it’s necessary to use broader rules and assumptions. Either way, the job is done when we can be confident we’ve distinguished the main intentions users have when searching in a category.
Frequently (often very highly searched) terms will fall into the “blurry middle” too, where no one intent stands out clearly. For example, a phrase like “running shoes” could be input throughout the customer’s journey to conversion. They could be in the early stage of research, comparing products, looking to visit a site or to make a purchase. For terms such as these it’s necessary to meet multiple intents simultaneously with our content; for example building a page that provides information, allows comparison and also fosters seamless conversion.
Comparison sites targeting vague terms like “travel insurance” do a good job of this. See the below from MoneySupermarket.com as an example. Their over 70s travel insurance page has FAQs, a prominent call to action, inspirational content and comparison details to suit the varied intentions a user may have when landing there.
What can we learn from customer intent data?
The first useful way to utilise intent data is to gauge broad characteristics of the category by ascertaining the search interest per intent stage. Depending on what’s being sold, sometimes it’s best to analyse this at a product level; other times reviewing the space as a whole is more valuable.
What intent level data categorised to this level allows us do is understand the overall “shape” of a search space. In some areas, the bulk of searches are carried out in the research phase; this shows that users require detailed information before converting, a need which we can tap into in strategy. Conversely, for other categories – especially simple purchases where price or availability are bigger determinants – it’s common to find more searches for comparison and purchase.
Below are a couple of examples of what we can find when aggregating data like this. The left hand chart shows a classic “research-led” search space, likely one with complex and high value products and services being offered. The one on the right looks more like a quick purchase-focused category that you’d likely see around cheaper FMCG purchases.
We’ll also address the seasonality of different intent areas at this point. It can inform not only SEO but also broader media decisions to be able to say things like “most people are researching at this point of year” or “we know that most purchase searches will happen in X (month) – let’s get our campaign live in Y (month) to prepare”.
How do we decide which intent areas to target?
With a good understanding of search market characteristics, it’s tempting to move straight to targeting areas of highest search volume at peak times. However, an effective strategy should also consider the likelihood of success from targeting different keyword sets.
Ascertaining likely performance is a matter of tackling two questions:
- How visible are we now across intent stages?
- Is new content we produce likely to perform?
The first question is a fairly simple one to answer. It’s a matter of calculating existing visibility and weighing against that which we’d see if leading the space – this gives a view of our currently tapped topic ownership. The key learning here is where search engines think our site is relevant already. Optimising in areas where we (or competitors) have existing market share is likely to require less investment than targeting entirely new ones.
Armed with current visibility, it’s time to address that second question of how likely we’d be to succeed in different intent areas. It’s all well and good knowing that there are untapped searches, but there’s no point going after a topic we’re never going to perform for.
A useful thing to remember when carrying out this kind of analysis is that Google itself is about the best intent categoriser on Earth. For the vast majority of searches, the engine will provide useful answers. It follows naturally then, that a clear view on what’s ranking now will help us assess how our site weighs up.
For this reason, the next step is to work out which types of content are most visible across different keyword sets. Using this data, it’s possible to gauge whether ours is likely to perform well if content is produced targeting different query types. Again, it’s a matter of answering some broad questions:
What type of sites are leading in search results?
Different types of content performs for different types of queries. Seeing that sites like ours don’t show up in a space doesn’t necessarily mean performing is impossible, but there is likely a higher barrier to entry if trying to rank where we’re not naturally an authority. Conversely, if the results pages are full of us and our adversaries, we’ll likely have an easier time.
One common mistakes is for companies to overly target vague but hugely searched informational terms. Search engines will give far more of a leg up to bipartisan publishers and official bodies for such queries, due to these sites’ “EAT” signals and established topical authority, so it’s often not worth your while to target them without serious investment. Seeing, for example, the search results for “New York Guide” below I’d be hesitant to create content targeting the topic with so many informational sites & specialists taking up the top spots.
What signals are we getting from ranking content?
It’s also important to understand how “strong” the content we’re up against is in search, and therefore how difficult it will be to perform there. As with any competitive pursuit, the stronger the competition is, the tougher it will be for us to win. We’ll use metrics like keyword difficulty, backlinks and word counts to ascertain this, as well as reviewing the degree to which competitors have invested in performing for a keyword set.
It’s often the case that competitors have setup whole on-site hubs, microsites or content series’ to tap in to a desired topic area. The level of competitor investment should give clues as to how much we’ll have to put in to lead the space, and whether it’s better targeting niche or high volume keyword sets. A simple way to graph this out is like the below, directly weighing difficulty against search opportunity. Areas where volumes are high and difficulty (however defined) is low are likely to be beneficial for targeting. I’d look created content with in the “Addressing a Need” or “Researching Features” categories in this specific example.
Combining data to make decisions
With a clear view on what people search, how often, our current market share and the relative competitiveness of different areas, we have all that’s required to make strategic targeting decisions. We’ll dig right into the data and take insights like:
- We already have intermediate visibility in Area 1, and the search results are full of sites like ours. We should invest in growing the position of our existing content at the peak time of year. If my Christmas pudding recipe was smashing it online, maybe I’d consider a “best Christmas pudding ice cream” article to complement it (pun intended).
- Area 2 is dominated by comparison sites, and we’re going to have a hard time being taken seriously here. We should partner with aggregators & affiliate sites when “vs” searches are high to piggyback on their visibility.
- We can see that the barrier to entry in Area 3 is very high, but search opportunity is huge and competitors are ranking. We need additional media support – and to build out a full on-site campaign – in order to perform.
- We should deprioritise Area 4; the search opportunity is relatively low and the barrier to entry is high. The only sites we see ranking are more specialist than we ourselves are. The example above around “New York City Guides” would be an example of such a topic.
I hope you’ve found this an interesting read. Let me know in the comments below if you have any questions or additional thoughts.