Category Archives: Keyword Research

Explicit and implicit search

Go to Google and search for “restaurant”.

Chances are that a number of the results are local to you.

This is the result that I get:

Restaurant Search

The third organic result is for

How can that be?

I searched for “restaurant” – not “restaurant Witney”.

This is as a result of implicit search.

In the past, search engines have typically worked like this:

The user searches for a keyword – for example “restaurant Witney”.

The search engine then returns a set of results based on what was searched for.

There are now two things that happen in search.

Explicit search and implicit search.

The old way was explicit.

The search engine user would explicitly state what they were looking for.

Implicit is the stuff that the user does not consciously provide.

It can be their location.

The device they are on.

Their search history.

All of the things that Google knows about the searcher that the searcher has not explicitly provided.

The combination of explicit and implicit search is starting to fundamentally change the way we use search engines (and in particular, Google).

Back to the example search above.

I am sitting in an office in Witney, Oxfordshire.

I carried out an explicit search for the keyword “restaurant” on Google.

To have a local restaurant rank number three for such a broad term was previously unheard of.

But when we consider implicit search, the term “restaurant” no longer just means what it explicitly says.

Google has used my implicit search details to show me results that it thinks I want to see.

Despite making no conscious effort to provide this information, Google knows where I am.

That is implicit search and it means that the keyword is no longer at the centre of the search.

It means that a local restaurant can appear in a vastly elevated position for an ultra competitive search term.

It means that measuring rankings is a wild goose chase (more than it was already).

I might be seeing one restaurant, you may be seeing something completely different.

The result of this is that the keyword research model of SEO is going to have to adapt.

Planning and analysis is going to have to rely more heavily on the data gleaned from analytical tools.

Implicit search is also going to bring online results more in-line with offline domination.

If a brand has 100 physical shops across the country and another has only one, the large brand is potentially 100 times more likely to be in the localised search results thanks to implicit search – regardless of who has the better website.

Implicit search means that it is not all about what you say but more about who you are.


This post was inspired by this video from Will Critchlow and Tom Anthony at Distilled:


A Simple Way to Improve Your Keyword Research

Out of all of the SEO tasks that fill my to-do list on a regular basis, keyword research is the one that I most love to see. Keyword research, as I am sure you are aware, is an important part of running an SEO project. It is so important that if it hasn’t been done properly I insist that a project stops until it has been carried out in a comprehensive way.

Keyword research has a very simple aim: to find keywords that have a high amount of searches are relevant and have a low amount of competition. All too often the keyword research is stifled by a lack of time and budget forcing SEOs to carry out half-baked work that causes them to miss one or more of these requirements. I want to present a new process that I believe can be as quick as any other but yields vastly more reliable results that always take into account the fundamental aims of keyword research.

The current keyword research process goes a little like this; log into a keyword gathering tool, enter a search using some keyword ideas, export the resulting list into Excel. The list is then looked over and with nothing more than subjective opinion and local monthly search volume a shortlist is chosen. Sometimes this process works but when it does I would wager that it was pure luck rather than SEO expertise that brought success.

Simple Keyword Research Process

The new process I want to present is designed to take all of the guesswork out of keyword research. Let’s go back to our list of aims that keyword research should focus on. Number one is to find keywords that have a high search volume. Taking care of this part of the process should be very familiar and it does involve simply using the keyword tool of your choice. Use the tool to gather as many keywords as you can handle, search and re-search using different variations and by all means use more than one tool. Don’t get bogged down by duplicates cropping up when you export lists, this can be easily filtered out in Excel. The aim here is to gather as many potentially relevant keywords as you can, from as many sources as you can. The key thing to remember is that the list of keywords you have need to have a localised search volume associated with them. You should only limit this stage of the process by the amount of time you have to play with.

The next aim of keyword research is to find keywords that are relevant. Normally the only way to judge relevancy is to go down the list of keywords and use your opinion to say whether a keyword is relevant or not. This binary approach does not work because some keywords are more relevant than others. This means we need to quantify the process to give each keyword a fair chance. It is here that I like to use the following method: imagine we are doing SEO for a pet supplies website – take a keyword, for example ‘dog house’ and give it a relevancy score of 1.0 – now for every other thing that the keyword can describe remove 0.1. It may be that there is a local pub called ‘The Dog House’ so remove 0.1, then there is a film of the same name – remove another 0.1. So that leaves this keyword with a relevancy score of 0.8, make a note of this in a new column within your spreadsheet next the local monthly search volume.

The third and final fundamental aim of keyword research is to find keywords that have low levels of competition. To quantify the level of competition we need to use a search engine and a bit of basic maths. Carry out a search for the keyword you are analysing. Again you must start this process by giving your keyword a competition score of 1.0. Once the search results are returned you must remove 0.1 from the competition score for every result that has the keyword you are analysing in either the title or description of the results. In the case of ‘dog house’ there are two organic results in that have the keyword in their title and / or description. This means the competition score for ‘dog house’ is 0.8. Add this to another column within your spreadsheet.

The Final Keyword Research Formula

We now have in our spreadsheet a list of keywords, their local search volumes, a relevancy score and a competition score. The information is all numerical which allows us to combine all of the data to score each keyword fairly. To combine the data you must simply complete the following formula:

Local monthly search volume X relevancy X competition = keyword opportunity score

In the case of our keyword ‘dog house’ the formula might be as follows:

2500 X 0.8 X 0.8 = 1600

What is the advantage of this method? Let’s consider that you also have the keyword ‘cheap dog house’ in your list. This keyword may have a local search volume of 6000 and on the face of it seem like an excellent keyword choice. Then imagine that we carry out the research and find it to have a relevancy score of 1.0 but a competition score of 0.3. The formula would look like this:

6000 X 1.0 X 0.2 = 1200

When we take into account all factors we can immediately see that this keyword ends up with a lower final score and that optimising for ‘cheap dog house’ could be a bad move because the high levels of competition on this keyword are not made up for by the relevancy or search volume. This is invaluable information and can turn a mediocre SEO project into a winning one.

I would strongly recommend that all SEOs employ this keyword research method on future projects when time is tight, it certainly isn’t as advanced as keyword research gets, to see what else can be done I would point you in the direction of Richard Baxter, however it is certainly a robust and fair way to test all keywords and end up with a reliable list.